Table of Contents About This Reference Manual ....................................................................................... 1 Who Should Use This Manual? ................................................................................... 2 Contact Language Logic ........................................................................................... 3 Third-Party Web Sites .............................................................................................. 3 Feedback on Documentation ..................................................................................... 3 Introduction to Ascribe™ ............................................................................................. 3 What is Coding? ...................................................................................................... 3 What is Ascribe™? ................................................................................................... 4 Access ................................................................................................................... 4 Ascribe™ Architecture .............................................................................................. 5 Data Security ....................................................................................................... 5 Additional Backup Provisions .................................................................................. 5 Security from Loss of Service ................................................................................. 6 Security from Server Intrusion ............................................................................... 6 Security from Website Intrusion ............................................................................. 6 Security in Data Transit ......................................................................................... 7 Multi-Lingual, Unicode Compliant Application ............................................................ 7 What's New ............................................................................................................ 7 Maintenance Schedule.............................................................................................. 8 Daily Maintenance Operations ................................................................................ 8 Weekly Maintenance Operations ............................................................................. 9 Monthly Maintenance Operations ............................................................................ 9 Legal Agreements .................................................................................................... 9 Get Started ............................................................................................................. 10 Configure Your Internet Browser ............................................................................. 10 Version of Microsoft Internet Explorer ................................................................... 10 AscribeClient.com Component .............................................................................. 10 Use Secure Communication.................................................................................. 10 i PrintDoc01 020813 Configure Ascribe™ as a Trusted Site .................................................................... 11 Security Setting for Trusted Sites ......................................................................... 11 Additional Configuration for IE 7 ........................................................................... 12 Set Your Language Preference .............................................................................. 12 Low Resolution Monitors ...................................................................................... 12 IE 8 Accelerator Icon ........................................................................................... 13 Configure Word for Spell-Check in Ascribe™ ............................................................. 13 Word Version 10.2 .............................................................................................. 13 Word Version 9.0 (Word 2000) ............................................................................. 14 FTP Setup ............................................................................................................. 15 Log In .................................................................................................................. 15 Home Page and Main Menu ..................................................................................... 16 Navigation (Popup Menus) ...................................................................................... 16 Application Help .................................................................................................... 16 Customization Features .......................................................................................... 17 Set User Options ................................................................................................ 17 Choose Columns ................................................................................................. 18 Change Column Widths ....................................................................................... 18 Resize Windows .................................................................................................. 19 Language Support .............................................................................................. 19 Open Sessions ...................................................................................................... 19 Downloads ............................................................................................................ 19 Log Off ................................................................................................................. 19 Study ..................................................................................................................... 20 Studies Page ......................................................................................................... 20 Filter Options ..................................................................................................... 20 Columns on the Studies Page ............................................................................... 20 Studies Page Right-Click Menu ............................................................................. 22 Assign a Coder to a Study ...................................................................................... 23 ii Table of Contents Assigned To Dialog Box from the Studies Page ....................................................... 23 More Notes on Assigning Users by Question Types .................................................. 24 How to Make Changes ......................................................................................... 25 Create a Study ...................................................................................................... 25 Add a Study ....................................................................................................... 25 General Tab (Study) ........................................................................................... 25 Data Tab (Output Data Formats) .......................................................................... 27 Date Tab ........................................................................................................... 28 Copy Study ........................................................................................................... 29 Save a Study ........................................................................................................ 30 Save Study Options ............................................................................................ 30 Job Submitted Page ............................................................................................ 31 Restore a Study .................................................................................................... 32 Restore Types .................................................................................................... 32 Restore Shared Codebooks .................................................................................. 33 How to Restore a Study ....................................................................................... 34 Merge Study ......................................................................................................... 35 How to Merge a Study ......................................................................................... 36 Restore and Merge via the FTP Site ....................................................................... 37 Study Questions Page ............................................................................................ 38 Add and Edit Questions for a Study .......................................................................... 41 Question Types .................................................................................................. 43 Question Components ......................................................................................... 44 General Tab (Question) ....................................................................................... 44 Tasks Tab .......................................................................................................... 45 Coding Tab ........................................................................................................ 46 Data Tab ........................................................................................................... 50 Related Questions ................................................................................................. 52 Delete a Study and/or Components ......................................................................... 52 iii PrintDoc01 020813 Delete a Study ................................................................................................... 53 Delete Questions ................................................................................................ 53 Delete Loaded Verbatims ..................................................................................... 53 Delete Coded Responses...................................................................................... 54 Codebooks .............................................................................................................. 55 Codebook Components (Codes and Nets) ................................................................. 55 Presentation in Ascribe™ ........................................................................................ 55 Create and Manage Codebooks ............................................................................... 56 Review Responses and Edit the Codebook Right-Click Menus ...................................... 57 Edit the Codebook Window ..................................................................................... 58 Codebook Editor .................................................................................................... 59 Left Pane Features .............................................................................................. 59 Codebook Editing Tool Menu Bar ........................................................................... 62 Codebook Display ............................................................................................... 66 Concurrent Edits ................................................................................................. 66 Right-Click Menus ............................................................................................... 66 Navigation Keystrokes ......................................................................................... 69 Set Color Dialog ................................................................................................. 69 Copy Text Dialog ................................................................................................ 70 Paste Text Dialog ................................................................................................ 70 Change Case Dialog ............................................................................................ 71 Renumber the Codebook ........................................................................................ 71 Renumber with Edit the Codebook or Review Responses .......................................... 71 Renumber with Codebook Editor ........................................................................... 72 Sort Codes ........................................................................................................... 74 Define Codebook Properties .................................................................................... 75 Edit the Properties of a Code ................................................................................... 76 Add Codes ............................................................................................................ 77 Spell Check........................................................................................................... 78 iv Table of Contents Regular Expressions............................................................................................... 80 Code in the Coder Window with Expressions .......................................................... 80 Simple Regular Expressions ................................................................................. 80 Match Words ...................................................................................................... 81 Verbatim Text Normalization ................................................................................ 82 Match Phrases .................................................................................................... 83 Match Negative Phrases ....................................................................................... 84 Match Commonly Misspelled Words ....................................................................... 84 Match Multiple Cases ........................................................................................... 84 Advanced Use of Regular Expressions.................................................................... 84 Check Expressions for a Specific Codebook ............................................................ 86 Copy and Paste Codebooks from Excel® or Word ...................................................... 87 Delete Codes ........................................................................................................ 89 Move Codes .......................................................................................................... 89 Share and Copy Codebooks .................................................................................... 91 Difference Between Copy and Share...................................................................... 92 Copy/Share Manager........................................................................................... 92 How to Unshare Questions ................................................................................... 96 Unshare Questions with the Copy/Share Manager ................................................... 96 Copy/Share From Dialog ...................................................................................... 96 Copy/Share To Dialog ......................................................................................... 97 Rules for Copy and Share Codebooks with Edit the Codebook ................................... 98 Copy and Share Codebooks with Edit the Codebook ................................................ 99 Copy and Share Codebooks to Multiple Questions with Edit the Codebook ................ 100 Find Where a Code Has Been Applied in Shared Codebooks .................................... 101 Unshare Codebooks with Edit the Codebook ......................................................... 101 Account Codebooks ............................................................................................. 102 Verbatims ............................................................................................................. 105 Studies in Progress .............................................................................................. 105 v PrintDoc01 020813 Columns on the Studies in Progress Page ............................................................ 106 Studies in Progress Right-Click Menu................................................................... 107 Questions in Progress .......................................................................................... 107 Fields on the Questions in Progress Page ............................................................. 108 Questions in Progress Right-Click Menu ............................................................... 109 Code Verbatims in the Coder Window..................................................................... 110 The Coder Window Defined ................................................................................ 110 Question Pane .................................................................................................. 110 Verbatim Pane.................................................................................................. 110 Codebook Pane (Coder Window) ......................................................................... 111 Notes Pane ...................................................................................................... 112 Selected Codes Pane ......................................................................................... 112 Related Responses Pane .................................................................................... 113 Summary Pane ................................................................................................. 113 How to Code in the Coder Window ...................................................................... 113 Right-Click Menu .............................................................................................. 114 Code Verbatims in the Review Responses Window ................................................... 121 The Review Responses Window Defined ............................................................... 122 Response Pane ................................................................................................. 122 Codebook Pane ................................................................................................ 122 Button Pane ..................................................................................................... 124 Select Responses and Apply Codes in Review Responses ....................................... 125 Right-Click Menu in the Response Pane ............................................................... 126 See Who Applied a Code to a Response ............................................................... 128 See the Question ID When You Use Shared Codebooks ......................................... 129 See How Many Responses You Have Left to Code in a Question .............................. 129 View Options .................................................................................................... 130 Swap One Code for Another ............................................................................... 133 Create a New Code from a Combination of Codes ................................................. 135 vi Table of Contents Code Verbatims in the Phrase Analyzer Window ...................................................... 135 The Phrase Analyzer Window Defined .................................................................. 136 The Parsed Responses Pane ............................................................................... 136 Right-Click Menu in the Parsed Responses Pane .................................................... 137 How to Parse Responses .................................................................................... 138 The Options Pane ............................................................................................. 140 The Codebook Pane .......................................................................................... 141 The Phrase Analyzer Coding Process ................................................................... 142 Multi-Media Coding .............................................................................................. 144 Transcribe IVR and Image Files .......................................................................... 145 Audio Keyboard Shortcuts .................................................................................. 146 Edit Verbatims .................................................................................................... 146 Edit Respondent and Edit Responses ................................................................... 146 Edit Respondent Window ................................................................................... 147 Edit Responses Window ..................................................................................... 151 Translate/Transcribe Verbatims ............................................................................. 156 Questions to Transcribe/Translate ....................................................................... 156 Considerations for Translation Using Ascribe™ ...................................................... 157 Translate/Transcribe Window ............................................................................. 157 Translate/Transcribe Settings ............................................................................. 158 Controls in the Banner Bar ................................................................................. 158 Controls in the Source Pane ............................................................................... 159 Controls in the Translation Pane ......................................................................... 159 Audio Keyboard Shortcuts .................................................................................. 161 Counts of Responses ......................................................................................... 161 Apply to Duplicates ........................................................................................... 161 Add or Edit a Translation ................................................................................... 162 Auto Translate.................................................................................................. 162 Filter - Set a Response Filter ................................................................................. 163 vii PrintDoc01 020813 Filter Dialog ..................................................................................................... 163 How to Tell if a Filter is Set ................................................................................ 164 Client Access to Data .............................................................................................. 166 Client Reports ..................................................................................................... 166 Coding Overview Report .................................................................................... 166 Study Summary Report ..................................................................................... 167 Question Summary Report ................................................................................. 169 Card Layout Report ........................................................................................... 172 Study Report .................................................................................................... 173 Study Quality Report ......................................................................................... 176 View Codes by Coder Report .............................................................................. 179 Respondents Report .......................................................................................... 180 Cost Accounting .................................................................................................. 183 Fee Schedule ................................................................................................... 183 Transaction Reports .......................................................................................... 184 Account Statement ........................................................................................... 186 Time Accounting .................................................................................................. 187 How Ascribe™ Tracks Time ................................................................................ 187 Session, Logon, and Tracked Hours ..................................................................... 188 Count of Responses Coded in Time Tracking Records ............................................ 189 Use Time Accounting for Job Cost Calculations ..................................................... 189 Time Accounting Reports ................................................................................... 189 Production Bands Report ................................................................................... 202 Ascribe Presentations™ ........................................................................................... 207 Base and Professional Editions .............................................................................. 207 Ascribe Presentations Main Page............................................................................ 208 Cross-Tabs ......................................................................................................... 209 Questions Tab .................................................................................................. 209 Codes Tab ....................................................................................................... 212 viii Table of Contents Words Tab ....................................................................................................... 214 Graphs Tab ...................................................................................................... 218 Options Tab ..................................................................................................... 220 Filters Tab ....................................................................................................... 223 Results Page .................................................................................................... 224 Multi-Media Files ............................................................................................... 225 Create Bar, Line, or Pie Charts .............................................................................. 226 Bar Chart Examples .......................................................................................... 229 Line Chart Examples ......................................................................................... 230 Pie Chart Example ............................................................................................ 231 Co-Occurrence Maps ............................................................................................ 232 Correlation Matrices ............................................................................................. 236 Co-Occurrence Maps and Correlation Matrices Across Questions ................................ 238 Proportional Code Font Views ................................................................................ 239 Multidimensional Scalings ..................................................................................... 241 Clusters ............................................................................................................. 243 Multiple Charts .................................................................................................... 245 Word Clouds ....................................................................................................... 245 Questions Tab .................................................................................................. 246 Words Tab ....................................................................................................... 247 Options Tab ..................................................................................................... 247 Filters Tab ....................................................................................................... 249 Results Page .................................................................................................... 249 Concept Clouds ................................................................................................... 251 Questions Tab .................................................................................................. 252 Options Tab ..................................................................................................... 252 Filters Tab ....................................................................................................... 254 Results Page .................................................................................................... 254 Cross-Tab (Legacy) ............................................................................................. 255 ix PrintDoc01 020813 Select Cross-Tab Questions ................................................................................ 256 Fill the Cross-Tab Table with Data ....................................................................... 257 Expand and Collapse Rows ................................................................................. 257 Drill Down to View Responses ............................................................................ 258 Filter the Cross-Tab View ................................................................................... 258 Save and Export Charts ........................................................................................ 258 Drill Down Tools .................................................................................................. 260 Ascribe Inspector™................................................................................................. 263 How to use Inspector ........................................................................................... 263 Inspector Privilege ............................................................................................... 263 Inspection Ownership........................................................................................... 263 Inspector/Inspections Page ................................................................................... 263 Inspections Page Layout .................................................................................... 264 Inspections Page Right-Click Menu ...................................................................... 264 Create an Inspection from a Study Question ........................................................... 265 Create an Inspection from Scratch ......................................................................... 266 Load Data to an Inspection ................................................................................... 267 Inspection Page................................................................................................... 267 General Information Section ................................................................................. 269 Reset Processors .............................................................................................. 270 Search ............................................................................................................... 271 Search Criteria ................................................................................................. 271 Advanced Search Options .................................................................................. 272 Search Options ................................................................................................. 273 Code This Search .............................................................................................. 274 Prior Coded Searches ........................................................................................ 276 Curate ............................................................................................................... 277 Right-Click Menu .............................................................................................. 277 Edit Individual Response .................................................................................... 278 x Table of Contents Options for Curate ............................................................................................ 278 Search and Replace .......................................................................................... 278 Update All ........................................................................................................ 280 Remove ........................................................................................................... 280 Sentiment Ratings ............................................................................................... 281 List View for Ratings ......................................................................................... 282 Chart View for Ratings....................................................................................... 284 Sentiment Topics ................................................................................................. 285 List View for Topics ........................................................................................... 287 Chart View for Topics ........................................................................................ 290 Chart Options ................................................................................................... 291 Group ............................................................................................................. 292 Export ............................................................................................................. 293 Adding and Analyzing Additional Data for Sentiment ............................................. 294 Concepts ............................................................................................................ 294 List View for Concepts ....................................................................................... 295 Chart View for Concepts .................................................................................... 298 Adding and Analyzing Additional Data for Concepts ............................................... 298 Entities .............................................................................................................. 299 List View for Entities ......................................................................................... 299 Accelerator ......................................................................................................... 300 Create an Accelerator from an Inspection ............................................................ 300 Edit Accelerator Training Examples ..................................................................... 301 Set Filters ........................................................................................................... 303 Filtering Within a Filter Variable .......................................................................... 306 Filtering With Multiple Filter Variables .................................................................. 306 Save to Ascribe or Export to Excel While Filtering ................................................. 306 Variables Page .................................................................................................... 307 Page Layout ..................................................................................................... 307 xi PrintDoc01 020813 Add a New Variable ........................................................................................... 308 Label the Data in a Variable ............................................................................... 308 Insert or Remove Labels .................................................................................... 308 Create an Account Codebook .............................................................................. 308 Navigate to Review Responses ........................................................................... 309 Remove a Variable ............................................................................................ 309 Variable Properties............................................................................................ 309 Return to Inspector/Inspections ......................................................................... 309 Data Management .................................................................................................. 309 How to Load Data ................................................................................................ 309 Load Tab Delimited Data ................................................................................... 309 Load Excel® Files ............................................................................................. 310 Load File Types ................................................................................................ 311 Define Your Own Load File Types ........................................................................ 314 Load Study Setup Information ............................................................................ 318 Load Data to a Study ........................................................................................ 319 View and Delete Loaded Data ............................................................................. 320 Loads by Date .................................................................................................. 321 Load Details ..................................................................................................... 321 Loads by Job .................................................................................................... 322 Load Voice Recordings ....................................................................................... 323 Load Image Files .............................................................................................. 323 Auto FTP Setup ................................................................................................ 323 Auto-Apply Codes ............................................................................................. 324 Download Data ................................................................................................... 329 Download Study Results .................................................................................... 329 User Defined Study Results ................................................................................ 331 Substitution Values for User Defined Study Results ............................................... 332 Download Codebooks ........................................................................................ 334 xii Table of Contents User Defined Codebook Formats ......................................................................... 335 Substitution Values for User Defined Codebook Formats ........................................ 336 Download Data with Scripted Output ................................................................... 340 Manage Jobs ....................................................................................................... 356 View Detailed Information about a Job ................................................................ 357 View Parameters for Scripted Output Jobs ........................................................... 358 Download the Results of a Job ............................................................................ 358 Resume a Paused Job ........................................................................................ 358 Schedule Jobs ..................................................................................................... 359 Set Up a Daily Schedule .................................................................................... 361 Set Up a Weekly Schedule ................................................................................. 361 Set Up a Monthly Schedule ................................................................................ 361 Repeat a Job .................................................................................................... 361 Change a Schedule ........................................................................................... 362 Delete a Schedule ............................................................................................. 362 View the Parameters for a Scheduled Job ............................................................ 362 Run a Job Now ................................................................................................. 363 Ascribe™ Study Shipping ...................................................................................... 363 Study Shipping Features .................................................................................... 364 Shipped Studies Page ........................................................................................ 364 Configure Study Shipping .................................................................................. 365 Set Up Your FTP Site ......................................................................................... 365 Maintain Your FTP Site ...................................................................................... 366 Use Encryption ................................................................................................. 366 File Management .............................................................................................. 366 Repository .......................................................................................................... 367 Display Studies ................................................................................................ 368 Find Box .......................................................................................................... 368 Show Details .................................................................................................... 368 xiii PrintDoc01 020813 Download a Study from the Repository ................................................................ 370 View the Questions in the Repository .................................................................. 370 Study Archive (Save and Restore) XML Specification ................................................ 370 Structure ......................................................................................................... 370 XML Document Outline ...................................................................................... 371 Tag Descriptions ............................................................................................... 372 Media Types ..................................................................................................... 383 Write XML Documents ....................................................................................... 384 Parse XML Documents ....................................................................................... 384 Example of a Study with No Questions ................................................................ 384 Example of a Study with Three Questions and Coded Responses............................. 385 Accelerator Coding Model ........................................................................................ 391 ACM Definitions ................................................................................................... 391 ACM Theory of Operation ...................................................................................... 391 Code Matching .................................................................................................... 392 ACM Architecture ................................................................................................. 392 VCS™ - Verbatim Coding System .......................................................................... 393 How the Verbatim Coding System Works ................................................................ 393 Creation of Binary Classifiers .............................................................................. 393 Coding Using Binary Classifiers ........................................................................... 394 Lessons Learned ............................................................................................... 395 Frequently Asked Questions About VCS™ ............................................................ 396 VCS™ Training Metrics ......................................................................................... 396 Accuracy at the Individual Level (F1) .................................................................. 397 Estimating Accuracy .......................................................................................... 398 Accuracy at the Aggregate Level ( PD) ................................................................ 398 Multi-Lingual Models ............................................................................................ 398 When to Use ACM ................................................................................................ 399 Trackers .......................................................................................................... 399 xiv Table of Contents Large One-Time Studies .................................................................................... 400 Codes Common to Multiple Codebooks ................................................................ 400 Questions Common to Multiple Studies ................................................................ 400 Multi-Lingual Projects ........................................................................................ 400 ACM Fees ........................................................................................................... 400 How ACM Fees are Applied ................................................................................. 401 ACM Transaction Reporting ................................................................................ 401 ACM Process on the Web ...................................................................................... 402 Create and Train a Model ................................................................................... 402 Use the Model for Coding ................................................................................... 402 Improve a Model .............................................................................................. 403 Maintain the Model............................................................................................ 403 Create a Model from a Question ............................................................................ 403 Accelerator Coding Models Page ............................................................................ 405 Right-Click Menu on the Accelerator Coding Models Page .......................................... 406 Noise Filters ........................................................................................................ 407 Constructing a Noise Filter ................................................................................. 408 Special Characters for a Noise Filter .................................................................... 408 ACM Properties Dialog .......................................................................................... 409 Training Example Editor ....................................................................................... 411 Filter Bar ......................................................................................................... 411 Training Results Section .................................................................................... 412 Training Examples Table .................................................................................... 413 Copy Examples from Shared Questions .................................................................. 415 Considerations Before Approving a Model ............................................................... 416 Assess the Quality of Training Results ................................................................. 417 Approve a Model ............................................................................................... 417 Unapprove a Model ........................................................................................... 417 Training Results Page ........................................................................................ 418 xv PrintDoc01 020813 Training Code Cleaner ....................................................................................... 419 Code Questions using a Model ............................................................................... 420 Run an ACM Coding Job ..................................................................................... 420 Assess the Quality of ACM Coding ....................................................................... 420 Improve ACM Coding Quality ................................................................................ 420 Number of Training Examples per Code ............................................................... 420 Quality of Training Examples .............................................................................. 421 Conceptual Clarity of Codes ............................................................................... 421 Specificity of Coding.......................................................................................... 421 Extraneous Text in Responses ............................................................................ 421 Negative Examples ........................................................................................... 421 Divide Responses into Concepts ......................................................................... 422 Accelerator Coding Model Page .............................................................................. 423 ACM Properties Table ........................................................................................ 423 Languages Table .............................................................................................. 423 Coding Questions Table ..................................................................................... 424 Right-Click Menu on the Accelerator Coding Model Page ........................................ 424 Right-Click Menu for the Coding Questions Table .................................................. 425 Question Selection Dialog .................................................................................. 426 Ascribe™ Desktop .................................................................................................. 426 Install Ascribe™ Client Applications: Desktop and Crossword .................................... 426 System Requirements ....................................................................................... 426 Installing Crossword/Desktop ............................................................................. 427 Common Key Usage............................................................................................. 428 Terminology ....................................................................................................... 428 Log On ............................................................................................................... 429 Log Off ............................................................................................................... 430 Refresh .............................................................................................................. 430 Grid Editing ........................................................................................................ 431 xvi Table of Contents Main Display and Navigation ................................................................................. 431 Main Menu ....................................................................................................... 431 Toolbar............................................................................................................ 432 Navigation Pane ............................................................................................... 433 Data Pane ........................................................................................................ 435 General Navigation Concepts .............................................................................. 435 Options and Customization ................................................................................ 439 Using the Workspace......................................................................................... 444 Using 'My Folders' ............................................................................................. 446 Account Codebooks in Desktop ........................................................................... 448 Desktop Codebook Editor ..................................................................................... 448 Codebook Properties Editor ................................................................................ 449 Codes Editor .................................................................................................... 450 Codebook Renumber Options ............................................................................. 456 Codebook Sort Options ...................................................................................... 457 Working with Spell Check Results ....................................................................... 457 Paste Codes into a Codebook ............................................................................. 458 Text AutoCoder ................................................................................................... 458 Start the Text AutoCoder ................................................................................... 458 Process Overview ............................................................................................. 458 Multiple Responses Coding Tool ............................................................................. 459 Top Toolbar ..................................................................................................... 460 Response Pane ................................................................................................. 461 Tabbed Pane .................................................................................................... 462 Set View Options .............................................................................................. 465 Code with Multiple Responses Coding Tool ........................................................... 466 Codebook Merge/Map Tool .................................................................................... 469 Merge Codebooks ............................................................................................. 469 Merge Mode Toolbar Options .............................................................................. 469 xvii PrintDoc01 020813 Map Codebook Codes ........................................................................................ 470 Mapping Mode Toolbar Options ........................................................................... 471 Working with the Match View ............................................................................. 472 Study Import Tool ............................................................................................... 473 Top Toolbar ..................................................................................................... 473 Left-Side Toolbar .............................................................................................. 473 Add a New Study .............................................................................................. 474 Add Questions to an Existing Study..................................................................... 474 Study Merge Tool ................................................................................................ 474 Study Merge Tool Functions ............................................................................... 475 Study Merge Tool Copy Process Description ......................................................... 475 Samples ............................................................................................................. 476 How to Create a Sample .................................................................................... 476 Create Quick Sample Dialog Box ......................................................................... 479 How to Use a Sample ........................................................................................ 479 View a List of Samples ...................................................................................... 480 Delete a Sample ............................................................................................... 480 Automated Coding Model (Desktop) .......................................................................... 481 ACM Process on Desktop ...................................................................................... 481 List of Automated Coding Models ........................................................................... 483 Edit ACM Dialog................................................................................................... 485 Training Example Languages ................................................................................. 487 Multi-Lingual Model .............................................................................................. 488 Provide Training Examples .................................................................................... 489 Copy Responses From a Study Question to the Model ............................................ 489 Load a File of Responses to a Model Using the Right-Click Menu in the Navigator ...... 490 Import Examples Using the Training Example Editor ............................................. 490 Listen for Copy Using the Training Example Editor ................................................ 490 Enter New Examples Using the Training Example Editor ......................................... 490 xviii Table of Contents Desktop Training Example Editor ........................................................................... 490 Quick Start ...................................................................................................... 490 Editor Toolbar .................................................................................................. 491 Training Example Editor View Options ................................................................. 492 Examples Pane ................................................................................................. 493 Delete Examples ............................................................................................... 494 Training Example Editor Tabbed Pane .................................................................. 495 ACM Coding ........................................................................................................ 499 Associate the Coding Question with the Model ...................................................... 499 Start the Coding Job ......................................................................................... 499 Coding Questions Table ..................................................................................... 499 Coding Time Options ......................................................................................... 501 Review ACM Coding .......................................................................................... 501 Review ACM Coding Window ................................................................................. 501 Accept Suggested Codes .................................................................................... 501 ACM Suggestion Confidence Threshold ................................................................ 502 Toolbar............................................................................................................ 502 Response Pane ................................................................................................. 503 View Options for ACM Review Coding Window ...................................................... 504 Tabbed Pane for Review ACM Coding Window ....................................................... 505 Ascribe™ Crossword ............................................................................................... 506 Installing Crossword ............................................................................................ 506 Crossword Logon ................................................................................................. 506 Understanding a Crossword Report ........................................................................ 507 Dimensions ...................................................................................................... 507 The Crossword Table ......................................................................................... 507 Crosswords Page ................................................................................................. 507 Open Crossword .................................................................................................. 508 New Crossword ................................................................................................... 508 xix PrintDoc01 020813 New Crossword Wizard ......................................................................................... 508 Select Study for Crossword ................................................................................ 509 Select Row Question(s) ..................................................................................... 509 Select Column Question(s)................................................................................. 509 Select Hidden Question(s) ................................................................................. 509 Troubleshooting................................................................................................ 509 Save Crossword .................................................................................................. 509 Crossword Table Right-Click Menu ......................................................................... 510 Responses Dialog ................................................................................................ 511 Filtering ........................................................................................................... 511 Saving ............................................................................................................ 511 Crossword Properties Dialog ................................................................................. 512 Show Dimensions ................................................................................................ 513 Dimensions List ................................................................................................ 513 Dimension View ................................................................................................ 514 New Dimension Wizard......................................................................................... 516 Publish a Crossword ............................................................................................. 516 Filter a Crossword Table ....................................................................................... 517 Graph a Crossword .............................................................................................. 517 t-Tests Between Proportions ................................................................................. 518 Create a Co-Occurrence Map from Crossword.......................................................... 518 User Administration ................................................................................................ 521 Privilege Levels ................................................................................................... 521 Associates .......................................................................................................... 521 Add/Edit an Associate ....................................................................................... 522 Delete an Associate........................................................................................... 523 Company ............................................................................................................ 523 Add/Edit a Company ......................................................................................... 523 Delete a Company ............................................................................................ 523 xx Table of Contents Manage Company Contacts ................................................................................ 523 Past Sessions ...................................................................................................... 525 Reset Passwords ................................................................................................. 525 Locked Responses ............................................................................................... 525 Site Administration ................................................................................................. 526 Translate the Ascribe™ Interface ........................................................................... 526 Change Ascribe™ Text ....................................................................................... 526 Reload Strings .................................................................................................. 528 Site Configuration ................................................................................................ 528 Frequently Asked Questions ..................................................................................... 531 General Questions ............................................................................................... 531 Limits in Ascribe .................................................................................................. 533 Coding and Codebook Questions............................................................................ 533 Review Responses Questions ................................................................................ 535 Translation and Transcription Questions ................................................................. 535 Load and Download Data Questions ....................................................................... 535 Performance Questions ........................................................................................ 538 Error Messages ...................................................................................................... 539 Codebook Needs Repair .......................................................................................... 543 Security Settings .................................................................................................... 547 Excel® Add-In ....................................................................................................... 549 Install the Excel Add-In ........................................................................................ 549 Extract the File ................................................................................................. 549 Format Excel 2003 ............................................................................................ 549 Format Excel 2007 ............................................................................................ 550 Use the Excel Add-In for Question Setup in Ascribe.................................................. 552 Excel Add-In Loading Tools ................................................................................... 553 I do not have full functionality with the Excel Add-In; how do I resolve this? .............. 553 Word Configuration for Spell Check in Ascribe™ ......................................................... 554 xxi PrintDoc01 020813 Word version 10.2: Configure Spell Check .............................................................. 554 Word version 9.0 (Microsoft Word 2000): Configure Spell Check ............................... 555 Multiple Language Support in Ascribe™ ..................................................................... 556 Loading Data ...................................................................................................... 556 Downloading Data ............................................................................................... 556 Configuring Your Computer For Multiple Language Support ....................................... 556 Windows XP ..................................................................................................... 556 Windows 2000.................................................................................................. 557 Other Versions of Windows ................................................................................ 557 Using Internet Explorer in Multiple Languages ......................................................... 557 Language Selection.............................................................................................. 557 Translating Your Ascribe™ Site .............................................................................. 557 Ascribe™ Basics ..................................................................................................... 558 Coder Interface (Coder Window) ........................................................................... 558 Review Responses ............................................................................................... 559 Edit the Codebook ............................................................................................... 559 Phrase Analyzer .................................................................................................. 560 Loading Data ...................................................................................................... 560 Copy and Paste Codebooks ................................................................................... 561 Support Procedures ................................................................................................ 561 Phone Support .................................................................................................... 561 Email Support ..................................................................................................... 561 After Hours/International Support ......................................................................... 562 Language Logic Staff .............................................................................................. 563 Glossary Terms ...................................................................................................... 564 xxii About This Reference Manual This reference manual explains how to use the Ascribe™ verbatim management application. It is designed for all user roles and privilege levels. Additional training is available from Language Logic. This reference manual contains the following: Chapter Description About This Reference Manual Describes the purpose and contents of the reference manual. Introduction to Ascribe™ Provides an overview of Ascribe™ and how it works. Get Started Describes how to configure your Internet browser, how to log in, how to navigate around Ascribe™, and how to customize features. Study Describes how to create, add, edit, merge and delete studies and study questions. Codebooks Describes how to create and manage codebooks. Verbatims Describes how to code verbatims with the Coder Window, Review Responses, and Phrase Analyzer, and how to translate and transcribe verbatims. Client Access to Data Describes client reports, cost accounting, and time accounting. Ascribe Presentations™ Describes how to create cross-tabs, bar charts, line charts, pie charts, co-occurrence maps, correlation matrices, multidimensional scalings, clusters, word clouds, and concept clouds. Ascribe Inspector™ Describes how to use Ascribe Inspector™ to explore sentiment ratings, topics, and concepts. Data Management Describes how to load and download data, how to schedule and manage jobs, and how to use study shipping and the repository. Ascribe™ Desktop Describes how to use Ascribe™ Desktop. Accelerator Coding Model (Web) Describes the theories behind the Accelerator Coding Model and Verbatim Coding System and how to use ACM to train models and code verbatims. Automated Coding Model (Desktop) Describes the ACM features available thru Desktop such as the Training Example Editor and the Review ACM Coding window. Ascribe™ Crossword Describes how to use Ascribe™ Crossword. User Administration Describes how to add and edit associates and companies and how to reset passwords. 1 PrintDoc01 020813 Site Administration Describes how to configure and translate the site. Frequently Asked Questions Provides answers to a list of common questions about Ascribe™. Error Messages Describes error messages and what they mean. Security Settings Describes how to set security settings for Ascribe™. Excel® Add-In Describes how to install and use the Excel ® Add-In to set up questions and load data. Word Configuration for Spell Check Describes how to set up Word to use it for spell check in Ascribe™. Ascribe™ Basics Provides a quick reference to the Coder Window, Review Responses, Edit the Codebook, and Phrase Analyzer and describes how to load data and copy and paste codebooks. Support Procedures Describes how to get help from the Language Logic Support Team. Language Logic Staff Provides contact information for the Language Logic staff. Multiple Language Support in Ascribe™ Describes how to use Ascribe™ in multiple languages. Glossary Terms Provides definitions of commonly used terms. Who Should Use This Manual? The following table lists the user roles available and the chapters in this manual that pertain to each. User Role All Roles Administrator Supervisor 2 Applicable Chapters About This Reference Manual Introduction to Ascribe™ Ascribe™ Housekeeping Get Started Frequently Asked Questions Security Settings Ascribe™ Basics Support Procedures Language Logic Staff Client Access to Data User Administration Study Codebooks Verbatims Data Management Codebooks Site Administration Error Messages Excel® Add-In Codebooks Verbatims Client Access to Data Data Management Editor Site Administration Transcriber Verbatims Coder Client Contact Language Logic If you have technical questions about this product that are not answered in this manual, you can contact Language Logic Customer Support: support@goascribe.com Third-Party Web Sites Language Logic is not responsible for the availability of third-party web sites mentioned in this document. Language Logic does not endorse and is not responsible or liable for any content, advertising, products, or other materials that are available on or through such sites or resources. Language Logic will not be responsible or liable for any actual or alleged damage or loss caused by or in connection with the use of or reliance on any such content, goods, or services that are available on or through such sites or resources. Feedback on Documentation Language Logic continually evaluates its documentation and looks for ways to improve it to directly benefit our global customers. We welcome your comments and suggestions. We normally do not send personal replies to feedback reports. However, we do read each and every message. We use the information to improve our products and services. Your comments, suggestions, and ideas for improvements are very important to us. We appreciate you taking the time to send us this information. Please email your comments to support@goascribe.com. Introduction to Ascribe™ The Ascribe™ application offers advanced technology to improve your research productivity in several areas, including market research coding. What is Coding? Open-ended coding refers to the process by which verbatim responses are assigned to categories. These categories represent the concepts that the verbatim responses are expressing. It is a common task in the survey research industry as many quantitative studies ask for verbatim responses from the people who are interviewed. Note Quantitative studies are those that use a scientifically chosen sample to be representative of some population; for example, it may be an entire country or it 3 PrintDoc01 020813 may be just people who want to buy a car. In many cases, some open-ended questions are distributed within the larger number of closed-ended questions. Note Closed-ended questions ask for a response from within a list of categories (for example, rate this from 1-5.) The open-ended questions might be there to: Clarify some of the closed-ended questions. "Why did you rate that so poorly?" Reveal information that is difficult to anticipate in a closed-ended category. "Why did you like that advertisement?" Act as a confirmation of the closed-ended responses. Generate ideas or concepts. "What do you think the company should do to improve their service?" Open-ended questions are used extensively in advertising and copy testing, attitudes about new products or product line extensions, concept testing, and in the development of difficult or expensive surveys to help understand the right kind of closed-ended questions to ask. Open-ended questions are used to explain, clarify, and diagnose the quantitative results of a survey. In many cases, they answer the why. What is Ascribe™? Ascribe™ is a verbatim management application designed to boost the productivity and accuracy of your coders. Ascribe™ provides several coding environments, each of which is optimized for specific coding tasks. Single verbatim coding - presents a single verbatim at a time. Multiple verbatim coding - designed to code multiple responses simultaneously. Verbatims of short to moderate length/common phrase coding - designed to process verbatims efficiently with common phrases. Automatic coding of closed-ended questions - although not required, this coding can be useful to reference related questions and filter verbatims. Automatic coding of text responses – designed to code exact matches of previously-coded verbatims. Access Access to Ascribe™ is controlled by user name and password and by user roles. You must have a password to get into Ascribe™. Passwords are case sensitive, must be at least 8 characters, and expire each month. Case sensitivity considers upper-case (capital) letters and lower-case letters to be different characters. You may use any combination of letters and numbers. Access to Ascribe™’s functions is controlled by defining each associate’s user role in the Associates table. Available levels include: Administrator, Supervisor, Coder, Client, Editor, and Transcriber. 4 Codebooks Ascribe™ Architecture Ascribe™ is a web application, hosted on our secure servers. This provides several advantages to you: Global accessibility - Ascribe™ is available from any location on the Internet. No installation required - Ascribe™ requires no installation on your network. Low IT costs - Language Logic manages the Ascribe™ servers. Multi-lingual - Ascribe™ supports any character set and is Unicode compliant. Data Security Ascribe™ provides a comprehensive security framework and strategy. All data is protected by a system of interlocking technologies that ensure maximum protection. Ascribe™ is built around SQL Server databases, housed on servers at Inetu, Inc (www.inetu.net) in Pennsylvania. Inetu provides 24x7 support for server and infrastructure failures. Ascribe™ provides multiple levels of security against data loss. Active Data Stores With a single exception, all Ascribe™ data is stored in SQL Server. That single exception is the media files which represent responses that are voice or image files. SQL Server Data Store The Ascribe™ application utilizes multiple, clustered instances of Microsoft SQL Server, version 2008R2. SQL Server is configured for the Full Recovery model. Each SQL Server instance executes on a dedicated server, clustered with another server in an active-passive configuration. The passive server is an exact duplicate of the active server. In the event of failure of an active database server, failover occurs automatically to the passive server. The SQL data files and active transaction logs are stored on a storage area network (SAN), connected to the clustered database server using a redundant optical fiber network. Media File Storage Media files are stored on a separate server, on protected (RAID) storage. Backup Data Stores The Ascribe™ SQL Server writes transaction log backups every 30 minutes. Full or differential backups of the active SQL Server data stores and file systems are written to secondary storage daily. Media Files Backup Media files are backed up daily. Additional Backup Provisions In addition to the provided levels of security against data loss, other options are available. 5 PrintDoc01 020813 Repository The Repository is a separately licensed feature. From the time the Repository is licensed to you, the Repository will save current studies in your Ascribe™ database on a nightly basis to a separate server. Each study is saved in the standard XML file format used for saving and restoring studies to Ascribe™. Each night a backup copy is made of each study, if that study has been modified during the past 24 hours. Backup copies of the study are deleted after 30 days, except for the most recent backup. The most recent backup is retained indefinitely. Study Shipping Study shipping is a standard feature of Ascribe™, available to any account. When the study shipping feature is used, Ascribe™ automatically saves each study in the Archived state to a file, and sends the file via FTP to a location designated by the client. Upon successful transfer, the study is deleted from Ascribe™. This feature allows convenient long-term storage of past work. Security from Loss of Service Ascribe™ provides multiple levels of security against loss of service. Clustered Database Server The Ascribe™ database server uses Microsoft SQL Server, running in a cluster in activepassive configuration. In the event of failure of the active database server, it will fail over automatically to a passive server. The passive server is configured identically to the active server. Redundant Web Servers Ascribe™ uses multiple, identically configured web servers. In the case of failure, another server will transparently be used. The web servers store no user information, so no data loss can occur from failure of a web server. Multiple Internet Backbone Connections Inetu uses multiple network carriers to prevent loss of service resulting from a network failure at a given carrier. Security from Server Intrusion Ascribe™ provides multiple levels of security from server intrusion. Firewall The Language Logic servers at Inetu are protected by dedicated Cisco firewalls. Security from Website Intrusion The Ascribe™ application has been carefully constructed to prevent intrusion to the website. Access to the website is possible only through the logon page, where the user must provide 6 Codebooks credentials consisting of account, username, and password. The Ascribe™ website is fully logged, and checked routinely for suspicious activity. The most likely scenario for website intrusion is by access via stolen credentials. As the administrator for your Ascribe™ site, you must be certain to educate your staff on the importance of keeping this information safe. You should also make certain to disable unused accounts. Ascribe™ prompts each user monthly for a new password, although Ascribe™ does not enforce complexity or historical uniqueness of the password. You should educate your staff on the importance of proper password selection and updates. It is Language Logic policy not to provide passwords to anyone. If a password is forgotten, you as the administrator can reset the user's password. Security in Data Transit Ascribe™ provides security for your data across the web. HTTPS Protocol The Ascribe™ web sites are equipped with a dedicated SSL (Secure Socket Layer) accelerator. This allows your users to use the https protocol with negligible effect on serverside performance. To use https, your users need only place "https//:" at the front of the URL used to access Ascribe™ from their browser. If you wish, Language Logic can modify your account so that all users are required to log in using https. Contact Language Logic Support if you would like this option enabled for your account. Security of Your Data Within the limits of their privilege level, your users can download information from the Ascribe™ web site. Once this information has left the Ascribe™ servers, Language Logic can of course do nothing to prevent unintended use. Again, you should instruct your staff about security of this information. Ascribe™ provides a way to download a zipped XML document containing all information in a study (the study save operation). Language Logic can provide an encryption password for such files. This password does not need to be known by users of Ascribe™. We recommend that you use a strong encryption password, and that you keep knowledge of this password very secure. Multi-Lingual, Unicode Compliant Application Since Ascribe™ is a multi-lingual, Unicode compliant application, you can translate Ascribe™ to other languages. When you translate the site, you change the text displayed by Ascribe™ in menus, buttons, headers, explanatory text, and so on. Tip: Translating the site does not translate verbatims! What's New 7 PrintDoc01 020813 The What's New page for current and previous versions of Ascribe™ can be found when you click "What's New" under the General heading on the Home page of Ascribe™. Maintenance Schedule The maintenance schedule is determined by the support team and is different for most accounts. Users can see their own Maintenance Schedule under the General heading on the Home page. The standard schedule for maintenance operations on the Ascribe™ database server is displayed below. During maintenance operations, Ascribe™ can run slowly. You can use the information in this report to help you plan your work schedule so that you avoid the maintenance periods to the extent possible. The report has this form: Last Run Daily Sun 10/5/2008 1:00:05 AM Weekly Monthly Sun 10/5/2008 4:09:54 PM Sun 10/5/2008 2:20:35 PM Run Time (Seconds) Next Run 1,085 0:18 Mon 10/6/2008 1:00:00 AM 3,720 1:02 Sun 10/12/2008 1:00:00 AM 1 0:00 Sat 11/1/2008 2:00:00 AM The maintenance operations run daily, weekly, and monthly. For each operation, this information is displayed: Last run – This column shows the date and time the maintenance operation last started. It is adjusted to your local time zone. You can hover over the date to see the information in UTC (Greenwich Mean Time.) Run time (seconds) – This column shows the number of seconds the maintenance operation took last time it ran, followed by the same information in hours:minutes. The amount of time the operation took last time is a good estimate of the time required next time. Next run – This column shows date and time the next maintenance operation will start. It is adjusted to your local time zone. You can hover over the date to see the information in UTC (Greenwich Mean Time.) Daily Maintenance Operations During daily maintenance, you can still use Ascribe™, and you should see only minimal performance reduction. The primary function of daily maintenance is to defragment or reorganize the database indexes. The Ascribe™ database uses indexes to improve speed. As the contents of the database change, these indexes can become "fragmented," meaning that they are not organized as efficiently as possible. First, the maintenance program checks for fragmentation on each index. This check can take several minutes, but typically less than 15 minutes. The program next fixes any fragmented indexes. Defragmentation can take from no time (if there are no fragmented indexes,) to more than an hour (if a large index is fragmented.) 8 Codebooks Ascribe™ is usable still during this operation. The check for the amount of fragmentation actually causes slower performance than the defragmentation procedure does. Weekly Maintenance Operations Weekly maintenance can affect Ascribe™’s performance severely. You should try to avoid using Ascribe™ during this maintenance. During weekly maintenance, Ascribe™ performs these tasks: deferred data deletion, shrink database, rebuild indexes, and consistency checks. Deferred Data Deletion To improve performance, Ascribe™ defers the deletion of certain database records. An example is the deletion of questions from a study. If you were to delete a question with a large number of responses, it could take a long time for the database to process this request. This would slow you down, as well as any others using Ascribe™. Instead of deleting the questions at the time you make the request, Ascribe™ marks the question for later deletion. Deleting the data can take many minutes, and of course depends on just how much data was marked for deletion. The performance of Ascribe™ is very poor during this operation. Shrink Database As data are deleted from the database, the size of the data files can become very large, which can adversely affect performance. During this maintenance task, the database files are reorganized to make them smaller. Ascribe™ is usable during this maintenance operation. Rebuild Indexes The daily defragmentation of indexes helps keep performance at optimum levels, but a full index rebuild is required periodically. This operation does a better job than defragmentation, at the cost of rendering Ascribe™ unusable while it runs. Ascribe™ is not usable during this maintenance operation. You do not risk data by trying to use Ascribe™ while the operation is in progress, but many operations will fail, and most will be intolerably slow. Consistency Checks A complete check of the database is performed to ensure the health of the database. Ascribe™ is usable during this operation, with degraded performance. Monthly Maintenance Operations The monthly maintenance does not affect performance of Ascribe™. Monthly maintenance performs the operation of usage calculation. Ascribe™ collects the transactions used for each account, and inserts this information in the statement for the account. Legal Agreements Navigate: General/Agreements 9 PrintDoc01 020813 Ascribe™ is licensed software, subject to license agreements. These agreements are available on the website. When you log on for the first time you are asked to review and accept the agreements. Both the End User Agreement and Ascribe™ Services Agreement can be viewed on the Ascribe™ website under the Agreements link. The date you accepted these agreements displays on the page. Get Started This chapter provides information about tasks and navigation to help you get started to successfully use Ascribe™. Configure Your Internet Browser Before using Ascribe™, it is very important for you to configure your browser properly. If you do not do this, some features of Ascribe™ will not work. This configuration allows your browser to perform certain operations on your computer (as opposed to on the Ascribe™ web server). This allows Ascribe™ to work faster, and enables Ascribe™ to do things (like drag-and-drop operations) that would be impossible without your computer being involved. By default, your browser will not allow such operations. This is to prevent a malicious website from doing things on your computer that you do not want. Part of the configuration is to set up the Ascribe™ website as a trusted site, meaning you trust that the Ascribe™ website to do its job without harm to your computer. Version of Microsoft Internet Explorer You must be using Microsoft Internet Explorer version 6.0 or later. Other browsers are not supported. We strongly recommend that you use Internet Explorer version 7.0 or later because Microsoft is no longer supporting IE 6.0. You can get a free update for your Internet Explorer program from the Microsoft website. AscribeClient.com Component Ascribe™ also uses a Language Logic component to communicate between Ascribe™ Web and Ascribe™ Desktop. This component is required for certain Ascribe™ Web menu options which redirect the user to the Desktop version. Ascribe™ Web will function perfectly well without it for all normal coding functionality. However, in order for Ascribe™ Web and Ascribe™ Desktop to communicate, this component must be installed. You can get this update from our website under General/Downloads. Use Secure Communication We recommend that you use the secure https protocol when connecting to the Ascribe™ website. You do this by using https: in the URL you enter in your browser to connect to Ascribe™. This URL has the secure protocol: https://ascribe.languagelogic.net/Ascribe 10 Codebooks This URL does not have the secure protocol: http://ascribe.languagelogic.net/Ascribe Your account may be configured to switch automatically to the secure protocol when you log on. If it is not configured this way, we still recommend that you use the secure protocol. You can tell if you are using the secure protocol by checking for a padlock icon on the status bar of your browser. Configure Ascribe™ as a Trusted Site Log onto Ascribe™. Check the status bar at the bottom of you browser. If you see a green check mark with the words Trusted sites next to it, Ascribe™ is already a trusted site, and you can move to the section Security Setting for Trusted Sites. If you do not see the check box, follow these steps: 1. From the Tools menu of Internet Explorer, select Internet Options... 2. Select the Security tab. 3. Highlight the Trusted Sites icon by clicking it. 4. Click the Sites... button. 5. Uncheck the Require server verification (https:) for all sites in this zone check box. 6. In the Add this Website to the zone box, type: *.languagelogic.net (the first two characters are an asterisk and a period). 7. Click the Add button. 8. Check the Require server verification (https:) for all sites in this zone check box. 9. Click the OK button. 10. From the Internet Explorer toolbar, select Tools and then Internet Options. 11. Select Pop-Up Blocker and then Pop-Up Blocker Settings. 12. Add *.languagelogic.net in the Address of Website to Allow box and Click Add. 13. Click Close to close out of Pop-Up Blocker Settings. Security Setting for Trusted Sites To configure your security settings for trusted sites so that Ascribe™ will work properly, follow these steps: 1. If the Internet Options dialog box is not displayed, select Internet Options from the Tools menu of Internet Explorer. 2. Click the Security tab. 3. On the Security tab, click the Custom level... button. 4. Locate the heading Active X controls and plug-ins. 5. Under this heading, there are several sub-headings. Set the Enable option for each of them. Make sure that you have set Enable for all of the Active X controls and plug-ins. If you miss one, you may still have problems using Ascribe™. 6. Click OK in the Security Settings dialog box. 11 PrintDoc01 020813 7. Click OK In the Internet Options dialog box. Additional Configuration for IE 7 If you are using Internet Explorer 7, it is recommended that you disable the use of "Tabbed Browsing." Use of tabs in Internet Explorer may confuse Ascribe™ because cookies are being managed in two tabs. 1. If the Internet Options dialog box is not displayed, select Internet Options from the Tools menu of Internet Explorer. 2. Click the General tab. 3. In the Tabs section, click the Settings button. 4. Make sure the Enable Tabbed Browsing check box is unchecked. If there is a check mark there, remove it. 5. Click OK. Set Your Language Preference The Ascribe™ website can be translated into multiple languages. To translate the Ascribe™ website, see Translate the Ascribe™ Interface. If your Ascribe™ site has been translated to a given language, you can choose to view Ascribe™ in that language. To view Ascribe™ in a translated language, perform the following steps: 1. Click Tools and then click Internet Options in your browser menu. 2. On the General tab, click the Languages... button. a. If the language you desire is not in the list displayed, click the Add button and add the desired language. b. Highlight the language in the list, and click the Move Up button until it is at the top of the list. c. Click OK to close all of the dialog boxes. 3. Log off of Ascribe™ and log back on for the language change to take effect. You can verify the current language setting on the Ascribe™ Home page. It is indicated by the blue abbreviation under the Ascribe™ logo. (For example, en-us represents United States English.) Low Resolution Monitors If your computer monitor is small or low resolution, it can be hard to see everything easily in this window. Here are some things you can do to help: 12 Codebooks Show full screen - You can tell Internet Explorer to use the full screen. From the View menu, select Full Screen. Check your display settings - You may be able to use a higher resolution setting for your monitor. From the Windows desktop, select Properties/Settings. Experiment with the Screen Resolution setting to see if you can use a higher resolution setting. Change your text size - Internet Explorer lets you adjust the text size in the window. From the View menu, select Text Size. Try a smaller size to see more in the window. IE 8 Accelerator Icon IE 8 introduced the use of an accelerator icon - a blue arrow that appears as you navigate around the screen. If you do not want to see this arrow, you can disable it by following these steps. 1. Open Internet Explorer. 2. Select Tools/Internet Options/Advanced. 3. Uncheck the Display Accelerator button on selection under the Browsing section. 4. Click Apply and then OK. Configure Word for Spell-Check in Ascribe™ Take advantage of the spelling-related functionality that Microsoft Word offers. Below are instructions for Word Version 10.2 and Word Version 9.0 (Microsoft Word 2000). If your version is not listed, use the appropriate process in Word to: Check spelling as you type. Add words to the custom dictionary. Add abbreviations or shortcuts to be autocorrected. Tip When you transcribe in Ascribe™, make sure you have the Check Spelling box checked at the top of the screen. Word Version 10.2 Perform the following steps: 1. First make sure that you have Word set up to check spelling as you type: a. Launch Word. b. Click the Tools menu. c. Click Options. d. Click the Spelling & Grammar tab. e. Make sure the first option is selected (Check spelling as you type). f. Click OK. 2. Add any desired words to your custom dictionary: 13 PrintDoc01 020813 a. Click the Tools menu. b. Click Options. c. Click the Spelling & Grammar tab. d. Click the Custom Dictionaries button. e. Click the dictionary in the left pane that you would like to add to (in most cases there will be just one). f. Click the Modify button on the right. g. Type the word you would like to add in the text box. h. Click the Add button. i. Click OK until you have exited out of all of the popup boxes. 2. Add any abbreviations or shortcuts you would like to auto-correct as you type: a. Click the Tools menu. b. Click AutoCorrect Options… c. Click the AutoCorrect tab. d. Type the abbreviation or shortcut you would like to use in the Replace text box. e. Type what the abbreviation means or the long version of your shortcut in the With text box. f. Click the Add button. g. Click OK until you have exited out of all of the popup boxes. Word Version 9.0 (Word 2000) Perform the following steps: 1. To add any abbreviations or shortcuts you would like to auto-correct as you type: a. Click the Tools menu. b. Click AutoCorrect Options… c. Click the AutoCorrect tab. d. Type the abbreviation or shortcut you would like to use in the Replace text box. e. Type what the abbreviation means or the long version of your shortcut in the With text box. f. Click the Add button. g. Click OK until you have exited out of all of the popup boxes. 14 Codebooks 2. To add any desired word(s) to your custom dictionary (so that Word and Ascribe™ will not view the word as a misspelling): a. Click the Tools menu. b. Click Options. c. Click the Spelling & Grammar tab. d. Click the Dictionaries button in the middle of the pane. e. Click the Edit button from the next popup box. A new Word document will open. f. Type the word or words you would like to add and then save and close the document. g. You will have to reset the Check Spelling as You Type function. It is de-selected every time you update your custom dictionary. h. Click the Tools menu. i. Click Options. j. Click the Spelling & Grammar tab. k. Select the Check spelling as you type option. l. Click OK until you have exited out of all of the popup boxes. FTP Setup Your FTP site should be configured according to your own best practices. See Auto FTP Setup for suggestions intended to help prevent malicious use of your FTP site. Log In To sign on to Ascribe™, enter the URL below in your browser address box: https://ascribe.languagelogic.net/ascribe Note Remember to use the s (https) in the address prefix so your traffic will be encrypted. Ascribe™ uses 128 bit, public key encryption from a subsidiary of Verisign called Thawte. The log-in screen displays. The account name is provided by Language Logic Support. Your systems administrator will assign you a user name (usually firstname.lastname). 15 PrintDoc01 020813 The first time you sign on, your password will be password. The system will ask you to select a new password. If you forget your password or user name, your administrator can reset your password to password. Passwords must be at least 8 characters and are case sensitive (upper/lower case matters). After you change and confirm your new password, you will be signed off and will have to sign on again. You will be required to change your password every month. When your password expires, you will be asked to enter a new one, confirm it, and sign back on. Home Page and Main Menu From the Ascribe™ Home page, you can navigate directly to the various sections of the site by clicking a link in the navigation bar at the left of the window. The features available to you depend on the privilege level of your login account. Your administrator sets the privilege level for each user. See Navigation for more information about how to move to various locations in Ascribe™. Navigation (Popup Menus) There are two navigation popup features: popup menus and popup menus in tables. To display a popup menu, click the right mouse button. When the popup menu appears, highlight the option you want and left-click the option. An arrow to the right of the option indicates you have additional choices on another popup menu. The contents of the navigation menu depend on your privilege level. Your administrator can change the privilege level for any account. You should not use the Back button on your browser to navigate around the Ascribe™ site. Because the content of the site is dynamic (it can change each time it is displayed), the Back button can show you incorrect results by displaying a page image saved by your browser. In many cases, you can select a row in a table (by clicking it) and then perform an action on the row. To perform the action, right-click the row. Select the desired action from the popup menu displayed. Application Help 16 Codebooks The help system is a feature that allows the user to obtain information about usage of the Ascribe™ software and updates. Ascribe™ Help provides the user with definitions about the different functions in Ascribe™. The help system is context sensitive. Users can right-click and choose Help to get information about the current page/function. The help system also can be accessed from various pop-up boxes throughout the software. When Help is selected, a new browser window opens with information about the current area of Ascribe™. Note Application Help opens a new browser window. It can be minimized or closed without interrupting your Ascribe™ session. To print a single page of the documentation, click the Print button in the Help screen toolbar. To more than a page, use the Help System PDF. Customization Features Set User Options On the Ascribe™ Home page, select User Options on the General menu. This option allows users to set default behaviors for selected aspects of Ascribe™ operation. Option Value Ascribe™ one-click page This option determines where Ascribe™ will be redirected when a user clicks the Ascribe™ logo in the upper right hand corner of most windows. Depending on user privilege and account licenses, some or all of the following will be available in the pull down list in this section. Home Page - Returns user to the Ascribe™ Home page Studies - Returns user to the Supervisor Studies page Last Study Questions - Returns user to the last Study Questions page visited Coding Studies - Returns user to the Coding Studies selection page Last Coding Questions - Returns user to the last Coding Questions page visited Transcribe Studies - Returns user to the Transcribing Studies selection page Last Transcribe Questions - Returns user to the last Transcribing Questions page visited Translate Studies - Returns user to the Translating Studies selection page Last Translate Questions - Returns user to the last Translating Questions page visited Client Studies - - Returns user to the Client Studies page visited Note: Users are always directed to the Home page after initial login. The selection of the re-direct options here are in effect only after the initial Home page is displayed. 17 PrintDoc01 020813 Disable spell check If checked, this option will disable the use of the spell checker in the for codebook editing codebook editor. Normally this will be selected if 1) MS Word does not exist on the user machine or 2) spell check is not desired in order to improve performance while in the codebook creation stage. Keep filters between This option, if checked, will keep response filters between sessions sessions of Ascribe™. Double-click to show If this option is checked, most of the drop-downs displayed (such as drop-downs on the Studies page) will open with a double-click instead of a single-click. It does not affect the drop-downs on the ACM Training Example Editor page. Choose Columns The Choose Columns option allows you to select which columns display. You can display as much or as little information as necessary. For instance, a coder might not want to see the date a study was created, but may be interested in seeing the start date and due date. To choose a column to display, click the box next to the column name. The boxes without check marks will not display. Click the OK box to save your changes and exit or click Cancel to exit without saving changes. After you change the column settings, the settings stick until you change the settings again. Some columns are not optional. For example, the study ID on the Studies page and the question ID on the Questions page must always display. Therefore, you will not see check boxes for those fields when customizing the Studies page or Questions page. Some boxes control the display of a group of columns. For example, the ACM check box on the Choose Columns dialog box for the Questions page hides or displays the Code Using, Approved, and Coding Job columns. Use the following links to view which columns are available on each page: Studies Page Questions Page Studies in Progress Page Questions in Progress Page Copy/Share Manager Account Codebooks Page Automated Coding Models Page Accelerator Coding Model Page Training Results Training Example Editor Inspections Page Shipped Studies Page Production Bands Report Associates Change Column Widths 18 Codebooks Navigate: Supervisor/Studies/Right-click a study Client/Studies/Right-click a study Supervisor/Studies/Questions/Right-click a question You can change the column widths for the columns on this page. Right-click a study, and select Change Column Widths. A dialog box displays with the defaults for each column. The widths are shown in pixels. Increase the number to make the column wider, and decrease the number to make the column more narrow. After you change the widths, click OK to save your changes. Click cancel if you do not wish to make any changes. Resize Windows You may drag the window bars to whatever height and width you want in any of the active windows; however, this resizing does not stick. When you leave the page and return, the default setting is in effect. Language Support The Language Logic system interrogates Internet Explorer and displays the site in your preferred language. You may see your preferred language by accessing Tools/Internet Options/Languages on Internet Explorer’s Tools menu or by navigating to the Home page and paying attention to the blue text under the Ascribe™ icon in the top right-hand corner of the page. Open Sessions The Open sessions report tells you who is logged on to Ascribe™. This page is visible only to associates. Users who log on as a contact cannot view this page. Downloads Navigate: General/Downloads The Downloads page contains links to files you may need to use Ascribe™. These files include components to install or documents. To download a file, click on the link. Log Off To log off Ascribe™ Web, use the Logoff option in the right-click menu in whatever window you are in. You should always log off from Ascribe™. If you do not log off, you may leave responses 'locked' so that others cannot code them. For example, if you simply close your browser in the Coder Window, the response you were coding will remained locked, awaiting your codes. This action prevents others from coding this response. (To unlock a response, see Locked Responses.) When you log off, Ascribe™ computes the amount of time you were logged on. When computing your logon time, Ascribe™ checks whether you had more than one open browser session. If you did, it calculates your logon time for the non-overlapping periods only. You can close Ascribe™ Desktop by any of these options: Select the Exit option from the Main Menu/File option Select the Log Off option from the Main Menu/File option 19 PrintDoc01 020813 Click the red "X" in the upper right corner of the Application Main window. Study Studies Page Navigate: Supervisor/Studies Client/Studies This page shows a filtered list of all studies. Filtering is based on the settings in the filter bar at the top of the page. To change the filter, check the desired boxes in the filter bar, and click the Update button. (When you want to refresh the Studies page, use the Update button rather than the Internet Explorer refresh button. If you use the IE refresh button, you may receive an error message from IE.) When you right-click a study, you have options to navigate to other parts of Ascribe™ or edit the study. See Studies Page Right-Click Menu for more information. Filter Options The text filter in the bar allows you to filter studies by study ID or study name. To show only studies whose ID or name starts with certain characters, enter those characters in the Filter box and click the Update button. For example, if you enter " abc" in the Filter box, only those studies whose ID or Name start with abc will be shown in the list. To find studies whose ID or name contains certain characters, put an asterisk as the first character of the filter. For example, if you enter "* abc" in the Filter box, only those studies whose ID or name contains abc will be shown in the study list. The Supervised by Me check box means "only display studies that are assigned to me as the supervisor." The Assigned to Me check box means "only display studies that are assigned to me as a coder (or a transcriber/translator.)" See Assign a Coder to a Study for more information. Columns on the Studies Page The Studies page contains these columns (you can choose the columns displayed by selecting Choose Columns from the right-click menu): Field Created Description The date the study was created on the web server. You can not modify this value. Start Date The date work is scheduled to start on the study. The start date can be set from the drop-down box on this field or from the Date Tab in the Study Edit dialog. The drop-down box contains a text box and a calendar. You can enter a date in MM/DD/YYYY format or select a date from the calendar. Press OK to save the date or press Close or Clear to exit without saving. Due Date The date this study is expected to be completed. The due date can be set from the drop-down box on this field or from the Date Tab in the Study Edit dialog. The drop-down box contains a text box and a calendar. You can enter a date in MM/DD/YYYY format or select a date from the calendar. Press OK to save the date or 20 Codebooks press Close or Clear to exit without saving. Study ID The unique identifier for the study. If you left-click this field, a drop-down box displays study information. You can navigate to the Question page if you click the Questions button on drop-down box toolbar. Study Name The name of the study. Quota The quota for the study. The field is just for information and is not used in any calculations or statistics. Edit this field on the General Tab of the Edit Studies dialog box. Questions The number of questions in the study. The drop-down box for this field displays information about the questions. Responses The total number of responses in the study. Responses Coded The number of responses coded in the study. Description A description of the study. Help On This Study Help for this study, intended to assist coders with information about the study. Status The study status. If you left-click this field, a drop-down box displays a list of status options. To change the status, left-click a status option. The drop-down box automatically closes when you click a status option. If you don't want to make any changes, click the Close button. Client The company assigned as the client for this study. End Customer The company assigned as the end customer for this study. Supervisor The associate designated as the supervisor of this study. The drop-down box displays supervisor details. Assigned To This column displays users assigned to this study. This column has a drop-down box. Click the Edit link to display the Asssigned To dialog box where you can assign coders to the study. For more information, see Assign a Coder to a Study. ACM Coding Questions This column displays the number of ACM coding questions. The drop-down box displays information about the coding questions, such as question ID and codebook ID. Layout Section Columns Per Card/Respondent The number of columns in each data row output. Card Number Column The column in which the card number starts. Card Number Columns The number of columns occupied by the card number. Respondent ID The column in which the respondent ID starts. 21 PrintDoc01 020813 Column Respondent ID Columns The number of columns occupied by the respondent ID. Note Drop-Down Boxes Several columns on this screen display drop-down boxes when left-clicked. A hand icon displays when you hover over these fields. The drop-down boxes can open with a single click or a double click. Choose the method you prefer in User Options. To close the drop-down box, click Close, left-click twice in the box, or click once anywhere on the other columns. Studies Page Right-Click Menu Navigate: Supervisor/Studies/Right-click a study Client/Studies/Right-click a study The right-click menu on the Studies page has the following options: Option Description Questions Navigates to the Questions page for this study. Copy/Share Manager Navigates to the Copy/Share Manager, where you can copy and share codebooks for the questions in this study. Ascribe Presentations Navigates to Ascribe Presentations, which offers tools for presenting Ascribe™ data in various graphical and tabular formats. Reports Navigates to one of the following reports: Respondents Loads Transactions by User - administrator only Coding Overview Study Summary Question Summary Card Layout Study Report New Crossword Study Quality Report Time by Study - administrator only, with these reports: Time by Question, Time by Activity, Time by Associate, Time by Associate, Question, and Activity, Time Details. Download Data Navigates to either Codebooks and Results or Scripted Output. Edit… Opens the Study Edit dialog box, where you can change the General Tab, Data Tab, or Date Tab information. Delete… Opens a dialog box where you can confirm you want to delete the study. 22 Codebooks Add a New Study… Opens the Study Edit dialog box where you can enter study information on the General Tab, Data Tab, or Date Tab. Set Quality Codes Navigates to the Set Quality Codes page. View Codes by Coder Navigates to the Codes by Coder Report. Load Responses / Process File Navigates to the Load Responses / Process File page where you can load data. File Navigates to one of the following options: Save Study page Restore Study page Merge Study page. Copy Study Navigates to the Copy Study page, where you can create a copy of a study, without the need to save and restore the study. Choose Columns Navigates to the Choose Columns dialog box, which allows you to choose the columns to be displayed on this page. Change Column Widths Navigates to the Change Column Widths dialog box, which allows you to choose the column widths for the columns on this page. Assign a Coder to a Study Navigate: Supervisor/Studies A question may be assigned to one or more users, signifying that these people are assigned to work on that question. These assignments serve two main purposes. First, they allow a coder to quickly determine the work to which he or she is assigned. Second, they allow you to distinguish time spent by a coder on assigned work versus unassigned work. This information is provided in the Time by Study reports. You can assign coders to a study in three different ways. The Assigned To column on the Studies page is used for study-level assignments for users with coder privilege. When you assign a coder, you also select the question type (open, closed, other specify, etc.) The coder is assigned to all questions with that question type. The Assigned To column on the Questions page is used to assign both users with coder privilege and users with only transcriber privilege. You can also assign coders and transcribers on the Coding Tab of the Questions Properties dialog box. With these two methods, you assign users to individual questions. Users can quickly see the studies they are assigned to by selecting the Assigned to me check boxes at the top of the Studies page, the Studies in Progress page, or the Questions in Progress page. Note When a study is saved or archived from your site, the Assigned To information is not retained. When a study is restored, the original Assigned To information is not available and will not display. Assigned To Dialog Box from the Studies Page 23 PrintDoc01 020813 First, make sure the Assigned To column is displayed on the Studies page. If it is not displayed, right-click a study and select Choose Columns. Click the Assigned To check box in the Choose Columns dialog. Next, click the Assigned To column for the desired study, and select the Edit link in the drop-down box. A dialog box displays. The Assigned To dialog box has this format: Section Coder List Description This list includes only users who have coder privilege. To select one coder, left-click the name. Use control-click or shift-click to select multiple users. The highlight indicates that you have selected the user or users. (If you need to assign users who have only transcriber privilege, use the Assigned To column on the Questions page.) This box is initially empty. Use the ">" button to move the selected user or users from the Coder List to the Assigned To box. (The user or users that are highlighted are the ones that will move.) Assigned To To remove a user or users from the Assigned To box, select the user or users by a left-click, control-click, or shift-click. Use the "<" to move a user or users from the Assigned To box to the Coder List. (The user or users that are highlighted are the ones that will move.) After you save the assignment and exit the dialog box, the Assigned To column on the Studies page displays the users who are assigned to this study. If you return to the Assigned To dialog box, the Assigned To box will display the coders assigned to the study. Question Type To select a question type, click the box next to it. You can select multiple question types. All of the users in the Assigned To box will assigned to whatever question types you select. Change Questions Already Assigned Select this box if you need to make a change to an assignment to a question type that has already been assigned and saved. (You do not select this box the first time you are making an assignment to a question type or types.) See How to Make Changes for more information. OK / Cancel Click the OK button after you assign a user or users to a question type. This action saves the assignment and closes the dialog box. Click the Cancel button if you do not want to save the assignment. More Notes on Assigning Users by Question Types If you need to assign a user or users to one question type and a different user or users to another question type, you must perform these actions separately. For example, User A should be assigned to the open questions, and User B should be assigned to the other specify questions. Here are the steps: 24 Assign User A to the open questions and then click the OK button to save and exit the dialog box. Codebooks Open the Assigned To dialog box again (click the Assigned To column and select Edit from the drop-down box.) User A will display in the Assigned To box. Move User A back to the Coder List. Move User B to the Assigned To box. Click the box next to Other Specify and click the OK button to save and exit the dialog box. (You do not have to click the box next to Change questions already assigned in this situation. Only the open questions had been assigned, and we did not want to change the coder assigned to those questions.) The Assigned To column on the Studies page will display both User A and User B. If you need to assign coders to individual questions, rather than by question type, use the Assigned To column on the Questions page or the Coding Tab of the Question Properties dialog. How to Make Changes Once you have assigned users to question types and clicked the OK button to save the assignment, you can make changes. Return to the Assigned To dialog box. The Assigned To box displays all coders assigned to the study. You have the option to add coders to and/or delete coders from question types. To add a coder, select the user from the Coder List and move that name to the Assigned To box. If there are any users who should not be assigned to the question type or types, move those names from the Assigned To box to the Coder List. Click the boxes next to the desired question types, and click the box next to Change questions already assigned. Only the names in the Assigned To box will be assigned to the question type or types. Click the OK button to save and exit the dialog. If you need to make more complex changes, you may want to use the Assigned To column on the Questions page or the Coding Tab of the Question Properties dialog. Create a Study You have several options to create studies in Ascribe™. You can create a completely new study with the Add Study function. You can create a study from an another study with the Restore Study function. Or, you can import files such as . qdi, . xml, . hrd, . vbs, or other input files as determined by an input script. These files contain information, usually from a questionnaire script, that defines a study’s questions, question text, output locations, etc. Add a Study Navigate: Supervisor/Studies/Right-click in white space/Select Add a New Study You must have supervisor or above access to create a study. From the Studies page, select Add a New Study. The Study Edit Screen displays. Enter information on the General Tab, the Data Tab, and the Date Tab. General Tab (Study) Navigate: Supervisor/Studies/Right-click study/Select Edit The General Tab contains general information about the study. First, enter the study ID and the study name. These fields identify the study and display in a number of lists. The other fields on the General Tab are optional and can be updated later. 25 PrintDoc01 020813 If you load data via auto-FTP, make sure the study name matches the data file name (see Auto FTP Setup). Study Status Study status refers to the point of completion of the study. Each coding department uses study status according to their internal processes. Some use all, and some use only Coding in Progress and Archived. Status Description Under Construction The study is not ready to code. Use this status while you prepare a study. For example, use this status if the data or the codebooks are not ready. The study will not appear on the Studies in Progress page or Studies to Transcribe page. (Coders and transcribers will not be able to access the study). Studies restored as "New" have a default status of "Under Construction." In Progress The study is ready to code or the coding has started. The study can be accessed from the Studies in Progress page or the Studies to Transcribe page (or any page). On Hold The study cannot be accessed from the Studies in Progress page, the Studies to Transcribe page or the Studies to Translate page. There may be a problem or a change to the study, and all work should be stopped until the problems are resolved. Review in Progress Use this status while the user does a quality check of the study. For example, use when the codes are checked for accuracy or low mentions. The study can not be accessed from the Studies in Progress page, the Studies to Transcribe page, or the Studies to Translate page Also, you can use this status when you create or deliver study output. Complete The study is complete; data is ready to be downloaded for analysis. Archived The study is removed from all lists of active studies. If study shipping is enabled, all archived studies are sent to your FTP site on the Saturday after the status is changed to archived. At that 26 Codebooks time, the study is deleted from Ascribe™. You can see a list of studies that have been shipped to your FTP site under Supervisor/Shipped Studies. Other fields on the General Tab: Field Description Quota The quota for the study. Study ID The unique identifier for the study. Study Name Name of the study. Description Text to explain or identify the study. Supervisor Select the supervisor assigned to the study. Client Select the company who requested the study. End Customer Select the company who requested the study. Help on This Study Enter information about the study. You can view this information from the Coder Window when you right-click in the codebook pane and select Question info. Data Tab (Output Data Formats) Navigate: Supervisor/Studies/Right-click study/Select Edit/Select Data Tab Data formats are defined at the study level and govern the way that data is written out for analysis. This information can be changed or updated at anytime during the coding process. Note: Administrators have access to set the defaults for any new study added to Ascribe™. See Site Configuration. The Data tab controls the format of the Column Binary Output Script. If you do not use this type of output, you normally do not need to change the settings on this tab (unless you make use of these settings in a user defined scripted output.) Here are the fields on the Data tab: 27 PrintDoc01 020813 Field Description Options are: File Format Binary: Write a binary file in 1130 binary format ASCII: Write an ASCII (text) file. If you use column binary output, you will almost always want ASCII format. See Layout Types for more information. Options are: Layout Punch: Write column binary data using punch format Numeric: Write column binary data using numeric format Punch Using Column Offset: Compute the column for the punch by adding the code column to the question column Use Question Settings: Use the layout setting specified in each question. This setting allows you to mix data formats among different questions. Columns Per The value specifies the number of columns (in characters) in each Card/Respondent data record written to the output file. In column binary format, these records are called cards. Card Number Column The value specifies the location of the card number on the card. The value 1 means the first column on the card. Card Number Columns The value specifies the number of columns used to write the card number on the card. The card number is left filled with 0's when written. Respondent ID Column The value specifies the location of the respondent ID on the card. The value 1 means the first column on the card. Respondent ID Columns The value specifies the number of columns used to write the respondent ID on the card. The respondent ID is left filled with 0's when written. Date Tab Navigate: Supervisor/Studies/Right-click study/Select Edit/Select Date Tab The Date tab allows you to change the start date and the due date properties of the study. Here are the fields on the Date tab: Field Description Start Date There are two ways to set dates. You can enter a date in MM/DD/YYYY format or use the calendar. (You can also set the Start Date from the Studies page.) To set the start date, click the Start Date check box. You can enter the date in MM/DD/YYYY format in the text box. Or, you can click the text box to display the calendar, and click the appropriate day on the calendar. After you select a day, the text box displays the date. To change the start date, click the text box and enter a new date in MM/DD/YYYY format. Or, you can click the text box, and click a date on the calendar. To display another the month on the calendar, click the backward or forward arrows next to the calendar header. Or click the 28 Codebooks calendar header to display a list of months. Click the header a second time to display a list of years. If the Start Date box is not checked, the start date will be blank in the Studies list. The start date can be changed at any time during the process of the study. Click OK after you set the start date and/or the due date. Due Date There are two ways to set dates. You can enter a date in MM/DD/YYYY format or use the calendar. (You can also set the Due Date from the Studies Page.) To set a due date, click the Due Date check box. You can enter the date in MM/DD/YYYY format in the text box. Or, you can click the text box to display the calendar, and click the appropriate day on the calendar. After you select a day, the text box displays the date. To change the due date, click the text box and enter a new date in MM/DD/YYYY format. Or, you can click the text box, and click a date on the calendar. To display another the month on the calendar, click the backward or forward arrows next to the calendar header. Or click the calendar header to display a list of months. Click the header a second time to display a list of years. If the Due Date box is not checked, the due date will be blank in the Studies list. The due date can be changed at any time during the process of the study. Click OK after you set the start date and/or the due date. Copy Study Navigate: Supervisor/Studies/Right-click a study/Select Copy Study You can make a copy of a study in a single operation, without the need to save and restore the study. On the Studies page, right-click the study you want to copy, and select Copy study... from the menu. Enter a value for the ID of the new study and click the OK button. If you check 29 PrintDoc01 020813 the Share codebooks with original study box, the codebooks in the new study will be shared with the original study. Otherwise copies of the codebooks in the original study will be made for the new study. The sharing relationships of codebooks within the study are maintained, regardless of the option you choose in the Share codebooks with original study box. The copy operation applies only to the meta-data for a study. Responses are not copied. The copy operation maintains all of the information in the original study, including the coders assigned to questions and related questions. The status of the new study will be set to Under Construction, so you must select to view studies with this status to see the newly created study. Save a Study The save and restore features under Supervisor/Studies enable you to save and restore entire studies or their components to your hard drive or network. This feature allows users to keep a "master" study or study templates in a central location and restore or merge them at anytime to have a similar study. Users generally save and restore the last version of a tracker or wave study to reduce the amount of study preparation time. When you save studies in another location, it also means you do not have to keep many studies in the Ascribe™ database. You can save them to your network and then restore them when needed. You also can take advantage of the study shipping option which automatically saves studies marked as archived to your FTP site. Save Study Options Navigate: Supervisor/Studies/Right-click study/Right-click/Select Save After you select to save a study, this screen displays: Select one of these options: Option Questions, Codebooks, Responses, Codes Applied, Media Files 30 Description This option saves all information in the study, and in addition saves any media files for the study. If there are no media files associated with the study, this option is the same as the default Questions, Codebooks, Responses, Codes applied option. The advantage of this option is that you can restore the study at a later time, along with the Codebooks media responses. The disadvantage is that the created file can be very large. If you keep the original media files yourself, you may want to save the study without the media files. The save and restore process will be faster, but it will not restore the media responses. You can restore those responses from your original load files, but this does require more work. Contact Language Logic Support for more information. Questions, Codebooks, Responses, Codes Applied This option saves all information in the study, but does not save media files. This option is the default, and is the one you normally want to use. Questions, Codebooks This option does not save the responses or any codes applied to responses. If you restore a study that is saved with this option, no responses will be restored. This option can be used to store a 'template' of a study, and can be appropriate for trackers. This option creates a much smaller file. The Job Submitted page displays after you select to save the study. Job Submitted Page The Job Submitted page displays after you select to save the study. You can remain on this page to wait for the completion of the job, or you can continue to do other work in Ascribe™. Your job continues to process if you leave this page or even if you log off from Ascribe™. To access the job later, navigate to General/Jobs. When the job is finished, right-click the link and select Save Target As. A popup box displays. Choose where to save your file (Save in), what to name your file (File name) and what file type to use (Save as type). 31 PrintDoc01 020813 Restore a Study The save and restore features under Supervisor/Studies enable you to save and restore entire studies or their components to your hard drive or network. This feature allows users to keep a "master" study or study templates in a central location and restore or merge them at anytime to have a similar study. Users generally save and restore the last version of a tracker or wave study to decrease the amount of time spent to set up the studies. To restore a study, you first must save the study. The restore operation returns the study to Ascribe™. Restore Types There are several types of study restore operations. Restore type Description Original This option restores the study completely. The intention of this restore type is to allow you to save a study, delete it from Ascribe™, and later restore it exactly as it was when you saved it. Because this restore type is intended to restore a deleted study to its original condition, it causes an error if the study is still present in Ascribe™. The study must be deleted from Ascribe™ before you can use this restore type. Setup 32 This restore type is available only when you send a saved study file to Ascribe™ via the loader FTP site. It creates a new study in Ascribe™ and loads all of the information from Codebooks the saved study. The restore operation will fail if there is a study in Ascribe™ with the same study ID as the one being restored. Questions, Codebooks, Responses, Codes Applied This restore type differs from the Original restore type only in the way Ascribe™ checks whether the study already exists. With the Original restore type, Ascribe™ checks an internal identifier for the study (the StudyGuid). If a study exists in Ascribe™ with the same StudyGuid as the study being "restored as Original," the restore operation fails. With the Setup restore type, Ascribe™ checks the study ID (the StudyIDClient tag in the XML document). If a study exists in Ascribe™ with the same study ID as the study being restored as Setup, the restore operation fails. This restore type, and the next ones listed in this table, create a new study in Ascribe™. With these restore types, you must provide a study ID and a study name. The study ID you provide may be changed by Ascribe™ if a study already exists with this study ID. This restore type loads all of the information from the saved study, with these exceptions: The study ID and study name are set to the values you provide; The start date of the study is set to the current date; The status of the study is set to under construction. Questions, Codebooks, Responses This restore type is identical to the Questions, Codebooks, Responses, Codes applied restore type, except that the codes applied are not loaded. All responses in the restored study have the coding removed. Questions, Codebooks This restore type is identical to the Questions, Codebooks, Responses restore type, except that the responses are not loaded. This restore type is useful when you want to use a saved study as a template for a new study. Questions This restore type is identical to the Questions, Codebooks restore type, except that codebooks are not restored. Each question is restored with an empty codebook. Restore Shared Codebooks 33 PrintDoc01 020813 When you restore a study, shared codebooks may or may not remain shared. Codebooks that are shared within a study remain shared. Codebooks that were shared across studies do not remain shared with those studies. For example, Q1 of Study A shares a codebook with Q1 of Study B. If we save Study A and then restore it as Study C, we find that Q1 of Study C no longer shares a codebook with Study B. The codebook is restored, but the shared connection is broken. It is possible to restore shared connection of codebooks across studies using a merge instead of a restore. How to Restore a Study Navigate: Supervisor/Studies/Click Restore Study After you select to restore a study, this screen displays: Enter the information below to restore a study. Field Description Study File Enter the name of the file to restore, or click Browse... to locate the file. You can use either the original zip file that was saved, or the XML file that you extracted from the zip file. Use of the original zip file is recommended because it is much smaller than the XML file and takes less time to transfer. New Study ID This entry is available only if you have not selected the Restore as original option. Type the study ID for the new study. New Study Name This entry is available only if you have not selected the Restore as original option. Type the study name for the new study. Restore as Original This option uses the study ID and study name from the saved study. Use this option if you want to restore a deleted study with the same study ID and name as the saved study. The restore operation will fail if the original 34 Codebooks study has not been deleted. Prior to 01/29/2010, Ascribe™ did not restore the name of the coder who applied a code to a response. After that date, the coder name is restored. The load information is not restored in either case. Account codebook designations are not restored. Codebook IDs are restored if the codebook ID is unique to the site. Restore as New Study Questions, Codebooks, Responses, Codes Applied This option restores all information in the saved study, creating a new study. Questions, Codebooks, Responses This option creates a new study, but does not include any codes applied from the saved study. Questions, Codebooks This option creates a new study, but does not restore the responses and codes applied from the saved study. After you enter the information, click OK to submit the job. When it finishes, the Jobs page displays. Merge Study Navigate: Supervisor/Studies/Right-click study/Click Merge Study When you merge a study, you take information from a saved study file and add it to an existing study. The merge does not create a new study. An item from the saved study file is added to the existing study if it does not already exist. Reasons to merge a study: Maintain codebooks across waves or trackers Automatically code text responses to help with the coding process Save time when you set up waves and trackers. The table below lists the components of an Ascribe™ study and how they are affected by a merge operation. Component Description Study The study setup information (the information entered in the study edit dialog box) is never modified by a merge. Codebook and Codes If the study has no questions, the questions and the codebooks from the merged study are added. The codes and the nets are added. Ascribe™ creates a share between the studies. If the study has the same question IDs as the merged study and has no codebooks, the codes are added without nets. Also, Ascribe™ does not create a share between the studies. If the study has the same question IDs as the merged study and has codebooks, the codes may be added. Nets are not added. Ascribe™ does not create a share between the studies. If a code 35 PrintDoc01 020813 is added, it is added to the end of the codebook. A code will be added if: o it is not a net o its outputID is not blank o there is no existing code (or net) with this outputID in the existing codebook. Question The question is added if there is no question in the existing study with the question ID of the merged question. Response The response is added if there is no existing response with the respondent ID of the merged response. Codes Applied to a Response Codes applied to responses are never added in a merge operation. Quality Codes Applied to a Response Quality codes applied to a response are never added in a merge operation. Question Relationship Question relationships are never added in a merge operation. See Related Questions. How to Merge a Study Navigate: Studies/Right-click a study/Select Merge Study When you merge a study, you have a master study and a target study. The master study has the information you want to add to the target study. Both the master study and the target study must exist on the Ascribe™ site. If the master study is not on the site, use the restore option to load it to the site. The target study can be a new study or an existing study. Here are the steps to merge a study: 1. Save the master study to your computer. 2. Right-click the target study and choose Merge Study. The Merge Study page displays. 3. Click the browse box to locate the master study you saved in step 1. 4. Choose Questions, Codebooks, Responses or Questions, Codebooks. 5. Click Ok. The job is submitted. When it is finished, the Jobs page displays to let you know the status. 36 Codebooks Restore and Merge via the FTP Site The merge and restore facilities allow you to conveniently develop automated systems for moving studies and response data into Ascribe™ from the FTP site. You can perform a complete setup of Ascribe™ studies automatically, and load data incrementally into existing studies. Data File Format The same file format is used for both the restore and merge operations. The name of the file is of no significance, other than that it must end with the extension .zip, and that the extension must not conflict with any load file types. The zip file must contain one file with the extension .XML. This file must be an XML file in the Ascribe™ study archive format. The name of the XML file is also of no significance, other than that it must end with .XML. The study in Ascribe™ is identified by the value contained in the StudyIDClient tag of the XML document. If no study exists with this ID, a restore operation is performed. If there is an existing study with this ID, a merge operation is performed. Study Restore A restore operation creates a new Ascribe™ study. All of the information in the XML document is restored to the new study. Study Merge Merge operations allow you to add information from a saved study file to an existing study. The merge operation adds new items found in the saved file, but does not affect any existing items in the study. Do not think of the merge operation as an "update" of the 37 PrintDoc01 020813 study, except that it adds new information. Existing items in the study are not changed, even if the same item exists in both files. Identification of Items to Merge To use merge effectively, you need to understand how Ascribe™ decides to add an item to the existing study. The decision is based on the identifier for each type of item. If there is an existing item in the study with the same identifier, that item is not merged. In this case, the existing item is not affected by the merge operation. The identifiers used for each item type is listed below: Item Tag Identifier Question QuestionID Question ID Codebook CodeBookGuid CodeBookGuid Response DRORespondent Respondent ID Codes in a codebook are added to the study only if the codebook was itself added. The codebook will be added if there is no codebook in Ascribe™ with the corresponding CodeBookGuid. Study Questions Page Navigate: Supervisor/Studies/Right-click study/Select Questions The Questions page displays the questions for a study. The page has many options for working with the questions. You can get information about the questions, such as the question text and how much has been coded. Also, you can add and edit questions from this page. The filter bar at the top of the page is used to filter the list of questions displayed, by question type and by text. The Filter text box matches questions whose ID or label starts with or contains the text. To display questions whose ID or label starts with certain characters, type the characters in the Filter box and click Update. To display questions whose ID or label contains certain characters, type an asterisk as the first character in the Filter box, followed by the characters you want to find. For example, "*like" displays those questions whose ID or label contains "like". The Study ID and Study Name display above the question fields, on the left side of the screen. Each question has several fields. Because there are so many fields, you can choose which ones display from the Choose Columns page. For information about other right-click menu options, see Add and Edit Questions for a Study. When the number of questions is large (say more than 250), the performance of page refreshes is greatly improved by hiding the Shared Codebook and Shared Questions columns. These columns require a much more expensive query. If the user wants the functionality of being able to see the codebook in a drop-down without the expense of displaying these columns, he or she can instead display the Codebook ID column. This is fast to render, and provides the codebook drop-down just like the Shared Codebook column. 38 Codebooks Note Several columns on this screen display drop-down boxes when leftclicked. A hand icon displays when you hover over these fields. The drop-down boxes can open with a single click or a double click. Choose the method you prefer in User Options. To close the drop-down box, click Close, left-click twice in the box, or click once anywhere on the other columns. Here are the available fields for the Questions page: Field Description Sort Order This field displays the sort order selected on the Coding Tab. You can change the sort order in the drop-down box on this field or on the Coding Tab. In the drop-down box, you can click Next to increment the number. Click OK to save your change or Close to exit without saving. Question Type This field displays the type of question. You can change the question type at any time through the drop-down box on this column or if you edit the question. You can select the type of questions that display on the Questions page with the check boxes in the gray bar at the top of the page. After you make your selections, click the Update button to refresh the screen. Question ID A short identifier for the question. Each question ID should be unique within a study. Click this field to display a drop-down box with study and question information. Question Label A short name for the question. This label is used in the cross-tab report and in other places where a short form of the question name is required for display purposes. Question Text The field displays the question text information from the General Tab. Assigned To The Assigned To field displays coders and transcribers assigned to the question. To create or change assignments, left-click the field. A drop-down box displays with a list of users with coder and/or transcriber privilege. Click the box next to a user to select or unselect the user. Click the OK link to save the assignment and exit the drop-down box. Click the Cancel link to exit the drop-down without saving any changes. You can also make assignments on the Coding Tab of the Questions Properties dialog box. To assign coders by question types, use the Assigned To column on the Studies page. NOTE: When a study is saved or archived from your site, the Assigned To information is not retained. When a study is restored, the original Assigned To information is not available and will not display. 39 PrintDoc01 020813 Question Help The field displays the question help information from the General Tab. Coding Source This field displays the information from the Coding Tab. It refers to the portion of the response (verbatim, transcription or translation) coded by the coder. Related Questions This field displays the number of related questions. A drop-down box displays the question information of the related question or questions. Codebook Section Codebook ID The ID of the codebook used to code this question. A drop-down box displays codebook information. Shared Codebook This ID of the codebook that shares this question. A drop-down box displays codebook information. Shared Questions The number of questions (in any study) that also use this codebook. If this column is blank, no other questions use this codebook. Shared Questions Drop-Down Box To see what questions share this codebook, left-click the column and open the Shared Questions drop-down box. You can unshare the question with the unshare option at the top of the box. Codebook ACM If a question is a coding question for an ACM, the name of the model displays in this field; otherwise, the field will be blank. Codebook is Empty If the codebook contains no codes, a dot appears in this field. The field is blank if the codebook contains codes. ACM Section Code Using The ID of the ACM training used to code this question. A drop-down box provides more information. Approved If the ACM training specified for this question is approved, a check mark appears in this column. Coding Job The status of the ACM coding job last run for this question. A dropdown box provides more information. Responses Section Total The number of responses currently loaded for this question. Coded The number of responses that have at least one code applied. Codes Applied The number of codes that were applied to all responses. % Coded The percentage of responses that have at least one code applied. Transcriptions Section Transcriptions The number of responses that have a transcription. This column will display only if transcriptions are licensed for your account. % Transcribed The percentage of responses that have a transcription. 40 Codebooks Translations Section Translations The number of responses that have a translation. This column will display only if translations are licensed for your account. % Translations The percentage of responses that have a translation. Quality Reviewed Section Responses The total number of responses in the question. Reviewed The number of responses marked as quality reviewed. % Reviewed The percentage of responses marked as quality reviewed. Layout Section Card The card number of the question. Column The starting column position for codes. Columns The number of columns occupied by each code. Max Codes The maximum number of codes to output for this question. Tasks Section Code A dot displays if this task is selected for this question. You can edit this field through the drop-down box. Left-click the field, and the drop-down box displays. Click the check box to select or unselect the task. You can also use the Tasks Tab on the Questions Properties dialog to edit this field. Transcribe A dot displays if this task is selected for this question. You can edit this field through the drop-down box. Left-click the field, and the drop-down box displays. Click the check box to select or unselect the task. You can also use the Tasks Tab on the Questions Properties dialog to edit this field. Translate A dot displays if this task is selected for this question. You can edit this field through the drop-down box. Left-click the field, and the drop-down box displays. Click the check box to select or unselect the task. You can also use the Tasks Tab on the Questions Properties dialog to edit this field. Cross-Tabs A dot displays if this task is selected for this question. You can edit this field through the drop-down box. Left-click the field, and the drop-down box displays. Click the check box to select or unselect the task. You can also use the Tasks Tab on the Questions Properties dialog to edit this field. Add and Edit Questions for a Study Navigate: Supervisor/Studies/Right-click study/Click Questions/Right-click a question Questions can be added automatically to a study through the load or import processes. See Load Data to a Study for more details. 41 PrintDoc01 020813 When you right-click a question, you have options to add or edit a question or to navigate to other pages. Here are the options on the right-click menu: Option Description Studies This option displays the Studies page. Add A New Question The option creates a new question for the study. You enter the question components in the dialog box to add the question ID, question label, etc. Edit This option displays the dialog box where you can change any of the question components like question text and question help. Delete Questions This option allows you to delete questions and question elements. Related Questions The option allows you to relate this question with other questions. Edit the Codebook This option displays the Edit the Codebook page. Codebook Editor (Preview) This option displays the Codebook Editor. Desktop Codebook Editor This option navigates to the Desktop Codebook Editor. Copy/Share You have the following choices: Copy/Share Manager Copy/Share Codebook From... Copy/Share Codebook To... Code This Question The option displays the Coder Window page, which allows you to code one response at a time. Review Responses This option displays the Review Responses page. Edit Responses This option displays the Edit Responses page. Auto-Apply Codes Automatically code closed-ended and other specify questions. Phrase Analyzer The option displays the Phrase Analyzer page, which has advanced capability for coding short verbatims. Load Details This option displays the number of responses loaded, combined, etc., for a question. Accelerator (ACM) This option allows you to navigate to the ACM area, if a question is associated with a model. Also allows you to create a model from a question. Inspector This option allows you to create an inspection or view any inspections created from this question. Change Column Widths Navigates to the Change Column Widths dialog box, which allows you to choose the column widths for the columns on this page. 42 Codebooks Question Types The basic types of questions include open, closed, other specify, value, and data. Every question type may be coded in Ascribe™. However, data questions usually are not coded and often do not appear in the question lists. Data questions are used usually for output. Open questions are answered with verbatim text. This text is entered in by the respondent over the Internet, described verbally to a phone operator, or written. Open-ended questions also can be verbally recorded or hand-written and scanned (image files). Closed questions limit the responses to a fixed set of answers. A closed-end question can help you understand a respondent’s answer to an open-ended question. For example, the open-ended question may be: "You said x about y, why did you say that?", where x and y are answers to a closed-ended question. Closed-ended questions may also be used in the Study Analysis function to build cross-tabulated reports (see Cross-Tab Report.) When you load a closed-ended question, the data can be automatically coded if the codebook for the question has been defined and the loaded data matches the code values in the codebook. Here's how it works: Define the question ID to match the question ID in the data to be loaded. Create a codebook that matches the codes in the data to be loaded. For instance, if you have a five-number rating scale on a purchase intent question, then your codebook should use the same numbers. After you build the codebook, load the data, and it will be automatically coded. Ascribe™ matches the codes in the incoming data to those in the codebook. If you load the data before you create a codebook, you can still automatically code the data. Create a codebook, then right-click the question, and select Auto-apply codes. 43 PrintDoc01 020813 Other specify questions may have both open-ended and closed-ended data. The closedended data can be automatically coded when the data is loaded, if the codebook is set up correctly. However, the verbatims for other specify questions often contain multiple responses. When Ascribe™ finds the same question ID and respondent ID combination in a load file, the data is concatenated and separated with the vertical bar symbol (|). These responses will not be automatically coded. Value questions have a numeric response which will be coded as a range. "How many miles do you drive your car in a week?" is an example. Ascribe™ has a special way to automatically code value questions in the correct range in Phrase Analyzer. Question Components Navigate: Supervisor/Studies/Right-click study/Click Questions/Right-click a question/Select Add a New Question or Edit A question is made up of four tabs of information: General – Allows the user to set up the general information about the question. Tasks – Allows the user to choose tasks to be performed. These tasks govern whether a question is accessible for the function indicated. The defaults come from the account option settings under Administrator (see Site Configuration.) Coding – Allows the user to define different aspects of the verbatim. Data – Allows the user to determine how and where the data will be reported or downloaded. General Tab (Question) The user can set up the general information about the question. These fields are on the General Tab: Field Description Question Type Select the question type from the list. Question ID Enter a short identifier for the question, typically the question number. This value must be unique within the study. If you enter a name that is already used by a question in this study, Ascribe™ automatically creates a unique name by adding a colon and a number to the name you entered. The question ID determines where a response is loaded in Ascribe™ by matching it to the question identifier in the data file. Question Label Enter a short name for the question. This label is used in cross-tab report and in other places where a short form of the question name is required for display purposes. Question Text Enter the text of the question from the survey. The question text displays in the top pane of the Coder Window if you select Show question from the Settings option. (See Customize the Coder Window with Settings.) It also displays in the top pane of Review Responses. 44 Codebooks Question Help Enter information about a question here. Coders can view the question help when they select Question Info on the right-click menu in the Coder Window. Tasks Tab The user can select the tasks to be performed on the question. These tasks govern whether a question is accessible for the function indicated. The default comes from the account option settings under Administrator (see Site Configuration.) Field Description Code This Question If checked, the question will appear in the coders' questions to code list. 45 PrintDoc01 020813 Transcribe If checked, the question will appear in the transcribers' questions to transcribe list. This check box is available only if transcriptions are licensed for your account. Translate If checked, the question will appear in the translators' questions to translate list. This check box is available only if translations are licensed for your account. Cross-Tab If checked, the question will appear in cross-tab reports. This check box is available only if cross-tab reports are licensed for your account. Coding Tab The user can define different aspects of the verbatim. These are the fields on the Coding Tab: Field Description Sort Order This field determines where the question will appear in lists of questions for this study. Question lists are sorted first by this numeric value, then by the question ID. If you leave the value at the default of zero, questions sort alpha-numerically by question ID alone. Use the Study Setup function from Language Logic Tools in Excel ® to update this information easily. The Excel ® Add-In can be found under Downloads on the Ascribe™ Home page. Coding Source This field determines the part of the response that is displayed for coders to code the response. The default is to code from the verbatim. If transcriptions are licensed, you may want coders to code from the transcription. Similarly, if translations are licensed, you may want coders to code from the translation. Phrase Analyzer uses the coding source. If there is a saved parse, the coding source cannot be changed here, and this option will be disabled. Alternate Coding Source This field determines a second response field to be displayed when the primary coding source is empty. It is useful when working with translated responses where not all responses are translated. For example, a survey may have responses in both English and French, and only the French get translated to English. By setting this field to the verbatim, the various coding windows will display either the translation if there is one or the verbatim if the translation is empty. For more information, see the Alternate Coding Source section. Allow Duplicate Codes (not shown in screen shot below) This check box is available only if duplicate code assignment is licensed for your account. If this box is checked, the same code may be applied more than one time to a response for this question. This list includes only users who have coder and/or transcriber Coder and Transcriber List privilege. To select one user, left-click the name. Use control-click or shift-click to select multiple users. The highlight indicates that you 46 Codebooks have selected the user or users. This box is initially empty. Use the ">" button to move the selected user or users from the Coder and Transcriber List to the Assigned To box. (The user or users that are highlighted are the ones that will move.) To remove a user or users from the Assigned To box, select the user or users by a left-click, control-click, or shift-click. Use the "<" to move a user or users from the Assigned To box to the Coder List. (The user or users that are highlighted are the ones that will move.) Assigned To After you save the assignment and exit the dialog box, the Assigned To column on the Studies page and the Questions page displays the users who are assigned to this study. You can make changes to assignments on this tab, through the Assigned To column drop-down box on the Questions page, or the Assigned To column drop-down box on the Studies page. NOTE: When a study is saved or archived from your site, the Assigned To information is not retained. When a study is restored, the original Assigned To information is not available and will not display. Sort Order The sort order specifies where the question will appear in lists of questions for this study. Question lists are sorted first by this numeric value, then by the question ID. If you leave the value at the default of zero, questions will sort by question ID alone. 47 PrintDoc01 020813 The sort order also can be changed by using the Excel ® Add-In. After downloading the page to Excel®, you can make changes to the sort order and then load the file back into Ascribe™. When you load the file back into Ascribe™, the changes in sort order will be applied. Alternate Coding Source Navigate: Supervisor/Studies/Right-click study/Click Questions/Right-click a question/Choose Edit The Alternate Coding Source function allows users to see the translation or transcription of a verbatim as the coding source if a translation or transcription exists. If the field set as the coding source is blank (meaning there is no translation or there is no transcription), then the alternate coding source will be displayed. Usually, the alternate coding source is the verbatim (which refers to the original file that was loaded into Ascribe™). This tool is helpful in Review Responses when coding multi-language studies where respondents can answer in multiple languages, but a language question does not exist to separate the languages. For example, a French/Canadian study may have some French responses to be translated and some English responses not requiring translation. The alternate coding source allows the user to display the coding source and notes in Review Responses. This is especially helpful for those companies that track their other mentions using notes. Another reason to use the alternate coding source would be to code mixed media studies. A mixed media study is one that has some verbatims that need to be transcribed (either scanned images or voice recordings) and other verbatims that are "normal" text. The alternate coding source allows the user to view the transcription if a transcription exists and view the verbatim when there is no transcription. This way, the user still can track other mentions using the notes field without changing the View Options in Review Responses. Also, the screen is less cluttered. Set the coding source to either Translation or Transcription depending on your situation. 48 Codebooks While in Review Responses, you will see either the translation (if one exists) or the verbatim (if no translation exists). Be sure to set your View Options to display the Coding Source with Notes as the alternate coding view. The response displayed will have the text [Verbatim] after it if it has not been translated. See respondent 13170 below. Notice that respondent 13588 has the response hippo. This is the translation. You can rightclick the response and choose Edit Response to see the native language verbatim. (While you are in the Edit Response screen, you also can add a note if necessary.) 49 PrintDoc01 020813 Another option is to right-click the verbatim and choose Other Responses to view the native language verbatim. Data Tab The user can determine how and where the data will be reported or downloaded on the client side. This information can be set up after coding has taken place. The question set-up function included with the Language Logic Tools Add-In for Excel® allows the user to set-up all the data components of a study easily. See Use the Excel ® Add-In for Question Setup in Ascribe™ for more information. Users with Client access can edit the Data tab from the Client/Studies/Reports/Card Layout Report page. 50 Codebooks Field Description The layout setting affects the placement of the data when downloaded from Ascribe™. It allows the user to determine how the data will be written for each individual question. The layout setting is only active if you choose ‘Use Question Settings’ on the Data Tab of the Study Edit screen. Layout Punch – Obtains the column location from the output code value rather than from the "column" field on the question edit page. Punch using column offset – Obtains the column location by offsetting the output code value by the value in the "column" field on the question edit page. Computes the column for the punch as the sum (column value specified in the code) + (Column value of the question) Numeric – Creates a flat data file, where the output code value is truly written to the column identified on the question edit screen. The following three settings are used for column binary output. Card - The card numbers where the coded data for this question will be written. There are no limits on the number of cards. Column - The starting column where coded data for this question will be written. It is important to note that this is only for data that will be written in a card column format as numeric codes. The starting column is not relevant for multi-punched data unless you are using the Punch using column offset option. Columns - The number of characters that each code will consume in the data file. So, if you’re using 3 digit numeric values for your codes, then use 3 here. It is important to note that you will need to reserve enough space to accommodate the largest number of codes that you think you will apply to any given response. The Card Layout Report helps the user determine the amount of space needed for a question. Output Data Map The following four values are used when loading data from a Mapped Ascii Multiple response file. These are files that contain one row per respondent with the responses to each question contained at specified positions in the row. Input Data Map Start - The starting locations for data for this question in the Mapped ASCII-Multi load file type. Length - The lengths of the data in a load file for this question. Together with the start value above, this field defines the position and size of the response field for this question. Width - For questions with multiple response values, this value represents the width of any given response. Spread - For questions with multiple response values, this setting determines whether the response values are positional indicators (e.g., 0/1 only in a column) or represent data values. 51 PrintDoc01 020813 Maximum Codes The maximum number of codes to be written to the output file. Used in column binary and user defined output. Also available in user defined output as an output value. This value does not restrict the number of codes that may be applied to a response. It is used to limit the number of codes that are output. A value of zero means any number of codes may be output. Related Questions Navigate: Supervisor/Studies/Right-click study/Click Questions/Right-click the question that you want to be related/Click Related Questions Related questions allow coders to see the response to another question while in the Coder Window. For example, you have an open-ended response and you would like to see the closed-ended response that relates to this question. To relate the questions, right-click the question and click Related Questions. On the next screen, click the box next to the questions you would like to relate and click OK. Delete a Study and/or Components You must be a supervisor to use the deletion options. You can choose to delete the entire study, questions, coded responses, or loaded verbatims. 52 Codebooks Delete a Study Navigate: Supervisor/Studies/Right-click the study you want to delete/Click Delete… After you select to delete a study, the dialog box below displays: Type OK into the box, click the OK button, and you will erase the entire study from Ascribe™. It is important to note that once deleted, the study is gone and can not be retrieved. We suggest that you use the save study function to store your study on your hard drive before you delete any studies. Delete Questions Navigate: Supervisor/Studies/Right-click study/Click Questions/Right-click any question/Click Delete Questions When you select delete questions, this function allows you to delete components of multiple questions at one time. The check boxes that correspond to question components allow you to delete just the codes that have been applied (basically uncoding the question), the quality codes (see Set Quality Codes), or the responses. Or you can delete everything about the question. Check a box or boxes in the Delete column to select a question or questions for deletion. Make your selection for each question from the Delete Options section. Click the OK button when finished. Delete Loaded Verbatims Navigate: Supervisor/Loads/Right-click the load/Click Delete Supervisor/Studies/Right-click study/Click Questions/Right-click a question/Click Load Details/Click Delete 53 PrintDoc01 020813 When you delete a load, you delete all of the responses for that load (and only for that load). When you select to delete a load, this screen displays: A dialog box will display and ask you to confirm the delete. Type OK in the box, click the OK button, and the load will be deleted from Ascribe™. This action will delete all responses for the selected question which were loaded from the file on the date and time indicated. You also can see the load details and delete the load if you navigate to Supervisor/Studies/Right-click study/Click Questions/Right-click a question/Click Load Details. Delete Coded Responses Navigate: Supervisor/Studies/Right-click your study/Click Questions/Right-click the question/Click Edit Responses When you reach the Edit Responses page, right-click the response and click Delete Response. This dialog box displays: 54 Codebooks Click OK, and you will delete the response and any codes applied to it. Note The Edit Responses page is the only place in Ascribe™ that you can delete a respondent from one question. If you reload data through a second wave, the respondent will be added back into the question. Codebooks The codebook refers to the list of categories the coder will choose from when coding a response. The list of categories (or codebook) is used by the coder to organize open-ended or other specify verbatims so they can be further analyzed by a researcher. The codebook also is referred to as the code frame, code map, code list, code manual, codes or various other names. Codebook Components (Codes and Nets) Codebooks have codes and nets. Nets are sometimes referred to as " undups" because, when tabulated, they typically describe unduplicated respondents that belong to the net. Note You can have a codebook without nets in Ascribe™. When a codebook contains only codes, the codebook is said to be "straight listed." Presentation in Ascribe™ In Ascribe™, codebooks are stored and displayed in a hierarchy of parents and children (similar to a family tree). A codebook with nets looks like this in Ascribe™: 55 PrintDoc01 020813 The question is shown at the top of the codebook, with the codebook itself below the line. The codebook is almost always displayed on the right side of the Ascribe™ window. Net – In Ascribe™, a net starts out as a code. In order for a code to become a net, it must have children or sub-codes which belong to it. The codebook above has three nets (Spicy, Smoky, and Sweet). In Ascribe™, the term parent is the same as net. Smoky is the parent of both Hickory and Mesquite. Codes – Codes display indented below their nets. In Ascribe™, the term children means sub-codes or codes which belong to a net (parent). In the example above, Hickory and Mesquite are children of Smoky. Finally, siblings are codes with the same parent. Hickory and Mesquite are siblings. So are Spicy and Sweet. Peppery and Honey are at the same level in the codebook, but are not siblings. Create and Manage Codebooks Several options exist to create a new codebook or obtain an existing codebook for a question in Ascribe™. Edit the Codebook and Review Responses offer the most options. The table lists which Ascribe™ functions you can use for codebook creation and management. Task Ascribe™ Function Create codebooks manually/add codes -Edit the Codebook -Review Responses -Phrase Analyzer Copy codebooks from an existing study -Edit the Codebook -Study Restore Copy codebooks from an existing question -Edit the Codebook in the same study -Supervisor/Studies/Questions page Copy/paste a codebook from another source -Edit the Codebook Share codebooks from an existing study -Edit the Codebook -Review Responses -Study Merge Share codebooks between questions in the -Edit the Codebook same study -Supervisor/Studies/Questions page Unshare codebooks -Edit the Codebook Load codebooks from another source -Loader Change codebooks/ add and delete codes -Edit the Codebook -Review Responses -Phrase Analyzer Change the order of the codebook -Edit the Codebook -Review Responses 56 Codebooks Create nets for a codebook -Edit the Codebook -Review Responses Renumber codebooks -Edit the Codebook -Review Responses Change code properties/add expressions -Edit the Codebook -Review Responses -Phrase Analyzer Change codebook properties -Edit the Codebook -Review Responses Delete empty codebook -Edit the Codebook Review Responses and Edit the Codebook Right-Click Menus The Review Responses Window and the Edit the Codebook Window have the most functions to create or change codebooks. Both windows use a right-click menu for these actions. These menu options are the same in each window: Action Result Questions Navigates to the Questions page. Code This Question Navigates to the Coder Window for this question. Phrase Analyzer Navigates to the Phrase Analyzer Window for this question. Codebook Editor (Preview) Navigates to the Codebook Editor. Question Info Provides help text for the question. Card Layout Navigates to the card layout report. Insert Inserts a code at the level of the code you rightclicked. Insert as Child Inserts a code as a child of the code or the net you right-clicked. Delete (when the user right-clicks Delete this code. a code) Properties (when the user rightclicks a code or net) Edit the properties of the code or the net. Properties (when the user rightclicks the question) Edit the properties of the codebook. Renumber Codebook If you right-click the question, you can renumber the codebook. If you right-click a net, you can renumber a net. The option is not available if you right-click a code. Paste Paste copied text from other sources like Excel ® or Word. These actions are exclusive to Review Responses: 57 PrintDoc01 020813 Action Result Match Expression When you right-click a code, it searches for the regular expression associated with that code. Searches the responses by the options set in Expression Tester this dialog box. These actions are exclusive to Edit the Codebook: Action Result Sort Allows the user to sort alphabetically or numerically. Refresh Refreshes the screen. Check Spelling Uses Word to spell check the codebook. Delete (when the user right-clicks Deletes the entire codebook. the question) Edit the Codebook Window Navigate: Supervisor/Studies/Right-click Questions/Right-click question/Select Edit the Codebook Coder/Studies in Progress/Questions in Progress/Code This Question/Right-click/Select Edit the Codebook The Edit the Codebook window has two panes. The right pane contains the codebook. The question ID displays above the codebook. You can change the codebook from this screen with the right-click menu. View pane Codebook pane The left pane provides various views that are useful when you work on the codebook. Here are the display options in the view pane: View Description Code Dictionary This dictionary can be used to store frequently used codes. The codes in the dictionary can be dragged onto a codebook. To add a code to the dictionary, right-click in the dictionary view or on a code in the dictionary. Select Insert to add a sibling code, or Insert as child to add a sub-code. 58 Codebooks Question List Displays a list of all of the studies that are currently active and gives you access to the codebook for each question. Useful to copy or share codebooks between studies. Responses Displays the responses for the question/codebook in the right-hand window. Expression List Displays a list of all of the expressions that are programmed into the codebook. Duplicates List Use this option to check for duplicate outputIDs (especially if you have manually numbered the codebook and have unchecked No duplicate outputID’s in Codebook Properties.) Where Used After you switch to this view, click a code in the codebook pane. A list of the studies and questions that use this code will display. Codebook Editor Supervisor/Studies/Right-click a study/Select Questions/Right-click a question/select Codebook Editor (preview) Supervisor/Account Codebooks/Right-click a codebook/Select Codebook Editor (preview) Supervisor/Copy/Share Manager/Right-click a question/Select Codebook Editor (preview) The Codebook Editor combines some features of Edit the Codebook with the new Codebook Editing Tool, which has more functionality for changing codebooks. The Codebook Editor will replace Edit the Codebook in a future release. You can access this page in context of a question (from the Questions page or the Copy/Share Manager) or without the context of a question (from the Account Codebooks page.) Certain features of the Codebook Editor depend on whether you are working in the context of a question. The right pane of the Codebook Editor contains the Codebook Editing Tool. The left pane contains display options for responses and questions. Codebook Editing Tool: Menu Bar Left pane Codebook Display Right-Click Menus Navigation Keystrokes Left Pane Features 59 PrintDoc01 020813 When you access the page in context of a question, the left pane displays the study ID, question ID, and question label. Drop-down boxes are available on the study and question ID's to see the details of the study or question. Other features: A slider control at the top of the pane allows you to change the size of the pane. The right-click menu for the left pane gives you the standard navigation menu (Home, Help, Supervisor, Logoff, etc.) You can use the Excel® icon ( View options that are useful when you work on the codebook. They are: ) to put any of the displayed lists into Excel ®. View Description Question List Displays a list of all questions in the current study (when working in the context of a question.) Drop-Down Information Drop-downs are available on the question ID for question information and on the Shared Codebook column (even if the question is not shared) for codebook information. Edit Any Codebook from the List If you right-click a question in the list and select Codebook Editor, you can edit the codebook for that question. Responses Displays unique responses, listed in decreasing order of the count of the response. Source of Responses If you accessed the Codebook Editor via a question, only the responses for that question will display. If you accessed the Codebook Editor via an account codebook, responses from all questions that share the account codebook will display. Use Regular Expression to Search for Responses You can enter a regular expression to search for specific responses or leave the text box blank to display all responses. Click the Test It button to see results. The number of responses and percent of total responses are displayed below the Test It button. The expression text is highlighted in the response. 60 Codebooks Select Number of Rows In addition, you can specify the number of rows to display by entering a number or a percent. (Note that the percent option displays responses in a random, repeatable fashion.) This option allows you to efficiently view responses to questions with a large number of responses. You can use Regular Expression option and the Select option together. Click the Test It button to see results. The number of responses and percent of total responses are displayed below the Test It button. Create Code from Text You can drag any text onto a code in right pane to insert the text as a new code. Expression List Displays only codes that have a regular expression. Click a code in the list to highlight it in the right pane. Duplicates List Use this option to check for duplicate input and output code values (especially if you have manually numbered the codebook and have unchecked No duplicate outputID’s in Codebook Properties.) Click a code in the list to find it in the right pane codebook. When you correct a duplicate input or output code value, the corrected duplicate will display a line through it after the correction. Blank input and output IDs are ignored when searching for duplicates. Update after Making Changes to Code Values If you're using the Duplicates List and making changes to code values in the right pane codebook, you need to click Update to refresh the Duplicates List display. Changes to code values do not automatically display. Word List Displays the unique words for the first 5,000 responses in the question, list in decreasing order of occurrence. Displays only words that are used twice or more. Unused Codes Displays codes that have not been applied to any response in this question (does not check across shared questions.) Click a code to highlight it in the right pane. Select Number of Mentions Enter a number or a percent to display codes with counts (number of times applied to responses) that are equal to or less than these values. Shared Questions If the codebook is shared, a Shared Questions view is available. It lists the questions sharing the codebook. Drop-Down Information 61 PrintDoc01 020813 Drop-downs are available on the study ID and question ID. Unshare Codebook Click the Unshare button to unshare this question from questions in the list. Access from Account Codebook If you access the Codebook Editor from an account codebook, you are not able to unshare questions. Codebook Editing Tool Menu Bar The Codebook Editing Tool allows you to create and maintain codebooks. In the Codebook Editor, the Codebook Editing Tool is in the right pane. At the top of the tool is a menu bar. An icon on the left side of the menu bar indicates the current mode of operation. The tool has multiple modes of operation, which can be selected in the drop-down menu. The tool remembers your current mode if you navigate away from this page and return to it. Modes of Operation Each mode of operation has an icon associated with it. The modes of operation are: Mode Drag/Drop Redline Description In this mode, drag/drop operations are supported in a fashion similar to that currently provided in Review Responses. The visual effect is somewhat different. It has been changed to guard against inadvertent "drops" on the wrong location. Larger Redline The line that appears when you drag a code or net is much larger than currently used in Review Responses. It also appears whenever you drag over a code or net, not just when the pointer is over the blank region between codes. This is intended to make your drop targets much easier to locate without error. Grey Line The line appears grey when the pointer is not over it, and becomes red when the pointer is over it. Promote Code to Net As in Review Responses, dropping directly on a code will promote the code to a net, and place the item dragged into the net at the bottom. 62 Codebooks Drag/Drop Ctrl-Key If you find the visual effect of the red line obtrusive, you may prefer this mode. It functions identically to the Drag/Drop redline mode, except that you use the control key to indicate whether the drop should be into the target or below the target. Use Control Key to Add Child to Net If you hold the ctrl key down as you drop the item, it is the same as dropping directly on the code or net in the Drag/Drop redline mode: the item dropped is placed into the target as a child. Think ctrl = child. If the ctrl key is not pressed at the time of the drop, the item dropped is placed below the target. Multi-Select In this mode, drag/drop operations are not supported. This mode instead uses cut/copy/paste operations to copy and move codes. This makes it very appropriate for large codebooks where you want to move codes more than a screen height of distance. It also allows selection of multiple sibling codes. When multiple codes are selected you can Cut or Copy them and then Paste the multiple codes. You can also Delete multiple codes, and Place into Net the multiple codes. Select Multiple Sibling Codes To select multiple codes, hold the shift or ctrl key as you click with the mouse, in a fashion similar to the File Explorer in Windows. The ctrl key toggles the selection of the codes you click. The shift key selects codes in a range. If you click on a code or net that is not a sibling of the current selection, the selection is cancelled. Cut/Copy/Paste Operations Once you have selected multiple codes,you may rightclick in the selection area for a menu of available commands. A Cut/Paste sequence performs the same operation as dragging and dropping the cut codes (but you can cut and paste multiple codes instead of dragging single codes.) A Copy/Paste operation makes a copy of the codes at the paste location, but does not affect the copied codes. Paste Options When pasting onto a net, you have two paste options available. Paste will paste the code below the net (like dropping on the red line below the net). Paste as Child will paste the code into the net (like dropping on the net itself). Direct Edit In this mode there are no right-click menus. It allows editing of codes by simply clicking the text of the code or net and typing the changes. When you first click a code, the cursor is positioned at the end of the text. Click again or use the right and left arrow keys to 63 PrintDoc01 020813 move the cursor. Move Around Codelist Use the tab key to move down the list of nets/codes. Shift-tab moves up the list. Keyboard Shortcuts You can also edit a code by pressing the Enter key, insert codes by pressing the Ins key, or insert codes into a net by pressing the ctrl-Ins keys. Check Spelling Use this mode to check the spelling of the description field of codes and nets. Uses Microsoft Word Spell checking uses Microsoft Word, and will use the spelling options and dictionary you have selected in that application. If Word is not installed, spell checking will not work, and the right-click menus will be disabled. Errors in Red Spell checking is performed on the entire codebook when you enter this mode. When spell checking is completed, the spelling errors are underlined in red. The number of errors display in the menu bar. Suggested Corrections Right-click on a misspelling to see the list of suggested corrections that Word has provided, and click one to apply it. Alternatively you can press the Enter key to edit the description in the Edit dialog. Move to Errors In this mode, the shift-left and shift-right arrow keys move to the previous and next spelling errors respectively. Properties This mode is provided as a default mode to be used when the user lacks permission to edit the codebook. It displays a popup showing the properties of a code when you right-click on the code. Go To Menu Use this menu to navigate to another page. Click the desired page in the drop-down menu. The right-click menu used in Ascribe™ for general navigation does not appear in this tool, but the left pane does have this menu. Format This menu item controls what is displayed in the codebook. The description of the code is always displayed. You can select the following optional items for display. Display Description Input Code Value Displayed with an underline. 64 Codebooks Output Code Value Displayed with a colon following the value. Regular Expression Displayed on a new line below the code description. Hover Help Displayed in italics on a new line below the code description. Long Help Displayed on a new line below the code description. Show Details (Counts And Percents) Displayed in red to the right of the code description. It is not displayed for nets. If you modify the codebook, the counts and percents may be removed by the edit operation. To restore the values, click the Update link in the Format menu. (Long Help ) Displays a question mark icon at the start of each code if long help is defined for the code. To view the help, either hover over the question mark or click it. + / - (Expand/Collapse All) This menu item allows you to fully collapse or fully expand the nets in the codebook. To expand or collapse an individual net, click the red + or – icon to the left of the net. Summary This menu item displays a summary view of the nets in the codebook (if there are no nets, this menu item does nothing). At the top of the summary appear numbers showing the available levels, or depth of nesting, of the nets. The current level is displayed in red. To change the level, click the number of the desired level. When you click a net in the summary, that net is expanded in the codebook and all other nets are collapsed. This allows you to navigate quickly to a desired location within a large codebook. Find The Find menu item allows you to locate codes in the codebook. Type the text you want to find in the box and then press Enter. If a match is found, the code will be highlighted. The search is performed only on the text displayed in the codebook. For example, if a code has an output code value of 10, searching for 10 will find that code only if the output code value is displayed (using the Format menu item). When you have entered text in the Find box, you can use the left and right arrow keys to search for the next occurrence of the text. Press the left key to search upward, and the right key to search downward. The search starts from the currently highlighted code, or from the top of the codebook if you press the right arrow and no code is currently highlighted. The menu below the Find item provides options for the search: Option Description Match Case Performs a case-sensitive search; for example, "likes" matches "likes" but not "Likes". Match Whole Word The text you type must match a complete word in the codebook. Regular Expression The text in the find box is a regular expression. For 65 PrintDoc01 020813 example, "good|bad" will match either "good" or "bad". Search Collapsed Nets If this option is selected, the search will look at nets that have been collapsed (they display the + icon.) If a match is found in a collapsed net, it will be expanded. If this option is not selected, the search applies only to items that are visible. Codebook Display Codes are displayed with either a hollow circle or solid circle. A hollow circle indicates that the code has not been used to code a response (in any question.) This makes it eligible to be deleted, or to be changed to a net. A solid circle indicates that the code has been used to code at least one response (in any question.) Such codes cannot be deleted, nor can they be promoted to nets. Nets are displayed with a red icon, either a minus or plus sign. A minus sign indicates that the children (if any) of the code are visible. If there are children, you can click the minus icon to collapse the net, hiding the children and changing the icon to a plus. Clicking the plus expands the net again. Concurrent Edits This tool checks for updates by other users while you are modifying the codebook. If another user makes a change to the codebook, the next change you make will be rejected. The tool will display a message indicating that another user has updated the codebook, and your display will be refreshed. Right-Click Menus Each of the following items have right-click menu options: question ID, nets, and codes. The options available depend on the mode of operation selected. For example, Direct Edit, Check Spelling, and Properties do not use the right-click menus, and Multi-Select has some options unique to it. Many of the options open dialog boxes. These boxes may be dismissed by pressing the Esc key. You can reposition a dialog box (except in IE6) by dragging the title bar. Question ID Right-Click Menu Note: Direct Edit, Check Spelling, and Properties do not use the right-click menus. Option Description Codebook Properties Opens Codebook Properties dialog where you can edit the properties. Question Properties Opens the Edit Question dialog where you can edit properties of the question. Set Color Opens the Set Color dialog, which allows you to change the color of the entire codebook. Copy Text Opens the Copy Text dialog, which allows you to copy the entire codebook. 66 Codebooks Paste Text Opens the Paste Text dialog, which allows you to paste codes into the codebook. Change Case Opens the Change Case dialog, which allows you to change the case of the entire codebook. Sort Opens the Sort Codes dialog. Renumber Opens the Renumber dialog. Insert As Child Opens the Code Properties dialog, to add a new code. Delete Deletes the entire codebook; only available when no codes are applied in the codebook. If the codebook is shared, and the codes have been used in one of the shared questions, the delete option is not available. Net and Code Right-Click Menus Note: Direct Edit, Check Spelling, and Properties do not use the right-click menus. When you right-click a net or a code, the menus offer similar options. However, there are a few options unique to each one, and they are noted below. Option Description Properties Opens the Code Properties dialog where you can edit the properties. Set Color Opens the Set Color dialog, which allows you to change the color of the net and its children and other items at the same level. Copy Text Opens the Copy Text dialog, which allows you to copy the net and its children. Paste Text Opens the Paste Text dialog, which allows you to paste codes into the net. Sort Opens the Sort Codes dialog. (Net right-click menu only.) Change Case Opens the Change Case dialog, which allows you to change the case of the net and its children. (Net right-click menu only.) Renumber Opens the Renumber dialog. (Net right-click menu only.) Place Into Net Opens the Code Properties dialog so you can create a new net. The operation moves the highlighted code or net into the new net as a child. In multi-select mode, you can select multiple codes or nets for the Place into net operation. 67 PrintDoc01 020813 Demote To Code Turns a net into a code. You may also press the - key to perform this operation. This command is available only if the net has no children. (Net right-click menu only.) Promote To Net Turns a code into a net. You may also press the + key to perform this operation on the highlighted code. This command is available only if the code has not been applied to a response in any question. (Code right-click menu only.) Insert Creates a new code or net below the highlighted code; opens the Code Properties dialog. The Input and Output code values presented when you open the dialog are calculated using the current settings in the Renumber dialog. The initial color presented is based on the color of the highlighted code when you open the dialog. Creates a new code or net and places it into the highlighted net; opens the Code Properties dialog, with the additional choice of Insert at start or end of net. Insert As Child The Input and Output code values presented when you open the dialog are calculated using the current settings in the Renumber dialog. The initial color presented is based on the color of the highlighted code when you open the dialog. (Available only on the net right-click menu.) Insert Copy Creates a copy of a code or net below the highlighted code; opens the Code Properties dialog. The Input and Output code values presented when you open the dialog are calculated using the current settings in the Renumber dialog. The initial color presented is based on the color of the highlighted code when you open the dialog. Creates a copy of a code or net and places it into the highlighted net; opens the Code Properties dialog, with the additional choice of Insert at start or end of net. Insert Copy As Child 68 The Input and Output code values presented when you open the dialog are calculated using the current settings in the Renumber dialog. The initial color presented is based on the color of the highlighted code when you open the dialog. Codebooks (Available only on the net right-click menu.) Deletes the highlighted code or net. When deleting a net, all of the descendents of the net are also deleted. In multi-select mode, you may select multiple codes or nets for this operation. Delete Only codes that have not been applied to any response may be deleted. Only nets that contain no codes that have been applied to a response may be deleted. If the deletion will affect more than one code, you are prompted for confirmation. Cut (Multi-Select mode only) Cuts selected items so you can move them with a copy operation (see Cut/Copy/Paste Operations.) Copy (Multi-Select mode only) Copies selected items (see Cut/Copy/Paste Operations.) Navigation Keystrokes You can use these keystrokes to navigate around the right pane: Option Description Up/Down Arrows Move the highlight to the next visible code upward or downward. Ctrl-Up/Down Arrows Hold the ctrl key and press the up and down arrows to move to the next sibling of the selected code. Shift-Up/Down Arrows Hold the shift key and press the up and down arrows to scroll the codebook up or down. Set Color Dialog Navigate: Supervisor/Studies/Right-click a study/Select Questions/Right-click a question/select Codebook Editor (preview)/Right-click question ID, a net or a code/Select Set Color Supervisor/Account Codebooks/Right-click a codebook/Select Codebook Editor (preview)/Right-click question ID, a net or a code/Select Set Color This dialog allows you to change the color of multiple codes at the same time. You can change the color of the entire codebook or only selected nets and/or codes. Change Color of Entire Codebook Right-click the question at the top of the codebook to change the color of the entire codebook. You can change the color of just the nets or just the codes or both. See Set Color Options below. Change Color of Selected Nets and/or Codes Right-click a net or a code and select Set Color. You can change the color of nets and/or codes. See Set Color Options below. 69 PrintDoc01 020813 Set Color Options Option Description Text Color Select the desired color by clicking the box to the left of Text Color. Apply Color to Children A checkmark indicates that the change will apply to all descendents of the highlighted net. Apply Color to Level A checkmark indicates that the change will apply to all siblings of the highlighted code or net. Nets A checkmark indicates that the color of nets will be set. Codes A checkmark indicates that the color of codes will be set. Copy Text Dialog Navigate: Supervisor/Studies/Right-click a study/Select Questions/Right-click a question/select Codebook Editor (preview)/Right-click question ID, a net or a code/Select Copy Text Supervisor/Account Codebooks/Right-click a codebook/Select Codebook Editor (preview)/Right-click question ID, a net or a code/Select Copy Text This dialog displays the highlighted item and allows copying it to the clipboard so that you can paste it into another application (such as Word). If you display this dialog from the question at the top of the codebook, the entire codebook is displayed. If you display the dialog from a net, only the net and its children are displayed. If you display the dialog from a code, only that code is displayed. The format of the display uses the current options in the Format dropdown menu; however, it does not copy the counts and percents. Initially the entire display is selected. Right-click it and select Copy to place it in the clipboard. You can create a different selection with the mouse. To restore the entire selection, click the Select All button. In IE, the right-click Select All option will select the entire browser window, which is probably not what you want to do. Paste Text Dialog Navigate: Supervisor/Studies/Right-click a study/Select Questions/Right-click a question/select Codebook Editor (preview)/Right-click question ID, a net or a code/Select Paste Text Supervisor/Account Codebooks/Right-click a codebook/Select Codebook Editor (preview)/Right-click question ID, a net or a code/Select Paste Text This dialog lets you paste text into the codebook as new codes and nets. Paste or type the desired text into the box on the left. A preview of the results are shown on the right. You can adjust the results by editing the text on the left until you are pleased with the results. When you press the OK button, the new codes and nets are inserted below the highlighted code (the one you displayed the dialog from). New lines separate codes. Blank lines are ignored. 70 Codebooks Use white space (tabs or spaces) in front of the codes to adjust the nesting level of the codes. Tabs and spaces are equivalent, each indicate one nesting level. While you can paste tabs into the box, the browser will not let you type them. Use blanks instead. If a line starts with digits and is followed by an optional colon then text, the digits are assumed to be the Output code value. This is displayed in the preview with the number followed by a colon. Only the code description, and optionally the output code value, can be entered using this dialog. Change Case Dialog Navigate: Supervisor/Studies/Right-click a study/Select Questions/Right-click a question/select Codebook Editor (preview)/Right-click question ID or a net/Select Change Case Supervisor/Account Codebooks/Right-click a codebook/Select Codebook Editor (preview)/Right-click question ID or a net/Select Change Case You can change the casing of net and code descriptions to lowercase, uppercase, title case, and first letter capital. Some rules: You must be in Drag/Drop Redline, Drag/Drop Ctrl-Key, or Multi-Select mode to use Change Case. If you right-click a net, this operation applies to all descendents of the selected net. If you right-click the question ID, this operation applies to the entire codebook. Here are the options in the dialog: Option Description Codes Changes only codes. Nets Changes only nets. lower case Changes text to all lowercase. UPPERCASE Changes text to all uppercase. Title Case Changes each word to start with a capital letter. First letter capital Changes only the first word to start with a capital letter. (If a word has a special character such as * in front of it, the word itself will not be capitalized.) Renumber the Codebook Navigate: Supervisor/Studies/Right-click study/Select Questions/Right-click question/Select Edit the Codebook or Select Review Responses Supervisor/Studies/Right-click study/Select Questions/Right-click question/Select Codebook Editor You can renumber the codebook from Edit the Codebook, Review Responses or the Codebook Editor. When you renumber, you can change the inputID and/or the outputID for the codes and/or nets in the codebook. Renumber with Edit the Codebook or Review Responses 71 PrintDoc01 020813 To renumber the entire codebook, right-click the question at the top of the codebook and select Renumber codebook. To renumber only the codes in a net, right-click the net and click Renumber codebook. Select an option from the Value section: o Input code value – changes only the input value for each code or net. o Output code value – changes only the output value for each code or net. o Both – changes both the input and output value and gives them the same value. Select an option from the Renumber section: o Nets – renumber only the nets; codes remain the same. o Codes – renumber only codes; nets remain the same. o Both – renumbers both nets and codes. o Blanks only – renumbers only those codes or nets with blank values; non-blank values remain the same. Enter a value in the Start code value field. The first code or net will start with this number. Enter a value in the Spread between codes field. The codes are increased by this number. For example, if the start code value is 1, and the spread between codes is 10, the values assigned will be 1, 11, 21, 31, 41, etc. Enter a value in the Digits field or leave it blank. This field defines the number of digits in the code values and whether leading zeros are used. A blank or zero means the value will not be given leading zeros. A number greater than zero means the value will be given leading zeros. The maximum value is 20. Click the OK button. A dialog confirms the renumber. Click OK to see the renumbered codebook. Renumber with Codebook Editor 72 Codebooks To renumber the entire codebook, right-click the question ID and select Renumber. To renumber just a net, right-click the net and select Renumber to display the dialog. The settings in this dialog are also used to calculate the proposed input and output code values when adding codes. The Renumber dialog has these options: Field Description Use the check boxs to select Input code value and/or Output code value: Renumber Input code value – changes the input value for each code and/or net. Output code value – changes the output value for each code and/or net. Items To Renumber Use the check boxes to select one or more of these: Nets, Codes, or Blanks only. Blanks only renumbers only those codes and/or nets with blank values; non-blank values remain the same. Start Enter a value in this field. The first code or net will start with this number. Spread Between Codes The codes are increased by the number in this field. For example, if the start code value is 1, and the spread between codes is 10, the values assigned will be 1, 11, 21, 31, 41, etc. The default value for this field is set in the Codebook Properties dialog. Digits This field defines the number of digits in the code values and whether leading zeros are used. A blank or zero means the value will not be given leading zeros. A number greater than zero means the value will be given leading zeros. The maximum value is 20. Ignore (Regular Expression) You can prevent codes from being renumbered by typing a regular expression in the Ignore box. Codes with values that match this regular expression will not be renumbered. The input and output code values and the code description are searched for the expression. If the expression is found anywhere in the code description, the code will not be renumbered. 73 PrintDoc01 020813 Sort Codes Navigate: Supervisor/Studies/Right-click study/Select Questions/Right-click question/Select Edit the Codebook Supervisor/Studies/Right-click a study/Select Questions/Right-click a question/select Codebook Editor (preview)/Right-click question ID or a net/Select Sort The sort feature changes the order of the codes in the codebook. If the codebook has nets and you access the Sort Codes dialog from Edit the Codebook, the sort feature will not work on the entire codebook; you will receive a message which says "Cannot reorder nets." The sort feature will work if you right-click a net (it sorts codes within a net.) If the codebook has nets and you access the Sort Codes dialog from the Codebook Editor, the sort feature will not work on the entire codebook; however, top-level codes (codes not in a net) will be sorted. The sort feature also works if you right-click a net (it sorts codes within a net.) To sort the codes: If you want to change the order of the codes in a net, right-click the net and select Sort. If you want to change the order of a codebook that has no nets, right-click the question ID and select Sort. Choose a sort option: o Code description sorts the codes alphanumerically by the code text or description fields. o Output code value sorts the codes alphanumerically by the output code values. o Input code value sorts the codes alphanumerically by the input code values. o Descending sorts the codes in descending order within the Sort By option (available only from the Codebook Editor.) Click OK. A dialog box displays to confirm the sort (in Edit the Codebook only.) Enter OK in the box and click OK. A message displays to tell you the sort is complete. Click OK. Note A sort can not be reversed. You can sort the codes again by one of the three choices. However, if the original codebook was not in numeric or alphabetical order, the original order will be lost. 74 Codebooks Define Codebook Properties Navigate: Supervisor/Studies/Right-click study/Select Questions/Right-click question/Select Edit the Codebook or Review Responses or Codebook Editor/Right-click question ID/Select Properties Supervisor/Account Codebooks/Right-click codebook/Select Codebook Properties Supervisor/Studies/Right-click study/Select Copy/Share Manager/Right-click question/Select Codebook Properties The Codebook Properties dialog contains basic information about a codebook such as the codebook ID and description. If you access the dialog from the Codebook Editor, Account Codebooks, or Copy/Share Manager, you will have two additional fields (Account Codebook and Notes.) Field Description Codebook ID This field is a unique identifier for the codebook and allows you to access the codebook from external sources. It should be unique within all codebooks. An error will occur if you try to set this ID to a value that already exists for another codebook. Description This field provides a description of the codebook. You can use this field to provide information about the codebook, which may be useful particularly for shared codebooks. This field determines how the default value for inputID and outputID are calculated. You may use any number greater than zero. Spread Between Codes When the No duplicate outputID's check box is cleared, the default value for a new code is determined by inspection of all of the siblings of the code. The default value will be the first whole multiple of Spread between codes greater than the outputID of all siblings. Only siblings with numeric outputIDs are considered. No Duplicate OutputID's If this check box is selected, Ascribe™ will guarantee that the outputID of each code in the codebook is unique. When you assign a new value to the outputID of a code, Ascribe™ will change the outputID if necessary to ensure that it is unique within the codebook. If Ascribe™ changes the outputID, it will assign a large number. This is a visual indication that you need to correct the outputID. This check box is disabled if the codebook contains duplicate codes. To enable the check box, you must first ensure that all codes have unique outputIDs. Note that uniqueness applies to 75 PrintDoc01 020813 all codes in the codebook, including nets. Account Codebook If this check box is selected, the codebook will be designated as an account codebook. The codebook must have an ID to use this option. (Only available when accessed from the Codebook Editor, Account Codebooks, or Copy/Share Manager.) Notes Use this optional field to enter additional information about the codebook. (Only available when accessed from the Codebook Editor, Account Codebooks, or Copy/Share Manager.) Edit the Properties of a Code The properties dialog for a code displays when you add a code, or when you right-click a code or a net and select Properties. After you have edited the properties, click the OK button to update the code. More information on Code Properties: Field Description Description The description is the text of the code which displays in the codebook. If you leave the field blank, you will not see the code. Input Code Value The input code value is used for automatically coding responses to closed-ended, other specify, and value questions. The field may be blank if you will not use this codebook for automatic coding of closedended responses. Ascribe™ writes this field to the output file when the coded responses are downloaded. It is the “value” or outputID of the code. The codebook displays the field to the left of the code description. It can be a number or text. Output Code Value If you select the no duplicate outputID’s option in the codebook properties dialog, Ascribe™ checks for duplicate values in the codebook. If a duplicate is found, Ascribe™ changes the value you entered here to a large number. The large number reminds you to select a different and unique value. The field may be blank. Even if the no duplicate outputID's option is set, you may have multiple codes with no outputID. These are not considered to be duplicates. When you open the Properties dialog, the box is checked if the code is a net (the IsNet property of the code is true). The box is disabled (meaning you can’t change whether it is checked or not) if any of the following are true: Net The code has children The code is currently applied to a response The code is currently applied as a quality code If a code has not been applied to any responses, you can click the box to make this code a net. If this code has been applied to responses, you cannot change it to be a net. Regular 76 An expression is used to match words in a response. It helps the Codebooks Expression coder to pick the proper code for a verbatim. During coding in the Coder Window, Ascribe™ checks the response against all of the expressions in the codebook. When a match is found, Ascribe™ highlights and underlines the expression word or words found in the response. The coder clicks the highlighted words, and the corresponding code is placed in the selected codes pane. The Regular Expressions section has more information. The field may be blank if you do not want to use regular expressions. Hover Help This field provides short instructions to the coders for this code. This text displays when you place the mouse over this code in the codebook pane of the Coder Window and the Review Responses Window. The field may be blank if you do not want to use hover help. This field provides longer instructions to the coders for this code. You can enter as much information as you want or the field may be blank. Long Help (HTML) If you enter text here, a yellow question mark appears next to the code in the codebook pane of the Coder Window and the Review Responses Window. In the Coder Window, click the question mark to view the long help. In Review Responses, right-click the code and select Properties to view the help. The browser interprets the long help text as HTML. You can use HTML commands to format the text for the printed version of the codebook. For example, if you enter this: Here is the long help.< br> It contains <b>HTML</b> tags. It will display in the printed version as: Here is the long help. It contains HTML tags. This field controls the color of the code in the codebook. Use color to highlight new codes or frequently used codes or to separate positive, negative and neutral nets. To change the color of a code: Click the text color box to display the color choices. Click the desired color. Optional: To change the color of this code and the children of this code, click Apply color to children. (Not available when you access code properties from the Codebook Editor.) Optional: To change the color of this code and all codes and nets at the same level, click Apply color to level. (Not available when you access code properties from the Codebook Editor.) Click OK. The code or codes and nets change color based on the options you selected. Text Color Insert At Choose Start or End to insert the new code at the start or end of children of the net. (Available only from the Codebook Editor when you choose Insert as child.) Add Codes Navigate: Supervisor/Studies/Right-click study/Select Questions/Right-click question/Select Edit the Codebook or Select Review Responses 77 PrintDoc01 020813 To add codes to an empty codebook: Right-click the question text to start a new codebook. Choose Insert. The properties dialog for the new code displays. Enter the information for the code properties, and the code is added. If a codebook exists, you have several choices to add codes. To add a sibling code: Right-click an existing code at the position where you want to add the code. Choose Insert. The properties dialog for the new code displays. Enter the information for the code properties, and the code is added below its sibling. To add a child to the end of a net: Right-click on the net. Choose Insert as child. The properties dialog for the new code displays. Enter the information for the code properties, and the code is added to the end of the net. To create a net when you add a code: Verify that the target code has not been used to code a response. (You cannot create a net from a code that has been used to code a response.) Choose Insert as child. The properties dialog for the new code displays. Enter the information for the code properties, and the code is added below the new net. Spell Check Navigate: Supervisor/Studies/Right-click study/Questions/Right-click question/Edit the Codebook Check spelling uses the Microsoft Word spell checker and dictionaries to test the correctness of the spelling of code descriptions. To check the spelling of a particular code, right-click the desired code and select Check Spelling. If the word is misspelled, it will be underlined and displayed in red. To check the entire codebook, right-click the question and select Check Spelling. Misspelled words will be underlined and displayed in red. The entire code will be displayed within a black frame. To correct the spelling, right-click the highlighted word and click a word from the menu. If the menu is empty, it means that Word could not find a suitable suggestion. To manually correct the spelling, right-click the code and select Properties. Change the spelling in the Description field and click OK. Note When you are in the Edit the Codebook Window, Ascribe™ also checks the spelling of code descriptions whenever you add codes or change code properties. The spell checker "learns" as it works, and will operate faster as you continue to work in the Edit the Codebook window. When you leave the page, the spelling checker "forgets" the words it has learned, and will start slowly again the next time you enter the page. 78 Codebooks Because Ascribe™ uses Microsoft Word to check spelling, you can check spelling in multiple languages by setting up Word for the desired language. Ascribe™ will use the spelling dictionary that you have selected in Word. The spell checker can be disabled through the User Options under the General heading on the Home Page. 79 PrintDoc01 020813 Regular Expressions A regular expression is a pattern used to match text. Each code in a codebook can have an associated regular expression. You can enter expressions in the Code Properties dialog from Edit the Codebook, Review Responses or Phrase Analyzer. Code in the Coder Window with Expressions When a verbatim is displayed to a coder in the Coder Window, Ascribe™ compares it to each regular expression defined in the codebook for the question. If a match is found, the matched text is underlined and highlighted. The coder can hover over the highlighted text to see which code will be selected. If the coder clicks the underlined text, the code that matched the text is selected and placed in the selected codes pane. Simple Regular Expressions Use lower case letters in regular expressions. Ascribe™ views lower case and upper case letters the same. Do not use any symbols (such as $, +, -, &) unless you are an expert in their use. They execute powerful commands that can have unforeseen consequences. Letters and digits in the regular expression match the corresponding text in the verbatim response. For example: 80 Codebooks Verbatim response Regular expression Result I love cats love I love cats I love dogs dog I love dogs I LOVE DOGS o I LOVE DOGS Notice that these regular expressions simply match the same sequence of characters in the verbatim text. Upper and lower case letters are treated the same. Match Words In many cases, you will want to match whole words. Use angle brackets around characters to mean "match this word". For example: Verbatim response Regular expression Result The cat likes Catawba melon cat The cat likes Catawba melon The cat likes Catawba melon <cat> The cat likes Catawba melon Notice that the first regular expression matches the "cat" in Catawba, which is probably not what you want. When we put the "cat" in angle brackets, we match only that exact word. Often you will want to match words that begin with a certain sequence of characters. Use two angle brackets at the end of the word to mean "match words that begin with these characters". For example: Verbatim response Regular expression Result I like Cadillacs and Catalinas <cad>> I like Cadillacs and Catalinas I like Cadillacs and Catalinas < ca>> I like Cadillacs and Catalinas You can also match words that end with a certain sequence of characters. Use two angle brackets at the start of the word to mean "match words that end with these characters." For example: Verbatim response Regular expression Result I use USMail, email, and SnailMail <<mail> I use USMail, email, and SnailMail I use US Mail, e-mail, and Snail Mail <<mail> I use US Mail, e-mail, and Snail Mail I use US Mail, e-mailing, and Snail Mail <<mail> I use US Mail, e-mailing, and Snail Mail Notice in these examples that the definition of a word for matching is a contiguous sequence of characters. Word matching stops at punctuation marks and spaces. Finally, you can use two angle brackets at the start and end of the word to mean "match words that contain these characters." For example: 81 PrintDoc01 020813 Verbatim response Regular expression Result I send mail by USMailing, emailing, and SnailMail <<mail>> I send mail by USMailing, emailing, and SnailMail It is critical that the opening angle brackets are matched with closing angle brackets. Any of these examples would cause the regular expression matching not to work: >cat< <cat cat>> Verbatim Text Normalization To make the use of regular expressions easier, the verbatim response is "normalized" before it is compared to your regular expressions. The type of normalization depends on the language you have selected for your browser. For English, the following normalization is performed: This word in the verbatim response Is replaced with this word 1st first 2 second nd 3rd third can not cannot dont do not don’t do not wont will not won’t will not isnt is not is'nt is not n't not cuz because & and If the browser language setting is English and the study is in English, the results of the normalization changes are displayed on the coder's screen and are also used for regular expression matching. For western languages other than English, the following normalization is performed: 82 Codebooks These characters in the verbatim response Are converted to these characters àáâãäå A ç C èéêë E ìíîï I ñ N òóôõö O ùúûü U The browser language determines the display of the results of normalization. If the language setting is for a western language other than English: The diacritical marks do not display in the Coder Window The diacritical marks display in Review Responses Searches and expression tests ignore the diacritical marks You do not need diacritical marks when you enter simple regular expressions for codes. Use the letter without the diacritical. If the language setting is for English: The diacritical marks display in the Coder Window The diacritical marks display in Review Responses Searches and expression tests use the diacritical marks You need diacritical marks when you enter simple regular expressions (if you want to find that letter with its diacritical.) Match Phrases You can match exact phrases when you place the phrase inside angle brackets: Verbatim response Regular expression Result I love cats <love cat>> I love cats You can also match phrases that are bound by certain words. Use three dots in the regular expression to mean "skip up to thirty characters within this phrase." The three dots match any sequence of up to 30 characters that do not contain the following characters: .,;:?! Verbatim response Regular expression Result I love cats <love…cat>> I love cats I love white cats <love…cat>> I love white cats I love dogs and hate cats <love…cat>> I love dogs and hate cats I love dogs, cats, and mice <love…cat>> I love dogs, cats, and mice 83 PrintDoc01 020813 Notice that the third and fourth examples do not give the desired result. The third one matched, but the phrase is not the intended phrase. The fourth did not match because of the comma between 'love' and 'cats'. Match Negative Phrases You can use a tilde character '~' directly in front of a < or << to match the word 'not' in the preceding portion of the phrase. It is another way of writing <not>… Remember that in English, the contraction n't is changed to 'not', so you do not need to match the contraction when you create an expression. For example: Verbatim Response Regular Expression Result I love cats ~<love…cat>> I love cats I don't love white cats ~<love…cat>> I do not love white cats Match Commonly Misspelled Words A single dot character matches any character in that location. Use this to handle commonly misspelled words. For example: Verbatim response Regular expression Result Cadillacs and Catalinas <cad.l>> Cadillacs and Catalinas Cadallacs and Catalinas <cad.l>> Cadallacs and Catalinas The use of brackets with the single dot can also help with misspelled words. The characters .{1,2} tells the expression tester to look for one or two missing characters between the letters before and after the brackets. For example: Verbatim response Regular expression Result Niether this nor that <n.{1,2}ther> Neither this nor that Nither this nor that <n.{1,2}ther> Nither this nor that Nether this nor that <n.{1,2}ther> Nether this nor that Match Multiple Cases You can join regular expressions together with a vertical bar. The vertical bar means "match either of these expressions." For example: Verbatim response Regular expression Result I love cats <love…cat>>|<love…dog>> I love cats I love dogs <love…cat>>|<love…dog>> I love dogs You should only use the vertical bar to separate words surrounded by angle brackets, unless you are using the advanced features described in the next section. Advanced Use of Regular Expressions 84 Codebooks Simple regular expressions are sufficient for most uses. However, this section covers the additional applications you can use with regular expressions. Use them only if you have a clear understanding of how they work. You can use the special characters described in the table below to build more complex regular expressions. Character Description \ Marks the next character as either a special character or a literal. For example, "n" matches the character "n". "\n" matches a newline character. The sequence "\\" matches "\" and "\(" matches "(". ^ Matches the beginning of input. $ Matches the end of input. * Matches the preceding character zero or more times. For example, "zo*" matches either "z" or "zoo". + Matches the preceding character one or more times. For example, "zo+" matches "zoo" but not "z". ? Matches the preceding character zero or one time. For example, "a?ve?" matches the "ve" in "never". . Matches any single character except a newline character. (pattern) Matches pattern and remembers the match. The matched substring can be retrieved from the resulting Matches collection, using Item [0]...[n]. To match parentheses characters ( ), use "\(" or "\)". x|y Matches either x or y. For example, "z|food" matches "z" or "food". "(z|f)oo" matches "zoo" or "food". {n} n is a nonnegative integer. Matches exactly n times. For example, "o{2}" does not match the "o" in "Bob," but matches the first two o's in "foooood". {n,} n is a nonnegative integer. Matches at least n times. For example, "o{2,}" does not match the "o" in "Bob" and matches all the o's in "foooood." "o{1,}" is equivalent to "o+". "o{0,}" is equivalent to "o*". {n,m} m and n are nonnegative integers. Matches at least n and at most m times. For example, "o{1,3}" matches the first three o's in "fooooood." "o{0,1}" is equivalent to "o?". [xyz] A character set. Matches any one of the enclosed characters. For example, "[abc]" matches the "a" in "plain". [^xyz] A negative character set. Matches any character not enclosed. For example, "[^abc]" matches the "p" in "plain". [a-z] A range of characters. Matches any character in the specified range. For example, "[a-z]" matches any lowercase alphabetic character in the range "a" through "z". [^m-z] A negative range characters. Matches any character not in the specified range. For example, "[m-z]" matches any character not in 85 PrintDoc01 020813 the range "m" through "z". \b Matches a word boundary, that is, the position between a word and a space. For example, "er\b" matches the "er" in "never" but not the "er" in "verb". \B Matches a nonword boundary. "ea*r\B" matches the "ear" in "never early". \d Matches a digit character. Equivalent to [0-9]. \D Matches a nondigit character. Equivalent to [^0-9]. \f Matches a form-feed character. \n Matches a newline character. \r Matches a carriage return character. \s Matches any white space including space, tab, form-feed, etc. Equivalent to "[ \f\n\r\t\v]". \S Matches any nonwhite space character. Equivalent to "[^ \f\n\r\t\v]". \t Matches a tab character. \v Matches a vertical tab character. \w Matches any word character including underscore. Equivalent to "[A-Za-z0-9_]". \W Matches any nonword character. Equivalent to "[^A-Za-z0-9_]". \num Matches num, where num is a positive integer. A reference back to remembered matches. For example, "(.)\1" matches two consecutive identical characters. \n Matches n, where n is an octal escape value. Octal escape values must be 1, 2, or 3 digits long. For example, "\11" and "\011" both match a tab character. "\0011" is the equivalent of "\001" & "1". Octal escape values must not exceed 256. If they do, only the first two digits comprise the expression. Allows ASCII codes to be used in regular expressions. \xn Matches n, where n is a hexadecimal escape value. Hexadecimal escape values must be exactly two digits long. For example, "\x41" matches "A". "\x041" is equivalent to "\x04" & "1". Allows ASCII codes to be used in regular expressions. Check Expressions for a Specific Codebook Navigate: Supervisor/Studies/Right-click study/Click Questions/Right-click question/Click Edit the Codebook/Select Expression List from the Current View List 86 Codebooks The codes and their expressions display in the view pane as this example shows: Copy and Paste Codebooks from Excel® or Word Ascribe™ allows the user to copy and paste codes into the codebook from a text document or an Excel® spreadsheet. 1. Before you copy and paste the codes, prepare the source document so that the outputID of the code appears to the left of the description of the code, separated by a tab (or in a different cell in Excel ®). You can also indent the codes with tabs or empty cells to indicate the net level. An example of a properly formatted codebook in Excel®: 87 PrintDoc01 020813 An example of a properly formatted codebook in Word (with all characters displayed): 88 Codebooks 2. To copy the document, select the desired portion of the document and select Copy from the Edit menu. 3. In Ascribe™, navigate: Supervisor/Studies/Right-click a study/Select Questions/Right-click a question/Select Edit the Codebook or Select Review Responses. 4. To paste the codes into an empty codebook, right-click the question text, select Paste and click the OK button to insert the copied codes. 5. To paste the codes into an existing codebook, right-click the net or code where you want to insert the copied codes. Select Paste and click the OK button to insert the copied codes. (The Paste option under Edit in the Internet Explorer toolbar does not work to paste codebooks into Ascribe™.) 6. The codebook is copied into Ascribe™. The nets, the subnets and the codes use the value given in the Excel® spreadsheet or Word document. Delete Codes Navigate: Supervisor/Studies/Right-click study/Select Questions/Right-click question/Select Edit the Codebook or Select Review Responses or Codebook Editor You can only delete empty codes and codes that have no children. If a code has responses coded to it, remove the code from the responses. Once the code is empty, you can delete it. If a code is a net, first delete the child codes. When the net is no longer a parent, you can delete it. Delete a code by one of these methods: o Right-click the code and select Delete. o Left-click the code and press the Delete key. Move Codes Navigate: Supervisor/Studies/Right-click study/Select Questions/Right-click question/Select Edit the Codebook or Select Review Responses 89 PrintDoc01 020813 Codes are moved around the codebook by the drag-and-drop method. Left-click and hold the mouse button down on the code to be moved. Drag the code to the new location and release the mouse button. Rules for Code Moves To move a code to the top of a net, you will need to place it in-between the top code and the second code, and then move the top code below it. To move a code to the bottom of the codebook, you can drag and drop it on the Question Text. You can not make a code a net if it has been used to code any responses. To delete a net, you must first delete or move all of the children (codes) that report to the net. Drag & Drop Actions This table explains what happens during the drag and drop actions. Action Result Drop a code between two codes The code is placed at the level of the upper code as a sibling. The code is inserted between the two codes. Drop a code on top of another code The dropped code becomes a child of the code you drop it on. This action creates a net or a subnet. (See Rules for Code Moves.) Drop a code just below a net The code is not placed where you drop it, but is placed at the same level as the net. Drop a code on a net The dropped code is placed as the last code in the net. Drop a net on top of another net The dropped net is placed below the last code of the existing net. The dropped net becomes a child or subnet of the first net. Drop a net just below a net The dropped net is not placed where you drop it, but is placed below the last code of the net. It is at the same level as the net. Drag & Drop Visuals The examples below show the drag-and-drop visual clues. In the first example, a blue line means the code or net is placed as a child of the code or the net that it is dropped on. This action is the same as insert as a child. 90 Codebooks In the next example, a blue line with a red line means the code or the net is placed between two codes or nets. This action is the same as insert. Note When you drop a code on top of another code, the code you drop it on becomes a net. A code can only become a net if it has not been used to code a response. For this reason, the solid highlight will not always appear. It will appear only on those codes that have not been used to code a response. Share and Copy Codebooks In Ascribe™, a codebook can be shared by multiple questions, even by questions that are not in the same study. You can also copy a codebook from one question to another. There are several places in Ascribe™ where you can copy and share codebooks: Supervisor/Studies/Copy/Share Manager Supervisor/Studies/Questions page (Copy/Share From dialog, Copy/Share To dialog, Copy/Share Manager) Edit the Codebook page. Here are some guidelines for which method to use: Use the drag/drop method (which opens the Copy/Share From dialog) for a simple share for one question in a study. The dialog does contain options for getting codebooks from other sources on your account. Use Copy/Share To dialog for sharing multiple questions within a study. 91 PrintDoc01 020813 Use Copy/Share Manager for sharing multiple questions within a study. This option also has the most flexibility for finding source questions and codebooks (lists all studies, account codebooks, and automated coding models on your account.) While Edit the Codebook offers options for copying and sharing codebooks, this page will be removed in a future release of Ascribe™. Difference Between Copy and Share When questions share a codebook, there is only one codebook. If you make a change to the codebook, that change will be seen in all questions that share the codebook. When you copy a codebook, you create a new, separate codebook. If you copy a codebook from question A to question B, the codebooks are initially identical. However, there are two codebooks, and they are separate from each other. If you change the codebook for question A, the codebook for question B remains unchanged since it is a separate codebook. Copy/Share Manager Navigate: Studies/Right-click a study/Select Copy/Share Manager Studies/Right-click a study/Select Questions/Right-click a question/Select Copy/Share>Copy/Share Manager The Copy/Share Manager allows you to manage the codebooks of a study. It allows you to share multiple questions in a study at one time. It also allows you to find any codebook on your site and to copy or share any of the codebooks with the questions in your study. Copy/Share Manager Right Pane The right side of the Copy/Share Manager displays all questions in the study that have no coded responses. These questions are eligible to have their codebooks replaced by a copy or share operation. Here are the fields at the top of the right pane: Field Description Study ID The study ID of the current study displays at the top of the pane. Drop-down information is available. Study Name The study name displays next to the study ID. 92 Codebooks Group By You can display groupings by questions or codebooks. SHARE The Codebooks Shares the codebooks. Make A COPY Of The Codebook Makes a copy of the codebook. SHARE The Codebooks Shares the codebooks. Questions View When you group by Questions, here are the fields available in the table (you can choose the columns displayed by selecting Choose Columns from the right-click menu): Field Description Use the check box to select one or more questions which will receive a codebook with the copy or share operation. Codebook The Codebook column has drop-down information about the codebook. The icons give the status of the codebook: - The open book means the codebook has codes. - The open book with arrows means the codebook has codes and the question is shared. - The closed book means the codebook is empty. - The closed book with arrows means the codebook is empty and the question is shared. Question Type This field displays the type of question. Question ID A short identifier for the question. Question Label A short name for the question. Question Text The field displays the question text information from the General Tab. Codebook ID The ID of the codebook (if one was given in the Codebook Properties dialog.) Codebook Description Description of the codebook (if one was given in the Codebook Properties dialog.) Length The number of items in the codebook. Shared Questions The number of questions that share the codebook. Click the number to view the questions which share the codebook. You also have the option to unshare the question. Notes Notes about the codebook (if any were given in the Codebook Properties dialog.) The right-click menu for the Questions View has these options: 93 PrintDoc01 020813 Option Description Questions This option navigates to the Questions page. Question Properties This option displays the dialog box where you can change any of the question components like question text and question help. (Available on the right pane only.) Add A New Question The option creates a new question for the study. You enter the question components in the dialog box to add the question ID, question label, etc. (Available on the right pane only.) Codebook Properties Opens the Codebook Properties dialog where you can edit the information. (Available on the right pane only.) Codebook Editor (Preview) This option displays the Codebook Editor. (Available on the right pane only.) Code This Question The option navigates to the Coder Window. Review Responses This option navigates to Review Responses. Choose Columns Navigates to the Choose Columns dialog box, which allows you to choose the columns to be displayed on this page. Change Column Widths Navigates to the Change Column Widths dialog box, which allows you to choose the column widths for the columns on this page. Codebooks View When you group by Codebooks, the questions that share the codebook display before the codebook. If a codebook ID exists, the ID is listed first, followed by the length of the codebook. If no codebook ID exists, only the length of the codebook displays. The icons have the same meaning as when you Group by Questions. There is one additional icon; if you click the question mark in a balloon, a drop-down list with study information displays. The right-click menu for the Codebooks View has these options: 94 Codebooks Option Description Questions This option navigates to the Questions page. Add A New Question The option creates a new question for the study. You enter the question components in the dialog box to add the question ID, question label, etc. (Available on the right pane only.) Properties If you right-clicked a codebook, the Codebook Properties dialog opens where you can edit the information. (Available on the right pane only.) If you right-clicked a question, the Edit Questions dialog opens where you can change any of the question components like question text and question help. (Available on the right pane only.) Codebook Editor (Preview) This option displays the Codebook Editor. (Available only when you right-click a question on the right pane only.) Code This Question The option navigates to the Coder Window. Review Responses This option navigates to Review Responses. Copy/Share Manager Left Pane The left pane allows you to choose a codebook from several sources via the Choose From drop-down list. You can group by questions or codebooks (see Questions View and Codebooks View for more information.) When you group by codebooks, questions that are shared with a codebook may belong to another study. Those study IDs will have brackets around them. When you select a source via the Choose From drop-down list, that choice will remain until the next time you use the Copy/Share Manager. How to Copy and Share with the Copy/Share Manager In the left pane, choose the source of the codebook by using the Choose From drop-down list: A question in this study - All questions in the study are listed. A question in another study - A drop-down list of studies displays. Select a study, and the list of questions/codebooks from that study display. Account codebooks - All account codebooks on the site are listed. Accelerator coding models - All accelerator coding models are listed. Empty Codebook - Use the empty codebook to delete a codebook in a question or questions. Select the source question or codebook on the left pane, and then select the target question or codebook on the right pane. Because the Copy/Share Manager allows you to copy/share multiple questions or codebooks, you can select as many questions or codebooks in the right pane as you would like. Select SHARE the codebooks or Make a COPY of the codebook, and click OK. Note that when you use the copy option, you could be unsharing a previously shared question. But, the Copy/Share Manager lets you re-establish the share if needed. It can 95 PrintDoc01 020813 save you time if you are aware of the shared relationships between questions before you do anything. The double-arrow icon can help remind you of these relationships. Note that using the Empty Codebook option also affects the shared relationships of questions. You can create a share between questions if you select Empty Codebook and multiple questions and choose Share. Also, you will unshare questions if you use the Copy option. How to Unshare Questions These pages have ways to unshare questions: Copy/Share Manager Codebook Editor Questions Page Edit the Codebook Unshare Questions with the Copy/Share Manager To unshare a question with the Copy/Share Manager, you need to Group by Questions on the right pane. Open the Shared Questions drop-down box for the question you want to unshare by clicking in the Shared Questions column. (If the Shared Questions column is not displayed, right-click the question and use Choose Columns to display that column.) The unshare option is at the top of the Shared Questions drop-down box. Click it to unshare the question. Copy/Share From Dialog Navigate: Studies/Right-click a study/Select Questions The Copy/Share From dialog allows you to copy or share a codebook from another question, replacing the codebook of the selected question. It is available only when the selected question has no coded responses. If the question's codebook is not empty but has not been used to code responses, you will get a message which asks if you want to replace the codebook. Within the dialog, you may select from questions in the current study, questions in another study, an automated coding model, or an account codebook. There are two ways to access the Copy/Share From dialog: Drag/Drop method: In this method, you have a source question which contains the codebook that you want to copy/share. You also have a target question which receives the codebook from the source question. Highlight text in the row of the source question. Left-click and drag the text to the target question and release the mouse button. The Copy/Share From dialog displays. (Note that you do not have to use the source question you originally selected; you can get a codebook from another source in the Copy/Share From dialog.) Right-click a target question (a question that needs a codebook) and choose Copy/Share>Copy/Share Codebook From. The Copy/Share From dialog displays. Here are the options in the dialog: 96 Codebooks Option Description Choose From You can choose what question/codebook you give the target question. Your choices are: A question in this study - choose a question from the Question text box. A question in another study - choose a study from the Study text box. Account codebooks - choose one from a dropdown list that displays. Automated coding model - choose one from a drop-down list that displays. Study This field displays if you choose to get a codebook from a question in another study. Enter the beginning character of the study ID or name, and a list of studies that match that character will display in a drop-down list. (Only the first 10 studies that begin with the character will display so you may need to enter more characters. If you need to see more than 10 studies, you may want to use the Copy/Share Manager, which lists all of the studies.) Question Type You can filter the list of questions by selecting or unselecting the question type boxes (open,closed, other specify, or value.) Question If you used the drag/drop method and chose A question in this study, the Question text box defaults to the source question. If you used the right-click method and chose A question in this study or you chose A question in another study, the Question text box defaults to the first question of the study. The Question text box has a drop-down list of questions so you can select whatever question you would like. Make A COPY Of The Codebook Makes a copy of the codebook. SHARE The Codebooks Shares the codebooks. Click OK to finish the procedure or click Cancel to exit without copying or sharing the questions. Copy/Share To Dialog Navigate: Studies/Right-click a study/Select Questions The Copy/Share To dialog allows you to share the codebook of the selected question with any number of other questions in the study. The questions you share the codebook to must have no coded responses. 97 PrintDoc01 020813 In this method, you select a source question which contains the codebook you want to copy/share. To start the process, right-click your source question and select Copy/Share>Copy/Share Codebook To. The Copy/Share To dialog opens, and the source question remains highlighted in gray behind the dialog. Here is the format of the dialog box: Option Questions With No Coded Responses Description We can also refer to these as target questions. These questions can have their codebooks replaced with the codebook of the source question. The target questions may have a codebook, but they have not been used to code responses. The number in brackets before the question ID shows the number of codes in the codebook. If there is no number in brackets, then the codebook is empty. If the question's codebook is not empty, you will get a message which asks if you want to replace the codebook. Click a question to highlight it. To select multiple questions, press the Ctrl key while you click different questions. If the questions are next to each other, click the first question and press the Shift key while you click the last question in the range. Click the forward arrow button to move the selected questions to the right pane. These are the questions which will have their codebooks replaced. If you need to remove a question from the list of Selected Questions, click the question and click the backwards arrow button. Selected Questions These are the questions which will have their codebooks replaced. Make A COPY Of The Codebook Makes a copy of the codebook. SHARE The Codebooks Shares the codebooks. Click OK to finish the procedure or click Cancel to exit without copying or sharing the questions. Rules for Copy and Share Codebooks with Edit the Codebook The question that receives the copied or shared codebook must have an empty codebook. 98 When you attempt to copy or share a codebook and the question has a codebook, you will receive a message that a codebook exists. The copy or share will not work. Codebooks If you still want to copy or share to this question, you can delete the existing codebook if no responses have been coded. If responses have been coded, you have to remove the coding before you can delete the codebook. To delete a codebook, navigate to Edit the Codebook for the question. Right-click the question and select Delete. Enter OK to confirm the delete and click OK. The codebook is deleted, and you can continue with the share or copy procedure. Copy and Share Codebooks with Edit the Codebook Navigate: Supervisor/Studies/Right-click question/Select Questions/Right-click question/Select Edit the Codebook From the Edit the Codebook page, you can copy and share codebooks from any question, including questions in different studies. To copy/share a codebook from any study question: Navigate to Edit the Codebook for the question that will receive the copied or shared codebook. Select Question List in the Current View box. A list of studies displays. Find the study that contains the question you want to copy or share. Click the plus sign next to the study ID. A list of questions for that study displays. Highlight the question you are copying from and drag it to the right pane. Drop it on the question ID on the top right pane. Select Copy or Share in the dialog box that displays. Click OK and the codebook displays. Note To see the codebook before you copy or share it, right-click the question ID in view pane and select View Codebook. This example shows the View Codebook option: 99 PrintDoc01 020813 Copy and Share Codebooks to Multiple Questions with Edit the Codebook To copy or share a codebook from the current question being edited to one or more study questions: Navigate to Edit the Codebook for the question you want to share or copy. Select Question List in the Current View box. A list of studies displays. Find the studies or questions that will be the targets. Click the plus sign in the left column to expand the question list. Click the left-most cell of the question row to select the questions you wish to copy or share. The selected question will display a in the cell as shown in the example below. Left-click the question ID in the codebook pane to highlight it. Drag it from codebook pane and drop onto any of the selected questions in the view pane. In the dialog box displayed, select whether you want to copy or share the codebook. Note As you move your mouse over the left-most cell of the question rows in the view pane (left pane), the cursor will display a "hand" image if the question has no 100 Codebooks codes in its codebook. This symbol indicates it can be selected for sharing. Find Where a Code Has Been Applied in Shared Codebooks Navigate: Supervisor/Studies/Right-click question/Select Questions/Right-click question/Select Edit the Codebook If you want to know where a code has been used in shared codebooks, you can use the Where Used function in Edit the Codebook. Select Where Used from the Current View list in the view pane. Left-click any code in the codebook pane. The table displays the study ID and the question ID where the highlighted code has been applied. This search looks at both quality codes applied and regular codes applied. Unshare Codebooks with Edit the Codebook Navigate: Supervisor/Studies/Right-click question/Select Questions/Right-click question/Select Edit the Codebook The share between codebooks can be broken in two ways from Edit the Codebook. One way is to click the Unshare link located just under the question ID in the codebook pane. A dialog box displays. Click the OK button to confirm. A second way is to select Question List from the Current View list in the view pane. Click the plus sign next to the study ID to show the questions which share this one. Then rightclick the question to be unshared and select Unshare. A dialog box displays. Click the OK button to confirm. 101 PrintDoc01 020813 Note After you break a share between codebooks, the questions contain separate codebooks. The questions cannot be shared again unless one of the codebooks is deleted. In order to delete a codebook, you first have to delete any coding. Account Codebooks Navigate: Supervisor/Account Codebooks Codebook Editor/Go To Menu/Select Account Codebooks The Account Codebooks page displays a list of codebooks that have been designated as account codebooks. The purpose of account codebooks is that you can use them as master codebooks to be shared or copied to other questions. The codebooks will not be deleted if even no question references them. You can designate a codebook as an account codebook by accessing the Codebook Properties dialog from the Codebook Editor (question ID right-click menu option) or the Copy/Share Manager. The codebook must have an ID in order to be designated as an account codebook. Each codebook ID must be unique. If you no longer want a codebook to be an account codebook, right-click the codebook and select Edit to open the Codebook Properties dialog. Click the check box next to Account Codebook to change the designation. You can also access the Codebook Properties dialog from the Codebook Editor (question ID right-click menu option) or the Copy/Share Manager to change the designation. When you restore a study that had an account codebook, the account codebook designation is not restored. The Account Codebooks page has these columns (you can choose the columns displayed by selecting Choose Columns from the right-click menu): Field ID 102 Description The codebook ID. Codebooks Description Notes An optional description of the codebook. An optional field which lists additional information about the codebook. Length The number of codes and nets in the codebook. Shared Questions The number of questions that share this codebook. No Duplicate OutputID's If there is a dot in this column, duplicate outputID's are not allowed in the codebook. If this column is empty, duplicate outputID's are allowed. For more information, see No Duplicate OutputID's on the Codebook Properties dialog. Spread Between Codes This field displays the value entered in the Spread Between Codes field on the Codebook Properties dialog. ACM If a codebook is shared with a model, the name of the model displays. The right-click menu options for this page are: Option Edit Codebook Editor (Preview) Description Opens the Codebook Properties dialog so you can edit the fields. If you no longer want a codebook to be an account codebook, click the check box next to Account Codebook to change the designation. Navigates to the Codebook Editor. Because you are navigating from an account codebook, you will not have access to a Questions List on the View Pane side or be able to navigate to a Questions page from Go To menu on the toolbar. Note Drop-Down Boxes Several columns on this screen display drop-down boxes when left-clicked. A hand icon displays when you hover over these fields. The drop-down boxes can open with a single click or a double click. Choose the method you prefer in User Options. To close the drop-down box, click Close, left-click twice in the box, or click once anywhere on the other columns. 103 Verbatims Ascribe™ has several ways for handling verbatims, typically based on the coding task at hand. Code Verbatims in the Coder Window— In this window, the coder is presented with one verbatim at a time and can easily access other information to help decide how to code the response. Code Verbatims in the Review Responses Window— This function has advanced capabilities to search and manage the codes and responses. For example, you can search the verbatims for specific phrases and then code all those responses at one time. Code Verbatims in the Phrase Analyzer Window— This function is an advanced version of the auto-coder. It provides flexible facilities to separate verbatims for short responses, to build codebooks, and to code the responses. Multi-Media Coding—Ascribe™ supports coding from images and sound files. Coding from images is designed for scanned paper surveys, where the images are pictures of the handwritten responses. Edit Verbatims— You can edit both the respondents and the responses in a variety of ways. Translate/Transcribe Verbatims— Ascribe™ supports the translation of data into many languages and is completely Unicode compliant. Ascribe™ also supports the transcription of verbatims from both visual and sound files. Filter – Set a Response Filter— You can set a filter to view only those responses that meet your criteria. For example, you may want to see just the respondents that answered "Yes" to a previous question. Studies in Progress Navigate: Coder/Studies in Progress A study displays in this list only if both of the following are true: the study status is In Progress, and the Code task is checked in the Questions Properties for at least one question in the study. This page can also be filtered, based on the settings in the filter bar at the top of the page. The text filter in the bar allows you to filter studies by study ID or study name. To show only studies whose ID or name starts with certain characters, enter those characters in the Filter box and click the Update button. For example, if you enter " abc" in the Filter box, only those studies whose ID or Name start with abc will be shown in the list. To find studies whose ID or name contains certain characters, put an asterisk as the first character of the filter. For example, if you enter "* abc" in the Filter box, only those studies whose ID or name contains abc will be shown in the study list. The Assigned to Me check box means "only display studies that are assigned to me a coder (or a transcriber/translator.)" See Assign a Coder to a Study for more information. 105 PrintDoc01 020813 When you right-click a study, you have options to navigate to other parts of Ascribe™ or edit the study. See Studies in Progress Right-Click Menu for more information. Columns on the Studies in Progress Page The Studies in Progress page contains these columns (you can choose the columns displayed by selecting Choose Columns from the right-click menu): Field Due Date Description The date this study is expected to be completed. Study ID The unique identifier for the study. If you left-click this field, a drop-down box displays study information. You can navigate to the Study Question page if you click the Questions button on drop-down box toolbar. Study Name The name of the study. Questions The number of questions in the study. The drop-down box for this field displays information about the questions. Description A description of the study. Assigned To This column displays users assigned to this study. For more information, see Assign a Coder to a Study. Responses Coded Section Coded The number of responses coded. Total The total number of responses in the study. % The percentage of responses coded. Responses Transcribed Section The number of responses transcribed. Transcribed Total The total number of responses in the study. % The percentage of responses transcribed. Responses Translated Section Translated The number of responses translated. Total The total number of responses in the study. % The percentage of responses translated. Client The company assigned as the client for this study. End Customer The company assigned as the end customer for this study. Supervisor The associate designated as the supervisor of this study. The 106 Client Access to Data drop-down box displays supervisor details. Help Help for this study, intended to assist coders with information about the study. Note Several columns on this screen display drop-down boxes when leftclicked. A hand icon displays when you hover over these fields. The drop-down boxes can open with a single click or a double click. Choose the method you prefer in User Options. To close the drop-down box, click Close, left-click twice in the box, or click once anywhere on the other columns. Studies in Progress Right-Click Menu The right-click menu on the Studies in Progress page has the following options: Option Description Questions in Progress Navigates to the Questions in Progress page for this study. Edit Respondent Choose Columns Navigates to Edit Respondent page. Navigates to the Choose Columns dialog box, which allows you to choose the columns to be displayed on this page. Questions in Progress Navigate: Coder/Studies in Progress/Right-click a study/Select Questions in Progress Select the question you want to code from the list. Right-click on the question you want and select Code this question from the list. A question will appear in this list only if the Code task is checked in the Question Properties dialog or on the Tasks field on the Questions page. If you have transcriber privilege, you can also edit the responses in the question, whether they are coded or not. To review and edit responses, select Edit Responses from the rightclick menu. The filter bar at the top of the page is used to filter the list of questions displayed, by question type and by text. The Filter text box matches questions whose ID or label starts with or contains the text. To display questions whose ID or label starts with certain characters, type the characters in the Filter box and click Update. To display questions whose ID or label contains certain characters, type an asterisk as the first character in the Filter box, followed by the characters you want to find. For example, "*like" displays those questions whose ID or label contains "like". The check boxes at the top of the screen control the type of questions displayed. Click the boxes to select the question types, and then click the Update button. Only the question types that have check marks in the boxes will display. 107 PrintDoc01 020813 The Assigned to Me check box means "only display studies that are assigned to me a coder (or a transcriber/translator.)" See Assign a Coder to a Study for more information. The Study ID and Study Name display above the question fields, on the left side of the screen. Each question has several fields. Because there are so many fields, you can choose which ones display from the Choose Columns page. For information about other right-click menu options, see Questions in Progress Right-Click Menu. Note Several columns on this screen display drop-down boxes when left-clicked. A hand icon displays when you hover over these fields. The drop-down boxes can open with a single click or a double click. Choose the method you prefer in User Options. To close the drop-down box, click Close, left-click twice in the box, or click once anywhere on the other columns. Fields on the Questions in Progress Page Here are the available fields for the Questions in Progress page: Field Description Question ID A short identifier for the question. Each question ID should be unique within a study. Click this field to display a drop-down box with study and question information. Question Label A short name for the question. This label is used in the cross-tab report and in other places where a short form of the question name is required for display purposes. Question Text The field displays the question text information from the General Tab. Question Help The field displays the question help information from the General Tab. Question Type This field displays the type of question. Users with supervisor privilege can change the question type at any time through the drop-down box on the Question page or through the Questions Properties dialog. You can select the type of questions that display on the Questions in Progress page with the check boxes in the gray bar at the top of the page. After you make your selections, click the Update button to refresh the screen. Coding Source This field displays the information from the Coding Tab. It refers to the portion of the response (verbatim, transcription or translation) coded by the coder. Codebook ID The ID of the codebook used to code this question. A drop-down box displays codebook information. 108 Client Access to Data Assigned To The Assigned To field displays coders and transcribers assigned to the question. For more information, see Assigned To column on the Studies page. Responses Coded Section Coded Total The number of responses that have at least one code applied. The total number of responses for this question. % The percentage of responses that have at least one code applied. Responses Transcribed Section The number of responses that have a transcription. This column will display only if transcriptions are licensed for your account. Transcribed Total The total number of responses for this question. % The percentage of responses that have a transcription. Responses Translated Section The number of responses that have a translation. This column will display only if translations are licensed for your account. Translated Total The total number of responses for this question. % The percentage of responses that have a translation. Responses Referred Max Codes The number of responses in this question that were referred to a supervisor for more information. The maximum number of codes to output for this question. Questions in Progress Right-Click Menu When you right-click a question, you have options to code a question or to navigate to other pages. Here are the options on the right-click menu: Option Code This Question Description This option displays the Coder Window. Review Responses This option displays the Review Responses page. Edit Responses This option displays the Edit Responses page. Phrase Analyzer The option displays the Phrase Analyzer page, which has advanced capability for coding short verbatims. Auto-Apply Codes This option allows you to automatically code closed-ended and other specify questions. Studies in Progress The option displays the Studies in Progress page. 109 PrintDoc01 020813 Choose Columns Navigates to the Choose Columns dialog box, which allows you to choose the columns to be displayed on this page. Code Verbatims in the Coder Window Navigate: Coder/Studies in Progress/Right-click a study/Select Questions in Progress/Right-click a question/Select Code This Question A user with coder privilege only must navigate to the Studies in Progress page to see a list of studies that are ready to be coded. To select a study, right-click the study and select Questions in Progress. On the Questions in Progress page, the coder selects a question to code. To select a question, right-click a question and select Code this question. The Coder Window displays. The Coder Window Defined The Coder Window has several panes. These panes always display: verbatim, selected codes, and codebook. The other panes display if certain settings are used. The settings are described in the sections that follow. Question Verbatim Codebook Notes Selected Codes Summary Related Responses Question Pane This pane displays the ID and text of the question you are coding. The question pane appears only if you have selected Show Question in the Settings dialog. Verbatim Pane This pane displays the normalized verbatim to code. The gray bar above the verbatim displays the respondent ID next to the arrows and the number of responses that remain to be coded. The current verbatim is not included in this count. The gray bar also has these buttons: Button Description Click the Apply Codes button, and you will code the response with the codes that appear in the selected codes pane. The next verbatim then displays. Once you have applied codes to a verbatim, the verbatim will not be displayed again in this window unless you uncode the verbatim. This button is the same as the 110 Client Access to Data Apply Codes button in the codebook pane. This button shows you the first uncoded verbatim. It does not apply codes to the currently displayed verbatim. If you have selected codes for the verbatim, they will be discarded. If you have made changes in the transcription pane, these changes will be saved when you click this button. This button shows you the next uncoded verbatim. It does not apply codes to the currently displayed verbatim. If you have selected codes for the verbatim, they will be discarded. If you have made changes in the transcription pane, these changes will be saved when you click this button. Codebook Pane (Coder Window) This pane displays the set of available codes and has a variety of options for the coder. Gray Bar Options The gray bar above the codebook has these options. Button/Box Description Click the Apply Codes button, and you will code the response with the codes that appear in the selected codes pane. The next verbatim then displays. This button is the same as Apply Codes button in the verbatim pane. Search for a code value or text in the codebook: Type the text you want to find in the box next to the Find Code button and click Find Code or press Enter. Ascribe™ highlights in yellow the code that you found. To continue the search through the codebook, press Enter. Each time you press Enter, a code is highlighted if it matches your search. If no code matches your search, nothing is highlighted. Once a code is found, the Settings dialog controls how the codes are selected and the responses are coded. Check this box to display only responses that are reviewed. This setting uses the coding source for the response to determine whether the response is reviewed. For example, if the coding source is the verbatim, then only responses with the reviewed setting checked for the verbatim are selected. If the coding source is the translation, then only responses with the reviewed setting checked for the translation are displayed. Change the Codebook Display You can change how you view the codebook: To collapse the view of a net, click the bullet to the left of the net. This action hides the codes in the net and displays only the name of the net. To expand the view of a collapse net, click the bullet to the left of the net. This action displays the codes that were hidden. 111 PrintDoc01 020813 View Coded Responses for a Code Also in the codebook pane, you can view how a code has been used. Right-click any code text and choose Coded Responses. A dialog displays with the responses in this question that have this code applied. To close the window, click the Close button. Uncode Responses The Coded Responses dialog box also has an option to uncode responses. To select a response, click the box next to the response. You can select multiple responses. Next, click the Responses to be Recoded button. All codes are removed from the response, and it is placed back in line to be coded. Click the Close button to close the window. Notes Pane This pane allows you to enter notes about a verbatim. The pane displays only if you have selected Show Notes in the Settings dialog. You can enter notes for any response. For example, you can make notes about a portion of the verbatim that requires later coding. Regular expressions in the codebook display for both the verbatim and the note as you code. Three Methods to Create a Note You can create a note in several ways: Click just below the gray bar at the top of the notes pane. A blinking cursor appears. Type the note text. Highlight a piece of the verbatim and drag it to the notes pane. Copy a piece of the verbatim or any text and paste it in the notes pane. After You Create a Note Once you create a note, you have several options: Move to the next response without coding. Press Enter with the cursor in the notes pane. Ascribe™ saves the note and moves to the next response. Code the response and save the note. Select codes for coding, and click the Apply Codes button. Ascribe™ saves the note and applies the selected codes. The next response displays. Once you have coded a response, you can not change the note in the Coder Window without uncoding the response. You can change notes for coded verbatims in Review Responses (use the Edit Verbatims option from the rightclick menu in the response pane.) Delete the note. Click the note, and a blinking cursor displays. Use the backspace or the delete key to remove the entire note. You can continue coding that response or move to the next response. Selected Codes Pane This pane displays the codes you have selected for this response. The codes shown in this pane are the ones applied to the verbatim when you click the Apply codes button in the verbatim pane or the codebook pane. You can select codes in several ways: 112 Client Access to Data Click a code in the codebook pane Click a highlighted expression in the verbatim pane Use Find Code in codebook pane. To remove a code, left-click the code in the selected codes pane. It is removed from the pane. Related Responses Pane This pane displays responses to other questions that are related to this question. This pane displays only if a supervisor has set up related questions for the question you are coding. Summary Pane The summary pane displays only when you choose a number other than 0 for the Summary setting. To change this setting, select Settings from the right-click menu in the codebook pane. The summary pane displays only nets, and only to the level you choose in the Summary setting. When you click an entry in the summary list, the corresponding net in the codebook is expanded. All other nets in the codebook are collapsed. This function can help you find codes in large codebooks. How to Code in the Coder Window The verbatim pane displays the response to be coded. You can either use the codebook pane or use expressions to select codes. You can use both methods to code the same response. Code with the Codebook Pane Use these steps to code in the codebook pane: 1. Left-click a code in the codebook pane to select it. The code displays in the selected code pane. 2. Continue to select codes as necessary. 3. After you select enough codes, click one of the Apply Codes buttons. The codes are applied, and the next verbatim displays. Note You can not select a net for coding. You can only select codes. Code with Expressions Use these steps to code with expressions: 1. If Ascribe™ finds an expression for the verbatim, that text is underlined and highlighted. 2. Use the cursor to hover over the text to see the suggested code. 3. If the suggested code is the right one for the text, click the highlighted text to select the code. The code displays in the selected code pane. 4. Continue to select codes as necessary. 113 PrintDoc01 020813 5. After you select enough codes, click one of the Apply Codes buttons. The codes are applied, and the next verbatim displays. Code by Number If you like to work from the numeric keypad only, select all four of the Find Code options in the Settings dialog. These options help you code efficiently from the numeric keypad. Use this method to code: 1. Enter the number of the first code you wish to apply in the Find Code search box. Press Enter. The code is selected and placed in the selected codes pane. 2. Continue to enter numbers of additional codes and press Enter after each number. 3. After you have selected all of the codes, press Enter again. This action applies the codes and moves to the next response. Remove Codes If you decide that a selected code is not right, you can remove it before you apply it. Leftclick the code in the selected code pane, and it is removed from the list. Right-Click Menu The codebook pane has additional options for the coder, such as a right-click menu. To see the right-click menu options, right-click in the white space next to the codes. A menu with these options displays: Option Description Help Connects to the online help for this page. Question List Navigates to the list of questions for the study. Edit the Codebook Navigates to this question in Edit the Codebook. Codebook Editor (Preview) Navigates to the Codebook Editor. Review Responses Navigates to this question in Review Responses. Print Prints the codebook with a time/date stamp and includes any help text under the code in italics. (Entire codebook will be printed). Postpone Sends the current response to the end of the responses to make it the last response you will see. Useful when you wish to delay coding a response. Undo code Allows you to "undo" all codes applied to the last response coded. You may uncode as many responses as you code during a single session. Refer to Supervisor Refer a response to the study’s supervisor. Reclassify Reclassify a response to another question Filter Allows you to see only the responses that meet your criteria. Settings Allows each individual user to define the default settings for the coder window. 114 Client Access to Data Question Info Displays information from the Study and Question set-up screens. Other Responses Displays the current respondent’s responses to all of the questions in the study. Use this before reclassifying a response or if the respondent says "See previous answer.” All Responses Displays all responses for the current question. A table displays respondent id, verbatim, translation, transcription, notes, and codes applied. Open Sessions Displays the users who are logged into Ascribe™. Go To Gives you the default right-click menu. Refer to Supervisor Choose Refer to Supervisor when you wish to communicate with a supervisor about a response. Your supervisor will receive an indication that you have referred this response and will then reply. Use referrals to communicate questions, to request a new code or net, or to train new coders. Create a Referral Right-click in the white space next to the codebook and select Refer to Supervisor. A dialog box displays. Enter your reason for the referral, and click OK. When you refer a response, it is not available for coding until the supervisor has responded to your referral. 115 PrintDoc01 020813 Reply to a Referral To see a list of referrals, navigate to Supervisor/Referrals. If the My studies only box is checked, you will see only those studies assigned to you as the supervisor. If this box is not checked, all studies are displayed. Any supervisor may respond to a referral, even if that person is not the supervisor assigned to the study. To reply to a referral, right-click the referral and choose Edit. The study information and verbatim display, along with a response box. Enter your answer in the box and click OK. The verbatim is removed from the list of referrals and is available to be coded. 116 Client Access to Data View a Response to a Referral The next time the referred response displays to a coder, a button appears to the left of the response. Left-click the Responses to Referrals button, and a window displays with the supervisor’s reply. Reclassify a Response 117 PrintDoc01 020813 To reclassify a response means to copy the response from the original question to another question. This action is useful when the respondent gives a response to one question that is really a response to a different question. For example, a respondent gives a 'dislikes' response to a 'likes' question. In this case, you can reclassify the response to the 'dislikes' question. When you reclassify a response, the text of the response is copied to the target question. However, the response remains in the original question and still needs to be coded. (How you code the original response depends on your department’s coding rules.) Note Only text responses can be reclassified. Numeric and media responses (such as pictures and audio) can not be reclassified. If you try to reclassify a non-textual response, the list of available questions will be empty. If you try to reclassify a text response to a question where the same respondent has a non-text response, the reclassify operation will have no effect. Append a Response or Create a Response Two situations can exist when you reclassify. First, the respondent answers both questions (the original question and the question that is the target of the reclassify.) Or, second, the respondent only answers the original question. When the respondent answers both questions, the response is copied and appended to respondent’s answer in the target question. When the respondent only answers the original question, the response is copied to target question. However, this action does two things. First, a response is created in the target question for the respondent. Second, the number of responses to the target question increases by one. Therefore, the quota and load totals may not match at the end of the study. It is possible to have more responses than were originally loaded because of reclassification. Because reclassification allows the coder to create data, it should be used with care. In most instances, you should check the response to the target question with the Other Responses option. You may not need to reclassify if the idea was covered in the respondent’s verbatim to the target question. How to Reclassify To reclassify a response: 1. Select Reclassify from the right-click menu. The Reclassify dialog displays. 2. Choose the target question from the drop-down menu and click the OK button. Display of a Reclassified Response When you reclassify a response, a note is added before the copied response in both the original question and the target question. For example, if a response is copied from question 7 to question 8, this note would display in question 7 before the copied response: {-- copied to q8 --}. A note also displays before the response in question 8: {-- copied from q7 --}. 118 Client Access to Data Customize the Coder Window with Settings Ascribe™ allows you to change aspects of the Coder Window through the Settings function of the right-click menu. Just adjust the check marks illustrated below, and you can customize how the window works. The setting dialog has the following options: Section Option Windows Description Allows you to choose which panes display on the Coder Window. Summary Levels Displays the nets of the codebook in the summary pane. If you have a large codebook, you can use this option to see only a section of it. The number in the box controls the nets that display. Summary level 1 displays all of the nets. Summary level 2 displays both nets and subnets. Summary level 0 hides the summary pane. Show Notes Displays the notes pane. Show Question Displays the question ID and question text in the top pane of the Coder Window. Wide Verbatim Pane Allows you to extend the verbatim pane across the 119 PrintDoc01 020813 entire screen. This option is when coding large responses or scanned image verbatims. Magnification Find Allows you customize how the Find Code button works. Select Code Found Selects the first code that meets your search criteria and places it in the selected code pane. The effect is the same as if you clicked the code after it is found. You can click the Find Code button or press the Enter key to search and select codes. Each time you press Enter, the next code that is found is selected for coding. Clear After Find Clears the Find Code area after a code has been found. Apply Codes When Blank When you press Enter or click the Find Code button after you find all of the codes and the search field is blank, all the selected codes are applied. This setting works in conjunction with Clear after Find setting. Whole Word Limits the search to whole words only. This setting has the same effect as placing angle brackets around the text in the search field. Coding Allows you to customize your coding style. Double-click to Apply Codes Allows you to apply codes when you double-click anywhere in the codebook pane. If you double-click a code, it selects the code and applies all selected codes. If you double-click a net or the white space next to the codes, it applies all selected codes. Code Duplicates If this option is selected, Ascribe™ checks all of the uncoded responses when it codes a response. If it finds an exact textual match, those matches are coded with the same code or codes that were selected. You may not want to use this option if the responses need to be reclassified or if the responses are very large. Codebook 120 Allows the user to determine the size of scanned image verbatims that display in verbatim pane. A setting of 100% means no magnification, and the image is displayed at normal size. Settings below 100 shrink the image. Settings above 100 expand image. Allows you to customize the codebook display. Show Code Values Allows you to toggle code values on and off. Show Net Values Allows you to toggle net values on and off. Use Short Menus Breaks the right-click menu into shorter menus by category. Client Access to Data The Go to… menu displays these navigation options: Question List, Edit the Codebook, Review Responses, and Other Menus. The View… menu displays these view options: Help, Print, Filter, Settings, Question Info, Other Responses, All Responses, and Open sessions. The Actions… menu displays these functions that work on the current respondent: Postpone, Undo Code, Refer to Supervisor, and Reclassify. The short menus are displayed below: Code Verbatims in the Review Responses Window The Review Responses Window allows you to review and code multiple responses. It is different from the Coder Window which only permits coding of a one response at a time. Review Responses also has advanced capabilities to find and handle codes and responses. There are several reasons to use the Review Responses Window: Change coding and perform quality review. You can check the coders’ work, change coding, check miscellaneous mentions codes for ‘codeable’ ideas, adjust netting, and perform other quality review. Build codebooks. Review Responses has tools to analyze the verbatims and allows you to build codebooks based on this analysis. Expert mode coding. Review Responses is often used for coding. Users can search for words or phrases, view all responses that match the search criteria, and code several responses at one time. Also, you can use Review Responses to: Test expressions Code brand lists Delete codes Add codes to the codebook Edit codes in the codebook Move code from one net to another 121 PrintDoc01 020813 Copy & paste a codebook from Microsoft Word, Excel ®, or Notepad Swap one code for another to change coding View notes, transcriptions, or translations Edit notes, transcriptions, or translations View respondent’s responses to all the other questions within the study Filter on other questions within the study View counts and percents for the codes in a codebook See who applied codes to a response Renumber codebook Create a word count The Review Responses Window Defined Navigate: Supervisor/Studies/Right-click study/Select Questions/Right-click question/Select Review Responses The Review Responses Window has three panes: the response pane, the codebook pane and the button pane. Response pane Codebook pane Button pane Response Pane The response pane displays responses and the codes applied to the responses. When you first display the Review Responses Window, the response pane is empty. To fill the response pane, use any of the commands in the codebook pane, such as the expression tester, and the buttons in the button pane. When responses display in the response pane, you can code them or change their coding. You can also change the display of the responses through View Options on the button pane and with the response pane right-click menu. Codebook Pane The codebook pane displays the codebook. The study ID, study name, question ID, and question text display above the codebook. The codebook pane has a right-click menu that provides many of the same features as Edit the Codebook. For example, you can add codes, move codes, and edit the properties of codes. In addition, Review Responses has two unique features for viewing responses. They are match expression and expression tester. Match Expressions You can locate responses that contain matches to the regular expressions for codes. (Expressions are entered on the code properties dialog.) This function displays only uncoded responses. Coded responses are not matched. 122 Client Access to Data To find responses that match a single code: 1. Click the Clear button to make sure no other codes are highlighted. 2. Right-click a code and select Match Expression(s). Responses that match will display in the response pane. To find responses that match any of multiple codes: 1. Click the Clear button to make sure no other codes are highlighted. 2. Double-click the codes to select them. 3. Right-click and select Match Expression(s). Responses that match any of the expressions will display in the response pane. Expression Tester The expression tester has several options to help you find responses. You can find responses with expressions or words. You can use it to show all coded responses or uncoded responses. To display the expression tester dialog, right-click a code or the white space next to the codebook. The options for expression tester are: Option Description Enter the expression The text box and buttons below it control the search for responses. Enter the word or expression you want to use to match responses. to be tested You can enter a simple expression, a complex expression, a single word, several words, or just a period. The text box defaults to a period if you right-click a code that has no expression. A period by itself means ‘show all’ that match the button you select. For example, if there is a period by itself in the text box and you click show uncoded responses only, all of the uncoded responses will display. The text box defaults to the code’s expression if you right-click a code that has an expression. The Regular Expression section shows many ways to enter expressions for codes. These methods also apply when you enter expressions in the text box. Show Uncoded Responses Only Displays responses that have not been coded and that match the expression in the text box. Show Coded Responses Only Displays responses that have been coded and that match the expression in the text box. Show All Responses Displays all responses that match the expression in the text box. Both coded and uncoded responses display. From Current Result Set The current result set means the responses that are displayed currently in the response pane. This option limits the search to just the responses that are displayed. Return First ( Responses Leave the field blank to see all responses that meet the criteria. You can limit the number of responses if you enter a number in the field. ) 123 PrintDoc01 020813 Test It Click the Test It button to start the search. Cancel Click the Cancel button to close the expression tester dialog and cancel the search. Button Pane The button pane has the following buttons: Button/Box Description Responses - All Displays all of the responses that are coded with the selected code. If you select more than one code, Ascribe™ displays responses that are coded with all of the selected codes. It acts like the ‘and’ operator in a Boolean search. To use this function: 1. Double-click one or more codes in the codebook pane. The code or codes are highlighted with yellow. 2. Click the Responses – All button. The responses display in the response pane. Responses – Any Displays all of the responses that are coded with the selected code or codes. If you select more than one code, Ascribe™ displays responses that are coded with any of the selected codes. It acts like the ‘or’ operator in a Boolean search. To use this function: 1. Double-click one or more codes in the codebook pane. The code or codes are highlighted with yellow. 2. Click the Responses – Any button. The responses display in the response pane. Clear Removes responses from the response pane and removes the highlight from all codes in the codebook pane. View Options Allows the user to choose display and view options for the responses and codebook. Word List Displays a list of words and the frequency that they appear in the responses. It is helpful when you build codebooks because you can see what words are used. The Word List also displays the number of unique words in the responses, the total number of words in a question, and the average number of words per response. The Word List Options in the View Options dialog controls how the Word List is displayed. The Word List also has a feature for displaying responses. When you double-click a word, Ascribe™ displays all responses (coded and uncoded) that contain the word. Set Response Filter 124 Displays the Response Filter dialog, which allows you to filter the responses displayed in the response pane. Client Access to Data Select Responses and Apply Codes in Review Responses The Review Responses Window provides more flexibility in coding than the Coder Window. For example, you can code one response at a time or you can code several. Code a Single Response You can code one response at a time through the drag-and-drop method. Left-click a code in the codebook pane and drag it to the response. Drop the code when the response turns red. Code Multiple Responses You can code more than one response at a time when you select multiple responses. 1. To select multiple responses, double-click each response or hold down Control (Ctrl) and left-click each response. The codes will be highlighted in blue. (To deselect a response and remove the highlight, double-click the response again.) 2. Drag a code from the codebook pane and drop it on one of the highlighted responses. Each response will display the selected code. You can continue to drag and drop as many codes as necessary. 3. After you have applied all of the codes, right-click in the response pane and select Clear all responses. This action deselects the responses and removes the highlight. You can continue coding other responses. If you do not remove the highlight, you will continue to add codes to these responses. If you want to code all of the responses in the response pane at the same time, there is a quick method to select all of them. Right-click in the response pane and choose Select all Responses. This option is useful when you expression test for one word answers like ‘none’ or ‘nothing.’ 125 PrintDoc01 020813 Avoid Partially-Coded Data After you apply a code to a response, the response is considered coded. The response will not display when you use the expression tester to see uncoded responses. You want to make sure you have completely coded a response or established a way to track the uncoded portion before you perform new actions. For example, you apply one code to a response, but notice it needs a second code. If you press the Clear button or refresh your screen at this point, all responses on the left side of the screen are cleared from the screen. Any response with a code is considered coded and won’t appear when you bring up uncoded responses. You can see coded responses with the show coded responses only button on the expression tester. Or you can select a code and click the responses all button to see the responses that use that code. Right-Click Menu in the Response Pane The response pane contains the responses that you’ve selected either through the expression tester, or use of the responses all or response any buttons. After the responses are on the screen, you can right-click, and the following menu displays. Option Description Find Code Allows you to highlight part of a response and find a code that matches it. To use this option, highlight a word in a verbatim, right-click the highlighted word, and select Find Code. Ascribe™ finds and highlights the code that matches the word. Add to Notes Allows you to highlight part of the response and add it to the Notes field. To use this option, highlight a word or phrase in a response, right-click the highlighted words, and select Add to Notes. Ascribe™ adds the highlighted text to the existing note or creates a new note. To see the notes, the user should have the additional display field under View Options set to notes. Edit Response Allows you to edit the text fields (translation, transcription, and notes) for a respondent. (You can not edit the response here. You can use the Edit Responses Window to change the verbatim.) To use this option, right-click any response and select Edit Response. The Edit Response dialog displays. The dialog displays the response, the transcription, the translation and the notes fields. You can add, delete, copy or paste text into all of the fields 126 Client Access to Data except the response field. Mark as Quality Reviewed Allows you to mark responses as quality reviewed. To use this option, double-click responses to select and highlight them. Rightclick a highlighted response and select Mark as Quality Reviewed. A dialog box displays with the number of responses processed. Click Ok. The respondent ID changes color. (You choose the color in the respondent ID color field under View Options.) Clear Quality Reviewed Clears the quality review status from a response. It also changes the respondent id back to its original color. To use this option, double-click responses to select and highlight them. Right-click a highlighted response and select Clear Quality Reviewed. A dialog box displays with the number of responses processed. Click Ok. The respondent ID changes color. ACM - Add to Training Examples This option is only available when the question is associated with the model. It allows you to add the selected responses as training examples for the model. A dialog opens with these options: Language - If there is more than one language in the model, you can select which language to use. If there is only one language, that language is the default and cannot be changed. Quality Reviewed Responses Only - Add only responses that have been quality reviewed. New Responses Only - Add only responses that do not already exist in the model; duplicate responses will not be added. Coded Only - Add only responses that have been coded. Apply Coding - Add the responses and the coding. ACM - Update Training Examples This option is only available when the question is associated with the model. It allows you to replace the coding of matching examples in the model with the coding of the selected responses. A dialog opens with these options: Language - If there is more than one language in the model, you can select which language to use. If there is only one language, that language is the default and cannot be changed. Quality Reviewed Responses Only - Use only responses that have been quality reviewed. Clear all Responses Removes the highlight from the responses. The highlight means that a response is selected. When you remove the highlight, the response is no longer selected. After you clear the responses, the responses stay on the screen. Clear all Codes Removes the highlight from the codes. The highlight means that a code is selected. When you remove the highlight, the code is no longer selected. 127 PrintDoc01 020813 Select all Responses Highlights in blue all responses in the response pane. The highlight means that the responses are selected. Select all Same Code Allows you to select the same code from all of the responses currently displayed in the response pane. This is helpful when you want to swap one code for another or delete a code from a number of responses. Remove Selected Codes Deletes all of the codes you have selected. (Selected codes are highlighted in yellow.) You can not undo this action. Hide Selected Responses You can hide responses to make the screen less cluttered. Show all Images Expands all image files loaded into the study. Allows the user to browse through the images. Hide all Images Collapses all image files back to a link. Other Responses Allows you to see a list of that respondent’s answers to all of the questions in the study. This option is helpful for responses that say ‘see previous answers.’ See Who Applied a Code to a Response When multiple coders are assigned to the same question, you may need to see which coder applied a code to a response. When you hover over an applied code, the coder’s name displays at the bottom left corner of your screen. Ascribe™ keeps track of who applied which code to each response. When saving and restoring studies prior to 01/29/2010, Ascribe™ did not restore the name of the coder. After that date, the coder name is restored. If you use shared codebooks and need to see which question the response is in, you can see the question ID when you hover over the response or the applied code. 128 Client Access to Data See the Question ID When You Use Shared Codebooks When you code questions that share a codebook, you can view the question ID for each response. When you hover over the response, the respondent ID, or the codes, the question ID displays. See How Many Responses You Have Left to Code in a Question When you use the expression tester to display all uncoded responses, the number of responses and the percentage display in the bottom left corner of the screen. This number is displayed every time you use a filter, the expression tester, or display responses. If you 129 PrintDoc01 020813 use shared codebooks, the number and percentage include all responses in the shared questions. View Options View options allow you to select how the codebook and responses are displayed. After you select the options, click OK. Option Description Coding Source Coding Source determines which field will be returned for coding. It defaults to display the coding source set on the coding tab of the Question Edit screen. It is usually the verbatim. The coding source can also be translation or transcription if the coder needs to code from either of those fields. The coding source can be toggled off if the coder only wishes to view the Additional Display. Additional Display The additional display selection allows the user to choose another text field to view along with the coding source. The additional display text is displayed after the coding source text. It is displayed in italics to allow the user to distinguish between the coding source and the additional display. The options are: Additional Display Filter The first filter is related to the additional display field. It allows users to display responses based on whether or not the additional display text boxes are empty. The default setting is "With or without" (additional display). The options are: 130 None – No additional display Verbatim – Displays the verbatim which is useful when the coding source is set to transcription or translation Transcription – Displays the text in the transcription field Translation – Displays the text in the translation field Notes – Displays the text in the notes field. With or without (additional display) – displays all Client Access to Data Review Status Filter responses whether they have information in the additional display field or not Without (additional display) – displays only those responses that do not have text in the additional display field With (additional display) – displays only those responses that have information in the additional display field. This filter allows users to display responses based on the quality reviewed status. The options are: Any Review Status – displays responses with any review status Not Reviewed – displays responses that have not been marked as quality reviewed Reviewed only – displays responses that have been marked as quality reviewed. Coder Filter This filter allows users to display responses that were coded by all coders or by one coder. The default is all coders. You can select one coder from the list of coders who worked on the study. Coded Since This filter allows users to display responses that have been coded since a specified date. It defaults to "Any Date." Click the text box to display a calendar and select a date. To reset the filter, click Clear Date. Note that the calendar is based on GMT. Coding Method Filter This filter allows users to display responses by coding method. The default is Any Coding Method. The options are: Sort Order Manually Coded - displays responses that were coded in the Coder Window, Review Responses, Phrase Analyzer, or the Multiple Responses Window in Desktop. Any Auto Coded - displays responses that were coded automatically when loaded, with the Auto Coder script, or with Text AutoCoder in Desktop. ACM Coded - displays responses that were coded with the Automated Coding Model. Inspector Coded - display responses that were coded by Inspector. This field determines the order that responses are displayed in the response pane. The default is by respondent ID, but most users change the field to by response for coding purposes. The options are: By Respondent ID – the responses are displayed in respondent ID order (alphanumerically) By Question ID, Respondent ID – the responses are displayed alphanumerically by question ID and by respondent ID. This option is only for shared codebooks. By Response – the responses are displayed alphabetically by verbatim By Transcription - the responses are displayed alphabetically by transcription 131 PrintDoc01 020813 Code Book Display These fields control the display of the codebook. The options are: Response Source By Translation – the responses are displayed alphabetically by translation By Notes – the responses are displayed alphabetically by the notes field By Text Length - the responses are displayed from shortest to longest. Display Codes Only – displays codes in alpha order without any nets or subnets. You cannot move the codes around in the codebook (drag and drop actions within the codebook are disabled.) You can apply codes, add codes, deleted unused codes, and change the properties of codes. However, code swapping is disabled. Show Details – displays the code counts and percentages. The code count is the number of respondents that use the code. The percentage is calculated as the code count divided by the total number of respondents coded. Show Code Values – displays the output code value to the right of the code text Show Net Values- displays the output value of the net to the right of the net text. This field controls the access of data from all questions that share this codebook. If the field is toggled off, responses display only from the current question. Respondent ID Color This field allows you to customize the color of the respondent ID of those verbatims marked as reviewed and not reviewed. Show Respondent ID This field allows you to display or not display the respondent ID. When you do not display the respondent ID, you will not know if the response has been marked as quality reviewed since the respondent ID color gives this indication. You will also not be able to use the right-click menu option to see the respondent's other responses. You can still code the responses. Word List Options These options control how the Word List is created from the responses. The options are: Delimiter – A list of characters that define where the responses are split before you view the word list. The default characters are the space character as well as ,./. Characters to Ignore – A list of characters to ignore when the responses are displayed in the word list Words to Ignore – A list of words to ignore when the responses are displayed in the word list. Here are some common words to ignore: a able about actually also always am an and any anything are as at be because been being but by can could did do does for from gave get had has have how I if im in is it its itself ive just like liked many may me might much my of on or really said say says seemed seems so such tend than that the their them then there these they think this through to 132 Client Access to Data too until very was we were what when where which who why will with within would yet you your Here are some common characters to ignore: 1=234567890+)(*&^%$#@!~’}{|\][“:;><,./!- Swap One Code for Another Ascribe™ has a quick way to swap codes or replace one code with another. For example, you may have multiple responses that with the same code that needs to be changed. A swap takes less steps than if you add a code to each response and then delete the code you did not want from each response. To swap the same code for multiple responses: 1. Find all of the responses with the code that needs to be changed. One option is double-click the problem code in the codebook pane and right-click the Responses All button. All of the responses coded with this code display in the response pane. 2. Select the responses you wish to swap. One method is to right-click in the response pane and choose select all responses. Each response is highlighted in blue. You can remove responses from the selection if you double-click them. 3. Find the code you want to swap below one of the highlighted responses. Left-click the code and drag it to the desired code in the codebook pane. (You must drag the code from the response pane to the codebook pane.) 133 PrintDoc01 020813 4. When the code in the codebook pane turns blue, and a plus sign displays, stop the left-click. A dialog box displays and asks, ‘Do you want to swap these codes for the selected responses?’ 5. Click OK. The codes are swapped, and the responses remain highlighted until you clear them. 134 Client Access to Data Create a New Code from a Combination of Codes Sometimes you may want to create a new code from some existing codes. For example, you notice a pattern among the respondents. You decide to create a new code called Really, Really Easy for all of the respondents who were coded with both WORKS WELL/QUICKLY and EASY/QUICK TO APPLY. In the codebook pane on the right-hand side: 1. Highlight both WORKS WELL/QUICKLY and EASY/QUICK TO APPLY. Click the Responses Any button. The responses that have been coded with either code display in the response pane. 2. Add a code called Really, Really Easy to the codebook. (To add a code, rightclick the codebook and select insert. Enter the code properties.) 3. Code each of the responses displayed in the response pane. (To code all of the responses, right-click any response and choose select all responses. Then drag the Really, Really Easy code to any of the responses.) 4. Delete the unwanted codes from the responses. Double-click the unwanted code WORKS WELL/QUICKLY beneath any of the responses to highlight it yellow. 5. Right-click the highlighted code and choose "Select all same code". Every WORKS WELL/ QUICKLY code on the left side of the screen is highlighted yellow. 6. Right-click again and choose remove selected codes. 7. Repeat steps 5 through 7 for the EASY/QUICK TO APPLY code. Code Verbatims in the Phrase Analyzer Window The Phrase Analyzer feature is useful for coding short response answers, like other specify and brand lists. With it, you can code responses to all questions that share a codebook, code responses automatically, and create codebooks. There are two basic ways to use Phrase Analyzer: When you already have a codebook in place, like a brand list When you create a codebook from the loaded data. 135 PrintDoc01 020813 In each case, it is useful to prepare the loaded data and combine the responses that have the same meaning, and then apply the codes or build the codebook. Phrase analyzer will also allow you to limit responses to a group of numerics only. The Phrase Analyzer Window Defined Navigate: Supervisor/Studies/Questions/Right-click question/Select Phrase Analyzer Phrase Analyzer has a window with three panes: the codebook displays in the right-hand pane, the parsed responses display in the left-hand pane, and the option buttons are in the bottom pane. Parsed Responses pane Codebook pane Options pane The Parsed Responses Pane The left-hand pane in Phrase Analyzer displays a list of "parsed" responses. It displays portions of responses which have been parsed or split apart using some specific character or characters as the splitting points. These characters are referred to as delimiters. The initial view shows responses for a question which have been split apart using the vertical bar character ("|"). This action splits apart multiple responses to the same question for a respondent. The display groups all the same parsed response elements on one line together with a count of how many were combined into that line. 136 Client Access to Data It is important to recognize that this display shows a list of pieces of responses. It is not necessarily a list of respondents’ verbatim answers to a given question. Depending on how it was parsed, a given respondent may actually be part of a number of lines in the list. In the left-hand pane, you are able to manipulate the response lines with these options: Re-parse them Visually combine the same ideas Create codebooks in an automated way from the response lines View respondents that make up the parsed response lines. Right-Click Menu in the Parsed Responses Pane The parsed responses pane has a right-click menu. Here are the available functions: Function Description Find Code This function finds a code that matches selected text. To use this function, select a piece of text in a response, right-click in the left pane, and select Find Code. Ascribe™ searches the codebook for that text. When the text is found, the code is highlighted in the codebook pane. Save Parse Delete This Parse This function saves your current parsed responses. You should use this function if you want to stop coding but want to use these parses again. This function deletes the current parsed segments. New Response Parse This function allows you to parse the responses and enter which characters to use as delimiters and which characters and words to ignore. Parse Line This function splits or parses the response. To use this function: Put the cursor where you want to split the response. Press Alt and then left-click. The delimiter character displays. Right-click in the parsed responses pane and select Parse Line. The line splits where the delimiter was placed. Parse Selection This function parses the responses with the selected text. To use this function, highlight some text and right-click in the parsed responses pane. Select parse selection. The lines are split before and after the selected text, with each part put on a new line. Parse Expression You may use an expression to parse the responses. The matched text is removed from each line and grouped into its own line, but the line itself is not split apart. The matched segment of the text is replaced with a ‘~’. The combined line segment is moved to the top of the list of segments. Combine Select the response lines that you want to combine, click the combine option, and the responses will be combined. The count of the combination is equal to the individual counts of the lines that you combined. 137 PrintDoc01 020813 Add to Notes This function adds selected text to the coding notes for a response. This option is only available if some text has been selected or highlighted. Add Selected This function adds the text of the currently uncoded segment to the Uncoded Segment to coding notes for a response. An uncoded segment is one that has a Notes strikethrough displayed. It is created when another segment from the same response has been marked for coding, and the Apply Codes button was clicked. Add All Uncoded Segments to Notes This function adds the text of all uncoded segments to the coding notes for a response. An uncoded segment is one that has a strikethrough displayed. It is created when another segment from the same respondent has been marked for coding, and the Apply Codes button was clicked. Ignore Current Uncoded Segment This function allows you to ignore this segment when you apply codes. Ascribe™ applies codes even though not all of response has been marked for coding. Ignore Selection This function causes the selected lines to be ignored. They remain unprocessed and uncoded. Clear Selection Codes This function removes the codes from that selection. Clear All Codes This function removes all codes from all of the responses. Create Code This function creates a code from the selected text and places it in the codebook pane. The created code will be applied to the selected text. Make Codebook This function creates an entire codebook from the response lines. You must enter how many mentions at which to build the codebook. View Respondents This function displays a list of all respondents included in the combined line item. Note Ascribe™ does not automatically save your parsed data. In order to leave the page and return later, save your work with the "Save Parse" option. How to Parse Responses To parse responses means to split apart words by a delimiter character. The initial view of a question’s set of responses is parsed by the "|" character. Ascribe™ uses this same character when it combines multiple responses from the same respondent during the loading of data. If there are no multiple responses in the data from a given respondent, the list simply shows the responses to a question grouped by response. Identical responses are grouped together without considering case. You can get a new parse of the entire response set when you choose New Response Parse from the right-click menu. You can enter delimiter characters as well as characters and 138 Client Access to Data words to ignore during the parse. The delimiters default to "|". Additional delimiters can be added. For example, if the response was loaded with quotes as part of the data, it would display just as it is, with no additional delimiters. However, if a comma is added to delimiters, it becomes: But, the data is not grouped into matches because the quotes are still part of the data. If you add quotes as a delimiter, Ascribe™ matches the data and groups it into three lines: You can enter as many delimiters and characters to ignore as you would like. You can process the existing parse further with the Parse Selection option on the right- click menu. This action can be useful when a codebook does not exist. For example, you have these responses in the display. Highlight the word "coke" in one of the lines: Right-click and select Parse Selection from the menu. You now have: 139 PrintDoc01 020813 Next, select Combine from the right-click menu. You know have: Now, right-click the line with the word coke and select Create Code. A new codebook entry is created, and the word coke will be marked for coding. You can also split up individual lines when you position the cursor at the point you wish to split, press Alt and left-click. A vertical bar ("|") appears. Repeat anywhere you wish to add additional line splits, including other lines. When ready, select the Parse Line option from the right-click menu. You also can group responses manually through a drag-and-drop process. To do this, highlight text within a row and drag it to a different row. The highlighted text disappears. The count of items in the target row increases by one. If you highlight an entire row and drag it, the entire row disappears. This dragging process is useful for words that re misspelled. For example, you have these responses: I like coke cokie is great The second line will not auto code or group together with the other "cokes". If you highlight the second line and drag it to the "I like coke line", the two will now be coded as one. Phrase analyzer allows you to limit responses to groups of numerics only. This action is useful when auto coding a question with both text and numeric responses, such as an Other Specify. To select this option, check the "Numerics Only" option in the new parse dialog window. If you have created a new set of parsed data with any of the above methods, you need to save if it you wish to leave this window and return. PARSED DATA AND TEXT MARKED FOR CODING ARE NOT SAVED AUTOMATICALLY. Save the parsed data with the Save Parse option from the right-click menu. The Options Pane The options pane provides these selections: 140 Client Access to Data Function Description From This Question The responses to be parsed come from only the selected question. From Shared Codebook Questions The responses to be parsed come from any question within the current study that share a codebook with the selected question. Respondent View This option displays the respondent ids of responses which have been marked for coding. Show Details This option displays the responses for each code. Show Nets This option displays the nets for the codebook. When it is not used, the codebook displays in alphabetical order with only the codes. Apply Codes When you press the Apply Codes button, you apply codes to all of the responses that are marked to be coded. This action is different than most other areas in Ascribe™, where codes are applied in real time. The Codebook Pane The right-hand codebook views are similar to those in the Review Responses and the Edit the Codebook windows with several significant differences. The default view displays only the codes without nets. It lists them in alphabetical order rather than codebook order. This display is useful when you code from brand lists and other non-concept type of codes. The Show Details view lists the codes in alphabetical order, but also will expand each code to show the responses coded with that code. To expand or contract the display, double click the code or click the +/- icons. Finally, the "Show Nets" view displays the codebook with nets. You can not rearrange the codebook, renumber the codes, or insert codes here since the codebook does not display in its true order. (Codes can be added with the Create Code 141 PrintDoc01 020813 option from the right-click menu in the parsed responses pane.) You can edit the properties of a code and delete unused codes. Here are the right-click menu options for the codebook pane: Function Description Questions This option navigates to the Questions page. Code This Question This option navigates to the Coder Window. Review Responses This option navigates to the Review Responses Window. Edit The Codebook This option navigates to the Edit the Codebook Window. Codebook Editor (Preview) The option navigates to the Codebook Editor. Auto Code This option attempts to match a code to a response segment. If matched, the response segment is tagged for coding. Delete This option deletes a code from the codebook if the code has not been used. Properties This option displays the code properties dialog. You can change the code properties here. The Phrase Analyzer Coding Process In general, the coding process is similar to the Coder Window. You select codes to be applied either by auto-coding or manually. When ready, you click the Apply Codes button to actually code the responses. As you select codes to be applied to particular text (either entire rows or portions of text within a row), the text is "marked" with the code. This action is similar to the Coder Window where codes are listed in the selected codes pane as they are selected. When you use the auto-coder option, the marked text is indicated by a color change. The first code applied changes the text to red, the second to green, and the third to yellow. You can see which codes are applied when you hover over the colored text. Auto Code in Phrase Analyzer When you auto code, Ascribe™ attempts to match a code to a response segment. If matched, the response segment is tagged for coding. Codes are matched to response segments with these rules. When the question is not closed ended: If a portion of the text is highlighted, Ascribe™ uses the highlighted text to match a code. If an expression exists for the code, Ascribe™ uses the expression to match a code. If the code text is highlighted, Ascribe™ uses the code text to match the responses. When the question is closed ended, Ascribe™ uses the input IDs of the codes to match the responses. You can also auto code in these ways. Place your cursor on a net, and you auto-code all of the codes in that net. Place it on a code, and you auto-code only that code. Place it on the 142 Client Access to Data question text, and you auto-code all the codes in the codebook. Any time a match is made, the matched text in the response line will be color-coded in red, green, or orange depending on the order that it has been coded. This screen was generated by auto-coding the entire codebook. If you wish to uncode something, use Clear Selection Codes or Clear All Codes options from the right-click menu. The former removes the marked codes from the current row, and the latter removes all marked codes from all rows. When ready, click the Apply Codes button. Any row with marked text will have the corresponding respondents coded with the marked code. Those respondents will be considered coded, and any text from their responses will no longer display in the parsed response list. If a response has been parsed into several segments, and not all of the segments have a code applied, no codes will be added for that response. Instead, you will receive a message that not all codes could be applied. Segments without a code will display with a strikethrough. To continue coding, either ignore the strikethrough or apply codes to those segments. Auto Code Value Questions Value questions have a numeric response which will be coded as a range. In order to auto code them with Phrase Analyzer, the input code value should be specified in this format: n:n, where the n's are the numbers in the range. For example, a code that represents the range of 1-10 should have an input value of 1:10, 11-20 should have an input value of 11:20, etc. Once the input values are correctly set, you can use Phrase Analyzer to auto code the responses. Right-click the code and select Auto Code. The selected responses will be highlighted. Repeat for each code, and click Apply Codes to apply the codes. Manual Coding in Phrase Analyzer In addition to the auto-code process, you can mark text for coding manually in several ways. 143 PrintDoc01 020813 First, you can drag a code from the codebook to a row of parsed text. This action marks that row with the appropriate code. You can also drag entire rows or highlighted portion of rows from the parsed response view to a code in the codebook. Finally, you can drag responses from a non-marked row to a marked row. This action codes the non-marked row’s respondents the same as the marked row. Remember to save your parsed views if you want to leave the screen and continue your work later. When you are finished, click the Apply Codes button. Any row with marked text will have the corresponding respondents coded with the marked code. Those respondents will be considered coded, and any text from their responses will no longer display in the parsed response list. If a response has been parsed into several segments, and not all of the segments have a code applied, no codes will be added for that response. Instead, you will receive a message that not all codes could be applied. Segments without a code will display with a strikethrough. To continue coding, either ignore the strikethrough or apply codes to those segments. Multi-Media Coding Ascribe™ supports coding from images and sound files. Coding from images is designed for scanned paper surveys, where the images are pictures of the handwritten responses. The user can code the image in the Coder Window or in the Review Responses Window. In the Coder Window, the written image will be visible. When you code in Review Responses, double-click the word "image" to view the written image. 144 Client Access to Data Coding from sound files is designed for working with voice files from IVR systems. The user can code sound files in the Coder Window or the Review Responses Window. In the Coder Window, you can listen to the sound file if you click the play button in the audio box. In Review Responses, click the blue text "Audio" to listen to the sound file. Transcribe IVR and Image Files Navigate: Transcriber/Transcribe/Right-click study/Select Questions/Right-click question/Select Transcribe Most users find it helpful to transcribe responses from IVR and image files prior to coding them. You can do transcription from the Transcribe page. 145 PrintDoc01 020813 Audio Keyboard Shortcuts The following default key combinations can be used to control the media player. You can change the keyboard key if you click the text box of the shortcut and enter a different key. Note that not all browsers or operating system will support all key combinations. Activity Keyboard Shortcut Start/Pause Ctrl + S Start Over Ctrl + B Advance 30% Ctrl + Right Arrow Rewind 30% Ctrl + Left Arrow Volume Up Ctrl + Up Arrow Volume Down Ctrl + Down Arrow Auto Translate Ctrl + T Next Response Ctrl + Enter Next ID/Respondent Ctrl + Page Up Previous ID/Respondent Ctrl + Page Down Check Spelling Ctrl + R Edit Verbatims Ascribe™ offers two ways to edit verbatims. Edit Respondent allows you to view and edit a single respondent’s answers to every question in a study. Edit Responses allows you to view and edit all responses to a question. Edit Respondent and Edit Responses 146 Client Access to Data Edit Respondent allows you to view and edit a single respondent’s answers to every question in a study at one time. Edit Responses allows you to view all responses to a question. Both windows have similar functions. The difference is that Edit Respondent displays one respondent at a time, and Edit Responses displays all responses to a single question. Here are the functions of the two windows: View and edit verbatims: Correct the spelling of verbatims, fix verbatims wrongly reclassified from the coder window, and update a response with new information. View and edit transcriptions: Transcribe audio and image files and quality check transcriptions (mark as reviewed.) View and edit translations: Translate foreign languages and quality check translations (mark as reviewed.) View and edit notes: Review notes and remove unwanted notes. Add a response. Delete a response (without deleting the entire respondent.) Edit Respondent Window Navigate: Coder/Studies in Progress/Right-click study/Edit Respondent Transcriber/Transcribe/Right-click study/Edit Respondent Transcriber/Translate/Right-click study/Edit Respondent The window is divided into two text columns. The column on the right can be edited, and the column on the left can not. Also, you can display the same column on both sides. For example, if you are changing a verbatim, you could select both columns to display verbatim. The left would show the original verbatim, and the right would show the changed verbatim. Controls and Options for the Edit Respondent Window The banner bar has these options: Move to the first respondent. Move to the previous respondent. Also available from the right-click menu. The Page Up key is a shortcut for this button. Move to the next respondent. Also available from the right-click menu. The Page Down key is a shortcut for this button. Move to the last respondent. The respondent ID. To move to a specific respondent, enter the respondent ID and press Enter. This box displays in red if there is no respondent with this ID, or if there is no available respondent in the direction selected by one of the buttons above. Note that in most cases, respondent IDs in Ascribe™ will have any leading zeros stripped. So a respondent ID of "001" will be in the Ascribe™ database as "1" and will not match "001". 147 PrintDoc01 020813 Click this button to display the filter dialog. In this dialog, you can select the respondents to display by their responses to a set of questions. The filter button is yellow when a filter is applied. Check spelling Check this box to enable the spell checker. The spell checker is available only if you have Microsoft Word installed on your computer. Hide N/A Check this box to hide any questions to which the respondent has no response. When this box is not checked, a question for which there is no response is displayed with a grey background. Row height When this selection is blank, the rows change height to accommodate the size of the response. You can select a fixed number of rows to display with one of the options in this list. These controls are available at the top of the left text columns: Coding Status Indictor The cell next to the question ID displays blue if the response is coded. It displays red if the response is locked because it is being coded, transcribed, or translated at the time you displayed this respondent. If cell has no color, the response has not been coded. This menu selects the portion of the response to display in this column. Verbatim and Notes are always available. Transcription and Translation are available if licensed. The center menu selects the reviewed status of the responses. If you select "All reviewed", only those respondents that have all responses reviewed will be displayed. This menu selects the blank status of the responses. If you select "None blank", only those respondents who have text in all responses will be displayed. There is not a requirement that the respondent have a response to all questions, but for each question with a response text must be present. This option has no effect when Verbatims are displayed (a response always has a verbatim, and it will be displayed even if it is blank). These controls are available at the top of the right column: This menu selects the portion of the response to display in this column. Verbatim and Notes are always available. Transcription and Translation are available if licensed. If you select Transcription, only those questions with the transcription task checked on the question task tab will display. Similarly, if you select Translation, only those questions with the translation task checked will display. 148 Client Access to Data The center menu down selects the reviewed status of the responses. If you select "Some not reviewed," only those respondents that have at least one response that is not reviewed will be displayed. This menu selects the blank status of the responses. If you select "Some blank", only those respondents who have text missing in at least one response will be displayed. When used with Verbatims, this option will cause no responses to be displayed (a response always has a verbatim; it can never be missing). Determine the Set of Questions Displayed The set of questions displayed (the rows in the table) is determined by two factors: The setting of the Hide N/A check box. If this box is checked, only those questions for which there is a response will be displayed. The properties of the question (selected on the Task Tab.) The Task Tab controls whether transcription or translation can be done for the question. When one of these is selected, only those questions with the corresponding transcribe or translate setting checked will be displayed Move to Next Respondent The set of respondents you can move to is determined by the combination of filters active on the page. These are: The codes applied filter set using the Filter... dialog The reviewed status filters set using the center menu in each text column The blanks filters set using the right menu in each column. If no filters are active, the set of respondents available is all respondents in the study. Otherwise, the set of respondents available is restricted to the set selected by the combination of filters. When you use one of the navigation buttons in the banner bar, the next available respondent is displayed. If there is no available respondent, the respondent ID box shows red. You can move to a specific respondent if you enter a respondent ID in the banner bar box, and press Enter. Again, that respondent will display only if available according to the active filters. Translate, Transcribe, or Edit a Verbatim You can translate, transcribe, or edit a verbatim on this page. If you click in the right column, a cursor appears, and you can type the desired text. The Enter key moves the cursor downward to the next response. To enter a line break in the text, press the Shift-Enter key. In this window, you can drag and drop text from one response to another. 149 PrintDoc01 020813 You can edit transcriptions, translations, and notes for a response only if the respondent has an existing verbatim response to that question. If you have supervisor privilege, you can edit the verbatim for a response even if the respondent has no existing response to that question. When you select Verbatims in the right column, a warning message displays. The message alerts you that when you edit verbatims, you have the ability to add responses. When you edit verbatims, the right column is editable even if the respondent has no existing response to the question. If you add text to a response where the leftmost column is grey (no existing response), a response is added to the question that respondent. Note If you delete all of the text for a verbatim, the response is not deleted. The response remains as a blank response. To delete the response, select Delete Response from the right-click menu. You can not edit verbatims that do not contain text. For example, you can not edit images or audio responses. However, you can delete those responses. Delete a Response Supervisor privilege is required to delete a response. Right-click a response and select Delete Response. Spell Check in the Edit Respondent Window If you have Microsoft Word installed on your computer, you can check spelling in this window. To enable the spell checker, select the check spelling box at the top of the window. When the spell checker is enabled, Ascribe™ checks the spelling of verbatims or transcriptions when you click on them or when you exit a cell. You can also request a spell check for the current cell by pressing the Esc key. Spell check is only performed for the right column and not the left column. Misspelled words are highlighted and underlined. You can correct the spelling if you type the box, or if you right-click the highlighted word and pick a word from the menu. (If the menu is empty, it means that Word could not find a suitable suggestion). The spell checker "learns" as it works, and will operate faster as you continue to work in this page. When you leave the page, the spell checker "forgets" the words it has learned, and will start slowly again the next time you enter the page. Because Ascribe™ uses Microsoft Word to check spelling, you can check spelling in multiple languages if you set up Word for the desired language. Ascribe™ will use the spelling dictionary that you have selected in Word. Ascribe™ also uses the Microsoft Word autocorrect feature, using the auto correct list maintained in Word. If you do not wish to use this feature, disable that option in Word. Mark as Reviewed Check the box to the left of a text area to mark it as reviewed. To set all responses in the left or right column as reviewed, right-click, select reviewed status, and select set all in left 150 Client Access to Data column or set all in right column. You can clear the reviewed status if you right-click, select reviewed status, and select clear all in left column or clear all in right column. Multi-User Considerations This page is not fully multi-user. It does not lock responses, so it is possible for two users to attempt to modify a response at the same time. However, this page does detect such conflicts. It will post an alert if you attempt to modify a response that was changed since it was displayed on this page. When a conflict is detected, your changes are discarded. It is safe for multiple users to edit responses simultaneously, but they must be prepared for occasional conflicts and possible loss of edits to a response. Edit Responses Window Navigate: Coder/Studies in Progress/Right-click study/Questions in Progress/Right-click question/Edit Responses Transcriber/Transcribe/Studies to Transcribe/Right-click study/Questions to Transcribe/Right-click question/Edit Responses Transcriber/Translate/Studies to Translate/Right-click study/Questions to Translate/Right-click question/Edit Responses On this page you can view and edit all of the responses to a question. All Edit Responses functions can take place before, after, or during coding. The Edit Responses page is designed to allow you to enter translations or transcriptions quickly for many responses. Controls and Options for the Edit Responses Window The Edit Responses Window has two columns and a number of controls. The text in the left column can not be edited. The text in the right column can be edited, except for multimedia verbatims. The text which displays in the columns are determined by the controls in the banner bar and above the left column. The banner bar has these options: Allows you to navigate through the responses. When there are more than 500 responses, the page does not display all of the responses. Use this field to select the range of responses to display. This field also works with the sort order field. For example, when you select 1-500 in this field, the first 500 responses display as determined by the current sort order. Click this button to display the filter dialog. In this dialog, you can select the respondents to display by their responses to a set of questions. The filter button is yellow when a filter is applied Check this box to enable the spell checker. The spell checker 151 PrintDoc01 020813 is available only if you have Microsoft Word installed on your computer. You can change the sort order of the responses with this field. The sort applies to the entire set of responses, not just those displayed. Here are the options: Respondent ID – Sort by the respondent ID. If your respondent IDs are numeric, they will sort in numeric order only if all respondent IDs are the same length. For example, 10 will sort above 2, and 10 will sort below 02. Left text – Sort by the response text in the left column. Only the first 100 characters of the response are used in the sort. Right text – Sort by the response text in the right column. Only the first 100 characters of the responses are used in the sort. Coded – Sort by the coding status, which is indicated by the colored bar to the right of the respondent ID. Left reviewed – Sort by reviewed check box for the responses in the left column. Right reviewed – Sort by the reviewed check box for the responses in the right column. The Descending check box reverses the sort order. When you change translations, transcriptions, or notes for a response, you can have that change applied to other duplicate responses. For this option, a duplicate response is one where the text in the left column is the same as the response you are changing. Responses where the text in the left column is blank are never matched. If more than one response is affected when you change the text, Ascribe™ displays the number of additional responses that were changed, and asks if you want the page to be refreshed with the changes. The options are: None – The change is not applied to duplicates. Only the response you change is affected. Blank – The change is applied to all duplicates that have no text in the right column. Not Blank – the change is applied to all duplicates that have some text in the right column. All – The change is applied to all duplicates. These controls are available at the top of the left text and right text columns: Coding Status Indictor 152 The cell next to the question ID displays blue if the response is coded. It displays red if the response is locked because it is being coded, transcribed, or translated at the time you displayed this respondent. If Client Access to Data cell has no color, the response has not been coded. This menu selects the portion of the response to display in this column. Verbatim and Notes are always available. Transcription and translation are available if licensed. The center menu selects the reviewed status of the responses. These are the options: Blank – All responses will be displayed. Reviewed - Only those responses that have the reviewed box checked will be displayed. Not Reviewed – Only those responses which are not reviewed will be displayed. This menu selects whether you display all responses or only those with the column blank or not blank. This option is only in the left text column. If this box is selected, the codes applied to the responses display below the text for each response. Codes are not displayed for multi-media verbatims. Codes are displayed for the transcriptions, translations, and notes for multi-media verbatims. Suggested Settings for Translation View Pane – Set to Verbatim to display the native language verbatim. Edit Pane – Set to Translation to allow translation to be entered. Filter – Set your language or country filter before you start. See Filter for more information. Check Spelling – Toggle this option on to check spelling. Apply to duplicates – Set to either All or Blank. For more information, see Apply to Duplicates. Review Filter – Leave this filter empty unless you need to see reviewed translations. Blank Filter – Set to Blank to view those respondents without a translation. Translate, Transcribe, or Edit a Verbatim You can translate, transcribe, or edit a verbatim on this page. If you click the right text column, a cursor appears. Enter text or modify the verbatim. To move to the next response, click another response, press the Enter key, or press the Tab key. The Ascribe™ database is updated immediately when you move the cursor from the text you are editing. 153 PrintDoc01 020813 Fix a Wrongly Reclassified Verbatim To fix a wrongly reclassified verbatim, first click the verbatim response in the right pane. Then you can highlight and delete the part of the response that has been reclassified from another question. Add a Response If you have supervisor privilege, you can add responses to the question. Right-click and select add response. You can enter in the desired respondent ID and response in the Add Response dialog. 154 Client Access to Data To enter in multiple responses with incremental ID’s, select the Increment Respondent ID option. The last Respondent ID will be replaced with the next highest RID. Also, the response field will be empty, and text can be entered. For example: if the increment response ID option was selected in the above screen, and you add a response, the next respondent ID field would contain 10003. Delete a Response Supervisor privilege is required to delete a response. Right-click a response and select Delete Response. Mark as Reviewed Check the box to the left of a text area to mark it as reviewed. To set all responses in the left or right column as reviewed, right-click, select reviewed status, and select set all in left column or set all in right column. You can clear the reviewed status if you right-click, select reviewed status, and select clear all in left column or clear all in right column. Spell Check in Edit Responses If you have Microsoft Word installed on your computer, you can check spelling in this window. To enable the spell checker, select the check spelling box at the top of the window. When check spelling is enabled, Ascribe™ checks the spelling of responses when you click on the text in the right column, or when you press the escape ( ESC) key. Spell check is performed only on the text in the right column, not in the left column. You can check the spelling of all responses if you right-click and select spell check all. Misspelled words are highlighted in red and underlined. Just right-click those words and a list of suggested, correct spellings display. You can choose a corrected spelling from the list or type the correction yourself. If you have “apply to duplicates" set to all, then duplicate responses in translations, transcriptions, and notes will be fixed also. You will find that the spell checker slows down the operation of the window. If you are not using the spell checker, you can turn it off to improve performance. Because Ascribe™ uses Microsoft Word to check spelling, you can check spelling in multiple languages by setting up Word for the desired language. Ascribe™ will use the spelling dictionary that you have selected in Word. If you change your dictionary options in Word, the changes may not appear in Ascribe™ until you logoff, close your browser window, and log back on. To spell check individual words, press the ESC key. 155 PrintDoc01 020813 Multi-Media Responses Media responses cannot be changed in this page and can not be modified even when they are displayed in the right column. When working with images, there is a difference between the way the Tab and Enter keys work when moving from one row to the next in the right column. If you press the Enter key and there is an image in the row you are moving to, that image is expanded and the current image is collapsed. If you use the Tab key, the cursor will move to the next response without changing the image display. You can use the audio control keys in this window to control the playback of the audio response for the current row. For example, the CTRL-B key restarts the playback of the audio. Translate/Transcribe Verbatims Ascribe™ provides several methods to translate or transcribe verbatims. Translate or transcribe at the study level – This method uses the Translate/Transcribe Window. It displays one response at a time in respondent ID order. If a respondent appears in more than one question, you will see each answer in order of the question ID. Translate or transcribe at the question level – This method uses the Translate/Transcribe Window. It displays one response at a time in respondent ID order for the question you choose. Edit respondent – This method uses the Edit Respondent Window. It allows you to translate or transcribe responses at the study level. Ascribe™ displays on one screen a respondent’s answers to all questions marked for translation. Edit responses – This method uses the Edit Responses Window. It allows you to translate or transcribe responses to a single question. All of the responses display in order by respondent ID. The Studies to Transcribe/Translate page lists studies available for these tasks. A study will appear in this list if both of the following are true: The Transcribe or Translate check box (depending on whether you are in Studies to Transcribe or Studies to Translate) is set in at least one question in the study The status of the study is In progress. Questions to Transcribe/Translate Navigate: Transcriber/Transcribe/Studies to Transcribe/Right-click a study/Select Questions to Transcribe Transcriber/Translate/Studies to Translate/Right-click a study/Select Questions to Translate 156 Client Access to Data This page lists the questions in the study to be transcribed or translated. A question will appear in this list only if the Transcribe or Translate task is checked in the Question Properties dialog. To Transcribe/Translate a Question - Right-click on the question and select Transcribe or Translate from the popup menu. To Edit the Responses to a Question - Right-click on the question and select Edit Responses from the popup menu. Considerations for Translation Using Ascribe™ Make sure the translators have transcriber privileges. Associates with transcriber privileges can see all studies marked for translation. Contacts with transcriber privileges can only see studies that have their company as the end customer or client. (See Add/Edit an Associate or Add/Edit a Company Contact.) Make sure the questions are marked for translation. You must mark a question for translation or transcription in order for it to display in the list of questions available for translation or transcription. (See Tasks Tab.) Decide whether the coding source should be set to translation or not. When you set the coding source to translation, it affects the Coder Window, Review Responses, and Phrase Analyzer. (See Coding Tab.) Make sure the study status is set to coding in progress. Only studies marked as "coding in progress" will appear in the list of studies to be translated or transcribed. (See Study Status on the General Tab.) Translate/Transcribe Window Navigate: Transcriber/Transcribe/Right-click study/Select Transcribe Transcriber/Translate/Right-click study/Select Translate Transcriber/Transcribe/Right-click study/Select Questions/Right-click question/Select Transcribe Transcriber/Translate/Right-click study/Select Questions/Right-click question/Select Translate This page is designed for transcription and translation of responses. Responses are presented one at a time. You simply type the desired text and press Enter to move to the next response. If you navigate to this page through Transcribe, the editable portion of the page is the transcription. Similarly, if you enter this page through Translate, the editable portion is the translation. If you navigate to this page from a list of studies, you transcribe or translate the entire study. Responses are presented by respondent. Each response for that respondent is presented before you move to the next respondent. If you navigate to this page from a list of questions, you transcribe or translate only the responses for that question. This page supports multi-user access, and more than one person may be work on the same study or question. A given response will never be presented to two people in this page at the same time. When you press Enter in the editable transcription or translation pane, your changes are saved, and the next available response displays. The response selected will be the next one with no text in the editable area. 157 PrintDoc01 020813 The right-click menu has these options: Questions The option will display Questions or Studies, depending on how you navigated to the page. If you are viewing the responses by question, the option will be Questions. If you are viewing all responses in the study, the option will be Studies. or Studies Settings This option opens a dialog with settings for the page. Keyboard Shortcuts This options opens a dialog with the keyboard shortcuts for the page. For more information, see: Controls in the Banner Bar Controls in the Source Pane Controls in the Translation Pane Counts of Responses Add or Edit a Translation Auto Translate Translate/Transcribe Settings Choose the following settings for the transcription/translation page: Show Notes Check this box to display and edit the notes for each response. Show Question Check this box to display the question ID and text at the top of the Transcribe/Translate page. Split Vertically Check this box to split the window vertically. Magnification Enter a numeric magnification percentage. This setting is used for image verbatims only. The value 100 displays the image at normal size. Use a value less than 100 to shrink the image, and a value greater than 100 to expand the image. Controls in the Banner Bar The bar above the source pane has these options: Move to the first response. Move to the previous response. Move to the next response. Move to the last response. 158 Client Access to Data Displays the respondent ID. To move to a specific respondent, enter the respondent ID and press Enter. Displays a red background if a respondent is not available. Note that in most cases, respondent IDs in Ascribe™ will have any leading zeros stripped. So a respondent ID of "001" will be in the Ascribe™ database as "1" and will not match "001". Move to the first response with a blank transcription or translation. Move to the next response with a blank transcription or translation. If you press Enter in the editable transcription or translation box, it is the same as if you click this button. Select a filter for responses based on codes applied. This option determines whether the changes you make in the editable transcription/translation pane are applied to duplicate responses in the non-editable pane. See Apply to Duplicates. Check this box to enable the spell checker. When the spell checker is enabled, you can use your keyboard shortcut to check what you have typed. You can also click anywhere on the page after you have typed a translation/transcription. Spell check uses Microsoft Word’s dictionary. You must have Word installed on your PC to use the spell check. Controls in the Source Pane The source pane displays what you transcribe or translate. The top of that pane has these controls: Verbatim Translation or This control is displayed only if you have licensed both transcriptions and translations. It allows you to select the other available source. For example, when you translate, you can select either the verbatim or the transcription as your source. Transcription Reviewed only Check this box to display only source items that have the reviewed check box set (you can not set that box in this page; it is set in other pages such as Edit Respondent or Edit Responses). When you change this option, it is in effect when you move to the next response. Uncoded only Check this box to display only uncoded responses. When you change this option, it is in effect when you move to the next response. Controls in the Translation Pane The translation pane has these controls: 159 PrintDoc01 020813 This field works in conjunction with the Auto Translate feature, but also controls whether the Source Language field is able to be changed. If Auto Detect Source Language is checked and you use the Auto Translate feature, the auto translator will determine the source language of the response. Auto Detect Source Language If Auto Detect Source Language is checked, you are not able to change the Source Language drop down. In order to change the Source Language field, unselect Auto Detect Source Language. If you unselect Auto Detect Source Language and the incorrect language is set in the Source Language field, the Auto Translate feature may not work correctly. You must select a Source Language and a Target Language if you do not use the Auto Translate feature. This field is used to display the source language of the response. You can change the language if Auto Detect Source Language is unchecked. Source Language Drop Down Target Language Drop Down If you do not use the Auto Translate feature, you must select a Source Language and a Target Language. The Source Language is used in conjunction with the translation library on each account. The translation library stores an entry for each verbatim text and its corresponding translation, Source Language, and Target Language. If the exact verbatim appears again in a question or study, a pop-up window displays the translation from the library, and you can decide to use that suggestion or enter a different translation. This field is used to display the language of the translation. You must select the Target Language before anything can be translated. The selection will stick from session to session, so you may only need to select it one time if you always translate to the same language. This feature allows you to use 'statistical machine translation,' which looks for patterns in hundreds of millions of documents to help decide on the best translation. These languages are supported: Auto Translate 160 Afrikaans, Albanian, Arabic, Belarusian, Bulgarian, Catalan, Chinese simplified or traditional, Croatian, Czech, Danish, Dutch, English, Estonian, Esperanto, Filipino, Finnish, French, Galician, German, Greek, Hebrew, Hindi, Hungarian, Icelandic, Indonesian, Irish, Italian, Japanese, Korean, Latvian, Lithuanian, Macedonian, Malay, Maltese, Norwegian, Persian, Polish, Portuguese, Romanian, Russian, Serbian, Slovak, Slovenian, Spanish, Swahili, Swedish, Thai, Turkish, Ukrainian, Vietnamese, Welsh, and Yiddish. Client Access to Data Note that transactions may apply in addition to the normal translation transaction schedule. Contact support@goascribe.com if you have questions. Audio Keyboard Shortcuts The following default key combinations can be used to control the media player. You can change the keyboard key if you click the text box of the shortcut and enter a different key. Note that not all browsers or operating system will support all key combinations. Activity Keyboard Shortcut Start/Pause Ctrl + S Start Over Ctrl + B Advance 30% Ctrl + Right Arrow Rewind 30% Ctrl + Left Arrow Volume Up Ctrl + Up Arrow Volume Down Ctrl + Down Arrow Auto Translate Ctrl + T Next Response Ctrl + Enter Next ID/Respondent Ctrl + Page Up Previous ID/Respondent Ctrl + Page Down Check Spelling Ctrl + R Counts of Responses In the banner above the translation or transcription, the counts of work completed are displayed. There are two sets of counts. The first is sensitive to any filter settings you have selected. The second is not. For example, suppose that you have a filter selected in the Filter... dialog, and also have the Reviewed only box checked. You might see this in the banner: Transcription 34/220 (55/500 Total) This means "You have completed 34 of 220 responses that are selected by your filter and have the reviewed box checked. Of all responses to be transcribed, 55 of 500 are completed." Apply to Duplicates This option in the banner bar determines whether changes made in the editable transcription/translation pane are applied to duplicate responses in the non-editable pane. The changes apply only to the transcription or translation. If you change the notes for a response, the change for the notes is always applied only to that response, regardless of this setting. 161 PrintDoc01 020813 If a response is being edited by another user in this window, your changes will not be applied to that response, even if it matches. The available options are: None - The changes you make will apply only to the current response. Blank - The changes you make will apply to the current response, and also to any blank responses in the editable pane, for all responses that match the text in the non-editable pane. Not blank - The changes you make will apply to the current response, and also to any non-blank responses in the editable pane, for all responses that match the text in the non-editable pane. All - The changes you make will apply to the current response and to all responses that match the text in the non-editable pane. Add or Edit a Translation To add or edit a translation, click the translation pane. A cursor appears. Enter or change the translation. To move to the next translation, press Enter or use the controls in the banner bar. As you are translating responses, you are building a library of translations for the account on which you are working. The translation library stores an entry for each verbatim text and its corresponding translation, Source Language, and Target Language. If the exact verbatim appears again in a question or study, a pop-up window displays the translation from the library, and you can decide to use that suggestion or enter a different translation. Note that it is best practice to have the Source and Target Languages set correctly in order to make the most of the translation library. The translation library cannot be edited by users - it is simply there to make suggestions based on previous translations. See Auto Translate for information about the Auto Translate feature. Auto Translate This feature allows you to use 'statistical machine translation,' which looks for patterns in hundreds of millions of documents to help decide on the best translation. These languages are supported: Afrikaans, Albanian, Arabic, Belarusian, Bulgarian, Catalan, Chinese simplified or traditional, Croatian, Czech, Danish, Dutch, English, Estonian, Esperanto, Filipino, Finnish, French, Galician, German, Greek, Hebrew, Hindi, Hungarian, Icelandic, Indonesian, Irish, Italian, Japanese, Korean, Latvian, Lithuanian, Macedonian, Malay, Maltese, Norwegian, Persian, Polish, Portuguese, Romanian, Russian, Serbian, Slovak, Slovenian, Spanish, Swahili, Swedish, Thai, Turkish, Ukrainian, Vietnamese, Welsh, and Yiddish. Note that transactions may apply in addition to the normal translation transaction schedule. Contact support@goascribe.com if you have questions. The Transactions reports display the number of transactions incurred for auto translations. When you click the Auto Translate button, the translated response displays in the Translation pane. You may edit or tweak the translation as necessary. 162 Client Access to Data If Auto Detect Source Language is checked and you use the Auto Translate feature, the auto translator will determine the source language of the response. If you unselect Auto Detect Source Language and the incorrect language is set in the Source Language field, the Auto Translate feature may not work correctly. You may receive a message indicating that no translation can be provided from the auto translator under these circumstances: the source response exceeds approximately 5000 characters (the response is exceedingly long) the source response is the same word in source and target languages the source or target language is not properly set. Filter - Set a Response Filter Navigate: Coder Window/Right-click/Filter Review Responses Window/Click the Response Filter Button (bottom left) Edit Responses or Edit Respondent Window/Click the Filter button (top right) Translation or Transcription Window/Click the Filter button (top right) You can set a filter to view only those responses that meet your criteria. For example, you may want to see just those respondents that answered "Yes" to a previous question. Or, if you have to translate respondents, you may want to filter by language. In order to use a filter, you must first code the question that acts as the filter. In most cases, this is a closed-end question. If you wish for your filter to stay from session to session, you can navigate to the home page and set your user options to keep filters between sessions. Filter Dialog You can set filters from the Coder Window, the Review Responses Window, the Edit Responses Window, the Edit Respondent Window, the Translation Window, and the Transcription Window. After you select to filter, the filter dialog displays. In the filter dialog, you select a question which acts as the filter. You further define the filter when you select a code or codes from that question’s codebook. Only respondents who have codes that match those selected as the filter will be displayed. The filter dialog has two panes at the top of the page. The left pane is the study questions pane and displays all of the questions in the study. The right pane is the filtering questions pane and is initially blank. It will display the question you select as the filter. To select a question to act as a filter: 1. Highlight the question in the study questions pane and click the “>” button or just double-click the question. After you select the question, it displays in the filtering questions pane. 2. Click the question in the filtering questions pane. The question’s codebook displays in the bottom window or codebook pane. 3. After you select enough codes, click one of the Apply Codes buttons. The codes are applied, and the next verbatim displays. To choose the code or codes you 163 PrintDoc01 020813 want to use as a filter, click the box next to the code. If you select multiple codes, Ascribe™ selects respondents who have any of the codes. 4. Click OK to return to the responses. Only respondents who have codes that match the filter criteria will be available for display. Multiple questions can be used as a filter if you repeat the process. The And/Or buttons in the center of the page determine the logical conditions for inclusion when multiple questions are chosen. Choose "Apply to all Questions" to have the filter stick for all questions in your study. Other options for this window include: “>>” button – this button selects all questions in the study questions panes and moves them to the filtering questions pane. “<” button – this button removes the highlighted question from the filtering questions pane (the question must be highlighted first.) You can also remove a question if you double-click it. “<<” button – this button removes all questions from the filtering questions pane. There is no limit to the number of questions to filter by, but the filter process will slow down as you add more and more questions. How to Tell if a Filter is Set Each window has an indicator to let you know a filter is set for the respondents. Here is a list of the indicators: 164 In the Coder Window, the forward and backward buttons are highlighted in yellow. Client Access to Data In the Review Responses Window, the response pane has the words “Question Filters: question ID: question label” In the Edit Respondent Window, the Edit Responses Window, the Translation Window, and the Transcription Window, the Filter button is highlighted yellow. 165 PrintDoc01 020813 Client Access to Data Navigate: Client The client section of Ascribe™ provides controlled access to each study and may be accessed by project directors, clients, tabulators, account executives, etc. The Client Studies page lists the studies available to you. The studies listed depend on the type of user account you for Ascribe™. If you have an associate account, you can see all of the studies except those that are archived. If you have a contact account, you can see only those studies for your company. Client Reports Navigate: Supervisor/Studies/Right-click study/Click Reports All items in the client results section display in "real time." Information is pulled directly from the database and includes all data/information in the database at the time the report was generated. Note Real-time is a much misused term. Here it means "you are getting the results as they are stored in the database." When you navigate to the Studies page, the list of studies is displayed. You can filter the list by the study status. In the grey bar at the top of the window, check the status boxes for the studies you want to view, and then click the Update button. You also have the option to change the properties of the study if you right-click the study and select Edit. Note Language Logic can optionally configure your site to allow contacts to view only completed studies. If this option is in effect, a contact will not see the grey bar to select study status. Coding Overview Report Navigate: Client/Right-click study/Click Coding Overview The Coding Overview Report lists all studies that are currently on the system except for those marked as archived. This report allows you to quickly gauge the level of completion for all live studies. The studies listed depend on the type of user account you have for Ascribe™. 166 Client Access to Data If you have an associate account, you can see all of the studies except those that are archived. If you have a contact account, you can see only those studies for your company. When you reach the Coding Overview report, you can filter the list of studies by the study status. In the grey bar at the top of the window, check the status boxes for the studies you want to view, and then click the Update button. This information displays at the beginning of the Coding Overview Report: Field Description Created The date the study was setup on Ascribe™. This date is based on the web server clock. Study ID Study identification number Study Study name Hours The number of hours recorded for this study in the Time Accounting records (only appears if you have administrative privileges) This information displays for all questions and then for each type of question individually in this order: All Questions, All Open Questions, All Other Specify Question, All Closed Questions, All Value Questions. Field Description Respondents Number of individual respondents in the study, in all open questions, in all other specify questions, etc. Responses Number of responses in the study, in all open questions, in all other specify questions, etc. Codes Applied Number of codes applied to all questions in the study, for all open questions, for all other specify questions, etc. Responses Coded Number of responses coded in all questions of the study, in all open questions, in all other specify questions, etc. % Percentage of responses coded in all questions, in all open questions, in all other specify questions, etc. This information displays at the end of the report: Field Description Client The company that requested the study End Client The end customer for the study Status The coding status of the study Study Summary Report Navigate: Supervisor/Studies/Right-click study/Select Reports/Click Study Summary 167 PrintDoc01 020813 The Study Summary Report displays two tables. Summary information about the study is at the top of the page. The second table lists coding information by question type. Here is the summary information on the table at the top of the page: Field Description Study Name The study name Study ID The study identification number Created The date the study was created Start Date The date work on the study was started Description The description of the study from the Study Edit screen Help The information about the study from the Study Edit screen Supervisor The supervisor of the study Client The company that requested the study End Customer The end customer of the study Status The coding status of the study from the Study Edit screen Coders The number of coders who worked on this study Loads The number of successful loads for this study Transactions The number of transactions used on this study; only administrators can view this information Hours The number of hours spent on this study; only administrators can view this information Here is the information on the table for question types: Field Description Type Each question type (open, closed, other specify) has a row in the table Questions Number of questions for this question type Respondents Total number of respondents for each question type Responses Total number of responses for each question type Responses Coded Total number of responses coded for each question type Codes Applied Number of codes applied for each question type 168 Client Access to Data Question Summary Report Navigate: Client/Right-click study/Select Reports/Click Question Summary The Question Summary Report lists summary information about each question. You can filter the list of questions by question type. In the grey bar at the top of the window, check the boxes for the question types you want to view, and then click the Update button. The report also has additional options for showing codebooks and verbatims. Here is the information on the Question Summary Report: Field Description Question ID The question identification number Question Label The question label Question Type The question type Layout The data layout Respondents The number of respondents for this question; the total of all respondents is displayed at the bottom of the table. Responses Coded The number of responses coded for this question, followed by the number of responses coded as a percentage of the number of responses; the total number of responses coded and its percentage display at the bottom of the table. Hours The number of hours recorded in time accounting for this question; only administrators can view this field. The total number of hours displays at the bottom of the table. Responses Per Hour The average number of responses per hour is determined by the number of responses coded divided by the number of hours worked on this question; only administrators can view this field. The average responses coded per hour displays at the bottom of the 169 PrintDoc01 020813 table. Question Detail Report within the Question Summary Report Navigate: Client/Right-click study/Select Reports/Click Question Summary The Question Detail Report is used to quality check a study, approve codebooks or just check on the status of a study. You can select any number of questions to see more information. From the Question Summary Report, click the box next to the question ID of one or more questions. Right-click the question to display these options: Option Description Detail This Question Displays a detail report of the selected question. From this option, you also can display a detail report for questions that share this question. Detail Selected Questions Displays detail reports for multiple selected questions Edit Question Displays the Edit Question dialog Select All Selects all of the questions and puts a checkmark in each box Clear All Removes the checkmark from each box Show Responses For Displays all of the responses for one question This Question Download Data Navigates to Codebooks and Results or Scripted Output Reports Navigates to other reports Export Table To Excel™ Copies the report to Excel™ Go To Navigates to other parts of Ascribe™ 170 Client Access to Data The Question Detail Report has general information about the question at the top of the page. This information includes question type, number of responses loaded, number of current responses, number of responses coded, number of responses currently referred, number of codes applied, card number, column number, number of columns and data layout. The codebook displays below the question information. The codebook information has these fields: Field Description Row The row number for this code or net Input The input ID for the code. This value is not displayed for nets. Output The output ID for the code. This value is not displayed for nets. Description The description of the codes. Nets display in bold. Number of Respondents The top of the column shows the total number of respondents in the question. For nets, this value is the unduplicated count of respondents who had any code of this net applied. For codes, this value is the number of responses to which this code was applied. Percentage Of Respondents This column displays the percentage of respondents who had this code applied. If the codebook is shared and if you select Detail this questions/with questions sharing codebook, a column is added for each question that shares the codebook within the study. Codebooks that are shared with questions in different studies are not displayed. You can also display responses for this question from the Question Detail Report. To view responses for a code, right-click the code and select Show responses for this code. To view all of the responses for a question, right-click anywhere in the table and select Show responses for this question. 171 PrintDoc01 020813 Card Layout Report Navigate: Client/Right-click study/Select Reports/Click Card Layout The Card Layout Report finds errors in your study set-up before you deliver data. It displays any problems detected in card/column/columns set-up and in numbering. The report is divided into two panes. The upper pane displays the questions in the study and is updated automatically if you change the properties of the study or the questions. The lower pane displays a picture of the data output in cards and columns. The lower pane is updated only if you select Update Card View in the upper pane and click the Update button. The question list in the upper pane has these fields: Field Description Question ID The question ID Question Label The question label Type The question type Card The card number where the data will be placed in column binary output. Layout Numeric, column punch, and punch using column offset are the three data layout types. Maximum Codes The maximum number of codes allowed for this question as defined on the Data tab of the Edit Question dialog. (0 = unlimited number of codes.) Max Codes Applied The maximum number of codes that actually have been applied to 172 Client Access to Data any one response. Column In numeric layout, codes for this question will start in this column. Punch using column offset layout uses the column field in a calculation. See Punch Using Column Offset for more information. Punch layout does not use this field. Columns Per Code The number of columns used to write the code which is also the number of digits in the output code value. This field only applies to the numeric layout type. Columns Used The maximum number of columns used for the output of codes for this question, based on the codes currently applied to this question. Applies only to the numeric layout type. The lower pane displays a diagram of the cards and columns to be output. The columns that are used display across the top of the card. The report has these fields: Field Description C Card number R Respondent ID Question ID The report displays the columns used by the questions. Problems A list of any problems detected by analysis of the study and question settings, the codebooks, and the codes currently applied in the study. This message displays if there are no errors - Card field : Specified column 0 starts before the first column of the card (1). Below is an example of the Static Hair Study set-up to be output numerically. Study Report Navigate: Client/Right-click study/Select Reports/Click Study Report The study report provides a real-time summary of the coding results for each question. 173 PrintDoc01 020813 Note Real-time is a much misused term. Here it means "you are getting the results as they are stored in the database." Typically this report is used by people who want to monitor progress, assess and approve codebooks, check for quality, and perform audits. It has two basic components: a study summary and a question summary for each question in the study. Study Summary in the Study Report The study summary provides this information: Field Description Study name The study name Study ID The study identifier Status The study status Created The date the study was created Due Date The due date of the study Responses The total number of responses in all of the questions (includes open ended, closed ended and other specify questions) Respondents The total number of respondents in all of the questions Responses Coded The total number of responses coded in this study Codes Applied The total number of codes applied for all questions Open Questions The number of open-ended questions in the study Closed Questions The number of closed-ended questions in the study Other Specify Questions The number of other specify questions in the study Value Questions The number of value questions in the study A table displays below the study summary information. It lists question type, number of responses and number of codes applied. This table can put copied to Excel™ if you click the Excel™ icon ( ). 174 Client Access to Data Question Summary within the Study Report Each question in the study is listed with summary and detail information. You can click the icon to view the verbatims and codes applied for the question. If you have supervisor privilege, you can specify responses to recode in this window. The question summary report has these fields: Field Description Question Type The question type (open, closed, other specify or value) Responses Loaded The total number of responses loaded for this question Responses Current The number of responses currently stored in Ascribe™. This number may differ from the number of responses loaded, because responses may have been reclassified. Responses Coded This field displays the number of responses that have been coded, the number of total responses loaded, and the percentage of responses that have been coded. Responses Referred The number of responses referred to the supervisor, but which have not gotten a reply yet Codes Applied The number of codes that have been applied to all responses in this question. Card The card where this question’s data will be written Column The starting column where the data (codes) will be written to upon output Columns The maximum width, in columns, for each code. It will be zero filled on output. Open Questions The number of open-ended questions in the study 175 PrintDoc01 020813 Closed Questions The number of closed-ended questions in the study Other Specify Questions The number of other specify questions in the study Value Questions The number of value questions in the study A table displays below the question summary information. This table can put copied to Excel™ if you click the Excel™ icon ( ). The table has this information: Field Description Row Number This column lets you resort the table in its original order if you have sorted by a column. Input Code Value The input value of the code Output Code Value The output value of the code Code Description The description of the code. If any responses have been coded with this code, you can click the description to see the verbatims and the codes applied. Codes Applied The number of times this code has been applied for this question Percent The number of times this code has been applied divided by the number of responses to this question. It displays as a percentage. Study Quality Report Navigate: Client/Right-click study/Select Reports/Click Quality The Study Quality Report typically is used for training. It shows the difference between codes applied to a study and the saved set of 'quality codes' for the study. The intention of the quality report is to allow you to store a set of 'known good' codes for a study, and then compare a coder's work against this set. 176 Client Access to Data Quality coding refers to a process by which an expert codes a study "perfectly," or as close to the way it should be. After the study has been coded, it is converted to quality mode, and a trainee codes it. The trainee's coding is then compared to and measured against the expert’s coding. A report is created that shows the differences in coding. Set Quality Codes Navigate: Supervisor/Studies/Right-click study/Select Set Quality Codes To create quality codes, first you must have coded (or partially coded) a study. Then navigate to the Set Quality Codes screen. Select each question or group of questions for which you want to set quality codes and click OK. When you set quality codes for a question, any existing quality codes are deleted and replaced with the current set of codes for the question. Questions you have not selected in the set quality codes window are not affected when you click the OK button. You can set quality codes only for open or other specify questions. After you set quality codes, you should delete the current coding. Navigate to the Questions screen, right-click any question, and choose Delete Questions. Select the questions for which quality codes have been set, and then choose "Codes" from the right-hand deletion column. See Delete Questions for more details. Now the study is ready for the trainee to proceed with coding. A few things to note: Any notes, translations, or transcriptions the expert coder added continue to display for the trainee unless you use Edit Responses to get rid of them. The codebook is already built for the trainee (otherwise the expert could not have coded the study.) Therefore, codes used for quality coding can not be deleted from the codebook. If you need to delete the quality codes that have been applied, you use Delete Questions and choose Quality Codes for deletion. About the Study Quality Report Navigate: Client/Right-click study/Select Reports/Click Quality The report lists each question in the study, and provides summary information about the difference between the codes applied and the quality codes. 177 PrintDoc01 020813 Summary Information Description Total Responses The number of responses to this question, both coded and uncoded. Responses Coded The number of responses that have at least one code applied, shown as a count and as a percentage of responses. Responses Quality Coded The number of responses that have at least one quality code applied, shown as a count and as a percentage of responses. Responses Both Coded and Quality Coded The number of responses that have at least one code and at least one quality code applied. Codes Applied The total number of codes (not quality codes) applied to responses in this question. Quality Codes Applied The total number of quality codes applied to responses in this question. Codes Added The number of codes applied to a response when there is no matching quality code. Codes Missing The number of quality codes applied to a response where there is no matching code. Quality Ranking An index that scores the coder on their conformity to an expert’s results. The actual score consists of codes missing vs. the expert and codes added vs. the expert. Missing codes count off more than codes added. At Language Logic, we begin to take action if the ranking is below the 80th percentile. The quality ranking is only displayed to people with supervisory access or above. Clients may not see the quality ranking. The idea of this report is to give an overview of how the coding of the question differs from the quality coding. The Codes added and Codes missing rows contain the important information. These are based only on those responses that have both a code and a quality code applied. In other words, it is based only on the responses listed in the Responses both coded and quality coded row of the table. The Codes added row lists the number of codes applied where there is no matching quality code. The coder has applied codes that the quality coder did not apply. The Codes missing row lists the number of codes the coder "missed." The quality coder applied codes that the coder did not apply. Percentile scores are assigned to the Codes added and Codes missing rows. These are intended merely to give a quick assessment of the overall quality. You should use the detailed report of the coding differences to assess quality issues, rather than relying on the percentile numbers. The percentile ranking for Codes added is calculated as: (1 - (Codes added) / (Correct codes)) * 100 178 Client Access to Data where (Correct codes) is the number of quality codes that match codes applied by the coder. Similarly, the percentile rank for Codes missing is: (1 - (Codes missing) / (Correct codes)) * 100 Finally the Quality ranking for the question is computed as the average of the Codes added and Codes missing percentiles. You can see a detailed report of those responses that were coded differently if you click the icon for the question. View Codes by Coder Report Navigate: Supervisor/Studies/Right-click study/Select View Codes by Coder This report shows how many codes have been applied by each coder in all questions of the study. The report is useful when multiple coders are assigned to a study. It allows you to see who has worked on each question. Also, it is useful to troubleshoot a problem code because you can see what codes the coder applied. 179 PrintDoc01 020813 Initially, the page shows each question in collapsed form. Double-click a question row to display the codebook. The codebook displays in the first column followed by a column for each coder who worked on this question. The columns contain the number of times a coder used that code. If you click that number, you can see the verbatims with that code. To collapse the codebook view, double-click the question. A column labeled ???? indicates that some codes were not applied by any coder. The most common reason for this is that the study was restored with some applied codes. These codes are not attributed to any specific coder. (Prior to 01/29/2010, Ascribe™ did not restore the name of the coder who applied a code to a response. After that date, the coder name is restored.) Respondents Report Navigate: Supervisor/Studies/Right-click study/Select Reports/Select Respondents This report lists the respondents for a selected study. The purpose of this page is to select one or more respondents and delete them. Note Deleting respondents from this page will remove the respondent from all questions in the study. To remove respondents from individual questions, see Edit Verbatims. At the top of the page is a toolbar which offers choices for the display of respondents. The Respondent page can contain a lot of data and require a long time to display. The toolbar options allow you to select what respondents and verbatims are displayed. Here are the options: 180 Client Access to Data Field Description Records Select a range of records to display. Ascribe™ initially determines how many respondents to display based on a combination of respondent count and the number of questions. It can vary from study to study. One range of numbers means that all records are displayed. At any time, all respondents in the study can be displayed by selecting the last option in the records box "All Respondents". For large studies with many questions, this option can require significant time to display. Search Search will find and display a respondent ID or a group of respondent IDs starting with a string of characters. For example, you can search for 10015, and respondent 10015 will be listed. If you are using 5 digit numeric respondent ids and you enter "1001", the list will display 10010 through 10019. Load List Load List displays a box where you can enter individual respondent IDs or copy and paste a list of respondent IDs from another source such as Word or Excel™. To prepare a list for copying to the Load List, enter the individual respondent IDs on separate lines as follows: 10001 10005 10010 10034 After you enter the respondent IDs, click OK. The respondents and the verbatims display. The text of the Load List button displays in red when items are in the list box. Data View Data View allows you to select what is displayed in the individual cells. The choices are: Abbreviated Data - displays the first 15 characters of the response Full Data - displays the entire text of the response Load File – displays the name of the file from which the data was loaded Coded Status – displays the coding status of the response Note There are only two ways to get a respondent back after you delete it. One way is to reload the data file. The Load File option gives you the name of that file. The other way is to save the study before you delete respondents and then restore the study. But, remember, you have to save the study before you delete any respondents. Below the toolbar are these columns: a selector column, the respondent ID column, and the question ID columns (one column for each question in the study.) 181 PrintDoc01 020813 The data (response) in each cell initially displays an abbreviated response for purposes of identifying the respondent. Other data display options are available if you select the Data View option on the top toolbar. To select respondents, you can click the box next to the respondent ID. Or you can use the right-click menu to select and delete respondents: Option Description Select All This option selects all of the respondents in the current list. Clear All This option de-selects all of the respondents in the current list. Delete Selected Respondents This option removes all of the selected respondents from all questions in the study. 182 Client Access to Data Cost Accounting The various cost accounting reports detail your usage of Ascribe™ transactions. If you have administrator privilege, you can view these reports to understand how you are using your transactions. Fee Schedule 183 PrintDoc01 020813 Navigate: Administrator/Select Transactions/Select Fee Schedule This page contains two tables: Transactions by Media Type and Licensed Product. Transactions by Media Type The first table shows the number of transactions charged for a response that is coded, or to which a transcription, translation, or notes have been added. The fee varies by the media type of the response. These types are listed in the first column. Numeric responses are those which contain only the digits 0 to 9, spaces, a period, or the hyphen character; text responses are those which contain any other characters. The other media types (such as image and audio) are for multi-media responses. The second column of the table lists the number of transactions charged for coding or transcribing the corresponding type of response. Ascribe™ will apply a charge only one time to a given response. Once a fee has been applied, it will never be applied again for that response. Licensed Product Surcharges The second table on the Fee Schedule page shows each of the licensed products currently enabled for your account. A charge may be applied for each of these products. The fee for the licensed products is deducted from your transaction balance at the end of each month. The deduction is calculated as the number of transactions used in the month, multiplied by the surcharge percentage. These deductions will appear on your statement at the end of each month. Transaction Reports The following reports are available for cost accounting if you have administrator privilege: Transactions by Period, Transactions by Study, Transactions by Month, and Transactions by User. Transactions by Period Navigate: Administrator/Select Transactions/Select Transactions by Period This report shows transactions used during a specific time period. While the format of this report is very similar to the Transactions by Study report, the contents of the report are quite different. This report shows all transactions for the specified period (and only transactions that fall in that period.) The Transactions by Study report shows all transactions for specific studies, regardless of when the transaction was recorded. The report shows transactions broken down by the type of the verbatim, with rows for each study. A study will appear in the report only if at least one transaction was recorded for the study in the specified time period. The bottom row of the report shows a total column for all rows. The report will show columns for each type of response for which a transaction was charged during this period. Here are the categories of transactions: Column Description Study ID The study ID 184 Client Access to Data Study Name The study name Total Responses and Transactions The total number of responses for which a transaction charge was applied during the specified period and the total number of transactions charged to this study during the specified period ACM Responses and Transactions The number of responses that were coded with ACM and the number of transactions incurred Text Analytics Responses and Transactions The number of responses that were saved when using Inspector and the number of transactions incurred Numeric Responses and Transactions The number of numeric verbatim responses for which a transaction charge was applied and the number of transactions incurred Text Responses and Transactions The number of textual verbatim responses for which a transaction charge was applied and the number of transactions incurred Image Responses and Transactions The number of image responses for which a transaction charge was applied and the number of transactions incurred Audio Responses and Transactions The number of audio responses for which a transaction charge was applied and the number of transactions incurred Automatic Translations Requests and Transactions The number of requests made to be auto translated and the number of transactions incurred Note Click the Excel™ icon ( ) to export the data to Excel™. Transactions by Study Navigate: Administrator/Select Transactions/Select Transactions by Study The format of this report is similar to the Transactions by Period report. If a study is included in this report, it will show all transactions for that study, regardless of whether the transaction was applied within the selected time period or not. Filter by Date Range or by Study ID/Name This report can be filtered by a date range or by study ID and name. (You cannot filter by both the date range and the study ID/name.) If the Search by study ID/name box is empty, a study is included in the list if the created date of the study is within the date range you enter. If the Search by study ID/name box has an entry, Ascribe™ ignores the date range and filters by study ID/name. Select Date Range 185 PrintDoc01 020813 There are two ways to select dates. You can use the drop-down list next to the Use local time box to select today, this week, this month, this year, last week, last month, or last year. When you select one of these options and click Update, the dates change to reflect your selection, and a list of studies displays. Another way to select dates is to click the date. A calendar displays. Use the arrows next to the calendar heading to scroll through the months. Or click the heading to see a list of months. Click the heading a second time to see a list of years. Use the arrows to scroll through the months or years. Note Click the Excel™ icon ( ) to export the data to Excel™. Transactions by Month Navigate: Administrator/Select Transactions/Select Transactions by Month The format of this report is similar to the Transactions by Period report. The difference is that the transactions are not shown by study. Transactions are reported by year, month and verbatim type. The bottom row of the report shows a total column for all rows. Note Click the Excel™ icon ( ) to export the data to Excel™. The columns of the report break out transaction usage by the type of response coded. See ACM Transaction Reporting for more information about how ACM transaction usage is reported. Transactions by User Navigate: Administrator/Transactions/Transactions by Study/right-click a study/Transactions by User Supervisor/Studies/right-click a study/Reports/Transactions by User This report shows the number of transactions by user for a study. Account Statement Navigate: Administrator/Select Transactions/Select Statement You pay for Ascribe™ usage on a per-transaction basis. This report gives a statement of your account. Column Example Description Date 1/8/2002 The date of the statement entry, according to the web server clock. 186 Client Access to Data Type Transaction purchase The type of the entry: Adjustment - an adjustment made to your statement by Language Logic. Transaction purchase - purchase of transactions credited to your account. Prior month usage - the number of transactions you used in the month preceding date. Month to date usage - the number of transactions you have used in the current month to date. Description Big block! A description of the entry, which may be added by Language Logic. Transactions 10,000,000.0 The change in your transaction balance for this statement entry. Positive numbers are increases in your transaction balance; negative numbers are reductions. Balance 105,784.2 Your current transaction balance as of this entry. The last row of the table gives your current transaction balance as of this report. Time Accounting Time accounting provides time and cost information about your studies. There are two types of time accounting reports available in Ascribe™: Time by Study and Time by Period. How Ascribe™ Tracks Time Time accounting in Ascribe™ is based on entering and leaving certain web pages (tracked pages) on the site. Ascribe™ maintains a clock for each tracked activity. The clock starts when the user enters a tracked page. The clock stops (and restarts for a new activity) when the user enters a different tracked page. The clock also stops when the user navigates to the Home page, or when the user logs off. When the clock stops for an activity, the time record for that activity is finished, and the start and end times for the activity are stored. Each time record contains: The person who performed the activity (based on the logon information) The activity performed (the name of the page) The study worked on (If the activity does not relate to a study, this entry is blank.) The question worked on (If the activity does not relate to a question, this entry is blank.) The start time of the activity The end time of the activity The number of responses for which codes were applied during the tracked activity (i.e., the difference between the number of responses coded at the end time and the start time; this will be zero if the activity does not relate to a question.) The time tracking reports display the time tracking information in various ways, allowing you to look at times by study, question, activity, and associate. 187 PrintDoc01 020813 There are several important issues to remember when you are looking at time accounting reports: The times recorded in these reports always will be less than the total session (logon) time of the people working. When the time accounting clock is stopped (because it was never started, or because the user navigated to the Home page), the time is not recorded in any time tracking record. Time records are assigned to users by logon ID. This means that if you want to distinguish activities by individuals, each person using Ascribe™ must log on with a unique user name. Time records are maintained for each browser session. If a person uses two browser sessions (two instances of Internet Explorer, both logged into Ascribe™), time records will be maintained separately for each session. If you want to track activities only and are not interested in time for costing purposes, it is fine for users to log on with multiple sessions. However, if you want Ascribe™ time accounting to accurately reflect your project costs for labor hours, you should instruct your staff to use only a single browser session at a time. Idle time when the clock is running is recorded the same way as non-idle time. For example, a worker takes a break while coding a study and does not stop the clock. The worker then resumes coding later. The break time will be recorded along with active time. To prevent recording idle time, instruct your users to navigate to the Home page or log off during idle periods. The only exception to the above is when a session times out from inactivity. When a session times out, the time record is finalized with the end time as the last time the user performed any activity. Session, Logon, and Tracked Hours In the time accounting reports, you will find three ways time is tracked: Session hours - A session corresponds to an internet browser session in Ascribe. A session starts when the user logs on, and ends when the user logs off, or the session times out. If the user logs off, the session end time is the time of the logoff. If the session times out, the session end time is the time of the last activity by the user, not the time that the timeout occurred. A session is tied to the user ID of the person who logged on. Logon hours - It is possible for a given user to have more than one simultaneous session. The user might open two browsers and logon in each of them. Or, two people might logon to Ascribe at the same time using the same user ID. Logon hours report the time a given user is logged on, eliminating any overlapping session hours for the user. For an example, a user opened two browsers and logged on with the same user ID in both, then worked for an hour and logged off in both browsers. The session hours recorded for this user would be 2 (one hour in each browser). The logon hours reported would be 1 (the overlapping time is discarded). Logon hours are a better indication of the total time a user worked in Ascribe, assuming no two users log on with the same user ID. If your users never work in two browsers at the same time, logon hours will be the same as session hours. One final note: Logon hours are computed when the session ends. Logon hours will therefore not contain times for any open sessions. Tracked hours - As described above, Ascribe keeps track of time spent in certain activities. These are tracked hours. Tracked hours, like session hours, are recorded 188 Client Access to Data for each open session. If more than one session is active for the same user, tracked hours will be recorded for each session. Generally, session hours will be greater than both logon hours and tracked hours. Count of Responses Coded in Time Tracking Records The count of responses coded in time tracking records gives information about production during a user's session. To interpret this information properly, you need to understand how this value is calculated. Each code applied is marked with two pieces of information used in time accounting: The user session in which the code was applied (this also identifies the user) The date and time the code was applied. This information is used by time accounting to determine the count of responses coded. When the user logs off, each code applied during the session is inspected. For each time slot in the time accounting records for that session, the number of responses with codes applied is totaled. This becomes the count of responses coded for that time accounting record. There are a few implications of this method of counting responses coded: The sum of responses coded in time accounting for a given question will normally not equal the number of coded responses for that question. There are many reasons for this. For example, a response may have one code applied in one session, and a second code applied in a different session. In this case, the response will be counted twice in the time accounting records. Equally, a response may be coded in one session. In a later session, the response may be uncoded and coded a second time. Again, this response will be counted twice in time accounting. Counts of responses coded are calculated when the user logs off. Hence time accounting reports will not show response count information for currently open sessions. If responses are coded and the codes are deleted in the same session, they will not appear in the time accounting reports. The count of responses coded is intended to give a feel for overall production by a worker during a session. Do not use time accounting records to try to determine the number of currently coded responses. Use Time Accounting for Job Cost Calculations You can configure Ascribe™ to give you job cost based on time accounting records. To do this, enter the hourly rate for each associate under Administrator/Associates. This allows Ascribe™ to calculate cost figures in the various time accounting reports. Some clients prefer to enter a flat rate for all associates to include the entire cost of the project to obtain this estimate. Time Accounting Reports Time accounting provides time and cost information about your studies. Only administrators have access to Time Accounting. Ascribe™ has three types of time accounting reports: Time Current (currently open sessions), Time by Study, and Time by Period. Within the Time by Study and Time by Period reports, there are additional options for more detail. 189 PrintDoc01 020813 Time Categories Navigate: Administrator/Time Accounting/Time Categories An administrator can create time categories that users can select as they perform different activities on the site. Reporting by time categories is available under the Time by Study reports (Time by Category.) Create Time Categories The first step is to set up the categories on the Time Categories page. Right-click and select New. Enter the following information in the dialog box: Field ID Description The name of the time category Description A short explanation of the category Active If checked, this category is available for users in the list of time categories on the Studies, Questions, and Studies in Progress pages. If unchecked, this category will not be available to users. Press OK to save and close. To edit the time category, right-click and select Edit. The dialog box opens, and you can make changes. Set a Time Category The time categories drop-down list displays on the Studies, Questions, and Studies in Progress pages underneath the Ascribe™ logo. The default is "no time category." To set a time category, click "no time category" (or whatever time category is listed) and select the appropriate category from the drop-down list. "No time category" is always in the list, so you have the option not to use any time categories. Time Current Navigate: Administrator/Select Time Accounting/Select Time Current This report shows time accounting information for currently open sessions. Time by Period and Time by Study reports (except for the Time Details report) do not show information for currently open sessions. The report has this information: Field 190 Description Client Access to Data ACM This column will have a dot if the table entry is for an automated coding model. The column will be blank if the table entry is for a study. ID The study or ACM ID Study Name The study name Question ID The question ID or the field will be blank if the activity is not associated with a specific question. User Name The user name of the person who performed the activity Activity The tracked activity Tracked Minutes The number of minutes spent in the activity Time by Study Reports The Time by Study reports should be used when you want to look at times by job, regardless of when the activity occurred. Use the Time by Study reports when you want to know how much time was spent on a study. From the initial Time by Study report, you can select other reports to display the time accounting details for the job. If you want to examine specific periods of time, use the Time by Period reports instead. The Time by Period and Time by Study reports do not show information for currently open sessions. Use the Time Current report to see information for open sessions. Time by Category Navigate: Administrator/Time Accounting/Time by Study/Right-click a study/Select Time by Category The Time by Category report displays responses coded by time category. This report is only available if an account uses time categories. The study ID and name display at the top of the page. The report has these fields: Field Description Category The time category the user selected while working on the study. If no time category was selected while a user was working on the study, the time is listed under the heading "No time category." Responses Coded The number of responses coded in tracked activities Hours The total number of tracked hours for this study hh:mm:ss The tracked hours also display in hours:minutes:seconds format. When tracked hours exceed 24, the number of days also displays. For example, 46.50 hours would display as 1 + 22:30:00, which means 1 day, 22 hours, and 30 minutes. Responses Per Hour The ratio of responses coded divided by hours 191 PrintDoc01 020813 Cost The cost of the recorded time, using the hourly rate information for each associate Time by Study Navigate: Administrator/Select Time Accounting/Select Time by Study This report displays the time accounting information for studies. It shows all time spent on a study, regardless of when the activity occurred. If you right-click a study, you can select Time by Question, Time by Activity, Time by Category (if an account uses time categories), Time by Associate, Time by Associate, Question and Activity, and Time Details for that study. You also have the option to choose which columns to display. Filter Options This page shows a filtered list of all studies. Filtering is based on the settings in the filter bar at the top of the page. To change the filter, check the desired boxes in the filter bar, and click the Update button. The text filter in the bar allows you to filter studies by study ID or study name. To show only studies whose ID or name starts with certain characters, enter those characters in the Filter box and click the Update button. For example, if you enter " abc" in the Filter box, only those studies whose ID or Name start with abc will be shown in the list. To find studies whose ID or name contains certain characters, put an asterisk as the first character of the filter. For example, if you enter "* abc" in the Filter box, only those studies whose ID or name contains abc will be shown in the study list. Report Columns The report has these columns: Field Created Date Description The date the study was created ACM This column will have a dot if the table entry is an ACM ID The ID of the study or ACM Name The name of the study or ACM Status The status of the study or ACM The next columns repeat for assigned work, unassigned work, and total time. Field Description Responses The number of responses in the study Responses Coded The number of responses coded in tracked activities Hours The total number of tracked hours for this study hh:mm:ss The tracked hours also display in hours:minutes:seconds format. When tracked hours exceed 24, the number of days also displays. For example, 46.50 hours would display as 1 + 22:30:00, which 192 Client Access to Data means 1 day, 22 hours, and 30 minutes. Responses Per Hour The ratio of responses coded divided by hours Cost The cost of the recorded time, using the hourly rate information for each associate Time by Question Within a Study Navigate: Administrator/Select Time Accounting/Select Time by Study/Right-click study/Select Time by Question This report displays the time accounting for the study, organized by question. It has this information: Field Description Question ID The question ID Question Label The question label Question Type The question type Deleted This column will have a dot if the question has been deleted. Codebook ID The codebook ID Responses The current number of responses for this question The next columns repeat for assigned work, unassigned work, and total work: Field Description Responses Coded The number of responses coded in tracked activities for this user Hours The number of hours spent on this question hh:mm:ss The hours also display in hours:minutes:seconds format. When hours exceed 24, the number of days also displays. For example, 46.50 hours would display as 1 + 22:30:00, which means 1 day, 22 hours, and 30 minutes. Responses Per Hour The number of responses coded divided by hours Cost The sum of the hourly rate for each associate who worked on the question, multiplied by the number of hours the associate spent on this question. To choose the columns which are displayed, right-click a user and select Choose Columns. The study ID and name display above the report. The study ID has a drop-down box with information about the study. You can navigate to the study question from the study ID drop-down box. Time by Activity Within a Study Navigate: Administrator/Select Time Accounting/Select Time by Study/Right-click study/Select Time by Activity 193 PrintDoc01 020813 This report displays time accounting for the study, organized by activity. The activity is the first field. The name of the activity generally corresponds to the page location in Ascribe™. The next columns repeat for assigned work, unassigned work, and total work: Field Description Responses Coded The number of responses coded in this activity Hours The number of hours spent by all users in this activity. hh:mm:ss The hours also display in hours:minutes:seconds format. When hours exceed 24, the number of days also displays. For example, 46.50 hours would display as 1 + 22:30:00, which means 1 day, 22 hours, and 30 minutes. Responses Per Hour The number of responses coded divided by hours Cost The sum of the hourly rate for each associate who worked on the question, multiplied by the number of hours the associate spent on this question. To choose the columns which are displayed, right-click a user and select Choose Columns. The study ID and name display above the report. The study ID has a drop-down box with information about the study. You can navigate to the study question from the study ID drop-down box. Time by Associate Within a Study Navigate: Administrator/Select Time Accounting/Select Time by Study/Right-click study/Select Time by Associate This report displays time accounting for the study, organized by associate. The user name is the first field. The information below repeats for assigned work, unassigned work, and total work: Field Description Responses Coded The number of responses coded in this study by the associate Hours The number of hours spent on this study by the associate hh:mm:ss The hours also display in hours:minutes:seconds format. When hours exceed 24, the number of days also displays. For example, 46.50 hours would display as 1 + 22:30:00, which means 1 day, 22 hours, and 30 minutes. Responses Per Hour The number of responses coded divided by hours Cost The sum of the hourly rate for each associate who worked on the question, multiplied by the number of hours the associate spent on this question. To choose the columns which are displayed, right-click a user and select Choose Columns. The study ID and name display above the report. The study ID has a drop-down box with information about the study. You can navigate to the study question from the study ID drop-down box. 194 Client Access to Data Time by Associate, Question, and Activity Within a Study Navigate: Administrator/Select Time Accounting/Select Time by Study/Right-click study/Select Time by Associate, Question, Activity This report displays time accounting for the study, organized by associate, then by question, and then by activity. It has this information: Field Description User Name The user name of the associate Question ID The question ID. If blank, this time accounting record is not associated with a specific question, or the question ID is blank. Question Label The question label. If blank, this time accounting record is not associated with a specific question, or the question label is blank. Activity The name of the activity. Activity names generally correspond to the page location in Ascribe™. The information below repeats for assigned work, unassigned work, and total work: Field Description Responses Coded The number of responses coded in this study by the associate Hours The number of hours spent on this study by the associate hh:mm:ss The hours also display in hours:minutes:seconds format. When hours exceed 24, the number of days also displays. For example, 46.50 hours would display as 1 + 22:30:00, which means 1 day, 22 hours, and 30 minutes. Responses Per Hour The number of responses coded divided by hours Cost The sum of the hourly rate for each associate who worked on the question, multiplied by the number of hours the associate spent on this question. To choose the columns which are displayed, right-click a user and select Choose Columns. The study ID and name display above the report. The study ID has a drop-down box with information about the study. You can navigate to the study question from the study ID drop-down box. Time Details Within a Study Navigate: Administrator/Select Time Accounting/Select Time by Study/Right-click study/Select Time Details This report displays all time accounting records for this study. This report can be used in Excel™ to create a pivot table to examine the time accounting records in a custom format. It has this information: Field Description Activity The name of the activity. Activity names generally correspond to the page location in Ascribe™. 195 PrintDoc01 020813 User Name The user name of the associate Question ID The question ID. If blank, this time accounting record is not associated with a specific question, or the question ID is blank. Label The question label. If blank, this time accounting record is not associated with a specific question, or the question label is blank. Type The question type. If blank, this time accounting record is not associated with a specific question. Assigned A dot in this columns means that the user was assigned to the study. Start The time the activity started, based on the server clock and adjusted to your local time zone Responses Coded The number of responses coded during this activity Hours Tracked The number of tracked hours spent in this activity hh:mm:ss The hours also display in hours:minutes:seconds format. When hours exceed 24, the number of days also displays. For example, 46.50 hours would display as 1 + 22:30:00, which means 1 day, 22 hours, and 30 minutes. Responses Per Hour The number of responses coded divided by hours Cost The cost is calculated as tracked hours multiplied by the rate for this user. To choose the columns which are displayed, right-click a user and select Choose Columns. The study ID and name display above the report. The study ID has a drop-down box with information about the study. You can navigate to the study question from the study ID drop-down box. Time by Period Reports The Time by Period reports should be used when you want to look at certain periods of time. For example, you might want to monitor how your coding staff spent their time last month. Time by Period reports can give you this information. From the initial Time by Associate report, you can select other reports to display the time accounting details across activities and studies. You can change the time period if you click the Change button at the top of the screen. A dialog box displays, and you can select the dates. Click OK to see the report for those dates. For reports that show time by study instead of period, use the Time by Study Reports instead. The Time by Period and Time by Study reports do not show information for currently open sessions. Use the Time Current report to see information for open sessions. Time by Period 196 Client Access to Data Navigate: Administrator/Select Time Accounting/Select Time by Period This report displays time accounting information for users within the selected time period. Select Date Range There are two ways to select dates. You can use the drop-down list next to the Use local time box to select today, this week, this month, this year, last week, last month, or last year. When you select one of these options and click Update, the dates change to reflect your selection, and the report displays. Another way to select dates is to click the date. A calendar displays. Use the arrows next to the calendar heading to scroll through the months. Or click the heading to see a list of months. Click the heading a second time to see a list of years. Use the arrows to scroll through the months or years. Report Columns The report has this information: Column Description User Name The user name of the associate Session Hours The total session hours for this associate for sessions in the specified period hh:mm:ss The session hours also display in hours:minutes:seconds format. When session hours exceed 24, the number of days also displays. For example, 46.50 hours would display as 1 + 22:30:00, which means 1 day, 22 hours, and 30 minutes. Logon Hours The total logon hours for this associate in the specified period. hh:mm:ss The logon hours also display in hours:minutes:seconds format. When logon hours exceed 24, the number of days also displays. For example, 46.50 hours would display as 1 + 22:30:00, which means 1 day, 22 hours, and 30 minutes. Tracked Hours The total tracked hours for this associate in the specified period. hh:mm:ss The tracked hours also display in hours:minutes:seconds format. When tracked hours exceed 24, the number of days also displays. For example, 46.50 hours would display as 1 + 22:30:00, which means 1 day, 22 hours, and 30 minutes. (last column has no title) The tracked hours are also displayed as a percentage of session hours. This report is similar to the default Time by Associate report, but differs in the way sessions are selected for inclusion. In this report, a session is included if the duration of the session overlaps the specified time period. In addition, if the duration of a session extends beyond the specified time period, the portion of the session outside the specified time period is omitted. 197 PrintDoc01 020813 This report is useful to get an accurate picture of session time by user when there are sessions whose duration crosses the midnight boundaries of the selected period, and when the period includes open sessions. Note that if sessions cross the midnight boundaries of the selected time period, the information presented in this report will differ from that in the Time by Associate report. Time by Activity Within a Period Navigate: Administrator/Select Time Accounting/Select Time by Period/Right-click user/Select Time by Activity This report displays time accounting information for activities during the specified period. Information for currently open sessions is not contained in the report. Select Date Range There are two ways to select dates. You can use the drop-down list next to the Use local time box to select today, this week, this month, this year, last week, last month, or last year. When you select one of these options and click Update, the dates change to reflect your selection, and the report displays. Another way to select dates is to click the date. A calendar displays. Use the arrows next to the calendar heading to scroll through the months. Or click the heading to see a list of months. Click the heading a second time to see a list of years. Use the arrows to scroll through the months or years. Report Columns The report has these columns: Field Description Activity The name of the activity. Activity names generally correspond to the page location in Ascribe™. Tracked Hours The total of tracked hours for this activity by all users during the specified period hh:mm:ss The tracked hours also display in hours:minutes:seconds format. When tracked hours exceed 24, the number of days also displays. For example, 46.50 hours would display as 1 + 22:30:00, which means 1 day, 22 hours, and 30 minutes. Cost The cost of the tracked hours, computed by multiplying the rate for each associate by the number of tracked hours. Responses The number of responses coded in this activity Responses Per Hour The number of responses divided by tracked hours The report also contains a “Not tracked” item, which lists the difference between session hours and tracked hours for the specified period. This is the user session time that is not captured as tracked time. Time by Study Within a Period Navigate: Administrator/Select Time Accounting/Select Time by Period/Right-click user/Select Time by Study 198 Client Access to Data This report displays time accounting information for studies during the selected period. Select Date Range There are two ways to select dates. You can use the drop-down list next to the Use local time box to select today, this week, this month, this year, last week, last month, or last year. When you select one of these options and click Update, the dates change to reflect your selection, and the report displays. Another way to select dates is to click the date. A calendar displays. Use the arrows next to the calendar heading to scroll through the months. Or click the heading to see a list of months. Click the heading a second time to see a list of years. Use the arrows to scroll through the months or years. Report Columns The report has these columns: Column Description Study ID The study ID Study Name The study name Tracked Hours The total number of tracked hours for this study during the selected period hh:mm:ss The tracked hours also display in hours:minutes:seconds format. When tracked hours exceed 24, the number of days also displays. For example, 46.50 hours would display as 1 + 22:30:00, which means 1 day, 22 hours, and 30 minutes. Cost The cost of the tracked hours, calculated as the associate rate multiplied by the tracked hours Responses The number of responses coded in the tracked activities Responses Per Hour The number of responses coded divided by tracked hours The report also contains a “Not tracked” item, which lists the difference between session hours and tracked hours for the specified period. This is the user session time that is not captured as tracked time. Time by User and Study Status Within a Period Navigate: Administrator/Select Time Accounting/Select Time by Period/Right-click in white space/Select Time by User and Study Status Coder/Time by User and Study Status This report displays time accounting records for a specified period of time. The report displays one row for each user who had tracked activities during the specified period. For each such user, the report displays time accounting information classified by the status of the study at the time the activity was tracked. This report allows you to view how coders spent their time by study status. Coders have access to this report, but the report will show information only for that coder. Administrators also have access to the report, and it will show information for all users. 199 PrintDoc01 020813 Select Date Range There are two ways to select dates. You can use the drop-down list next to the Use local time box to select today, this week, this month, this year, last week, last month, or last year. When you select one of these options and click Update, the dates change to reflect your selection, and the report displays. Another way to select dates is to click the date. A calendar displays. Use the arrows next to the calendar heading to scroll through the months. Or click the heading to see a list of months. Click the heading a second time to see a list of years. Use the arrows to scroll through the months or years. Report Columns The first five columns of the report are: Field Description User Name The user name of the associate Session Hours The total number of session hours for this user during the period hh:mm:ss The session hours also display in hours:minutes:seconds format. When session hours exceed 24, the number of days also displays. For example, 46.50 hours would display as 1 + 22:30:00, which means 1 day, 22 hours, and 30 minutes. Logon Hours The total number of logon hours for this user during the period hh:mm:ss The logon hours also display in hours:minutes:seconds format. When logon hours exceed 24, the number of days also displays. For example, 46.50 hours would display as 1 + 22:30:00, which means 1 day, 22 hours, and 30 minutes. The next columns repeat for the total of all study status and for each study status classification: Field Description Responses Coded The number of responses coded in tracked activities for this user Tracked Hours The number of tracked hours spent for this user hh:mm:ss The tracked hours also display in hours:minutes:seconds format. When tracked hours exceed 24, the number of days also displays. For example, 46.50 hours would display as 1 + 22:30:00, which means 1 day, 22 hours, and 30 minutes. Responses Per Hour The number of responses coded divided by hours Cost The cost is calculated as tracked hours multiplied by the rate for this user Like the Time by Associate report, sessions are selected for inclusion in this report if the start time of the session is within the period specified. Open sessions are not included in the report. 200 Client Access to Data Note Tracking of activities by study status was introduced in Ascribe™ version 5.0. Any activities recorded before this release are assigned a study status of "coding in progress.” Time by Study, User and Activity Within a Period Navigate: Administrator/Select Time Accounting/Select Time by Period/Right-click user/Select Time by Study, User, Activity This report displays time accounting for the study, organized by study ID, then by user, and then by activity. Select Date Range There are two ways to select dates. You can use the drop-down list next to the Use local time box to select today, this week, this month, this year, last week, last month, or last year. When you select one of these options and click Update, the dates change to reflect your selection, and the report displays. Another way to select dates is to click the date. A calendar displays. Use the arrows next to the calendar heading to scroll through the months. Or click the heading to see a list of months. Click the heading a second time to see a list of years. Use the arrows to scroll through the months or years. Report Columns The report has these columns: Column Description Study ID The study ID Study Name The study name User Name The user name Activity The name of the activity. Activity names generally correspond to the page location in Ascribe™ Tracked Hours The tracked hours for this study and activity hh:mm:ss The tracked hours also display in hours:minutes:seconds format. When tracked hours exceed 24, the number of days also displays. For example, 46.50 hours would display as 1 + 22:30:00, which means 1 day, 22 hours, and 30 minutes. Cost The cost of the activity, calculated as the tracked hours multiplied by the associate rate Responses The number of responses coded in the tracked activity Responses Per Hour The number of responses coded divided by hours The report also contains a “Not tracked” item, which lists the difference between session hours and tracked hours for the specified period. This is the user session time that is not captured as tracked time. 201 PrintDoc01 020813 Time Details Within a Period Navigate: Administrator/Select Time Accounting/Select Time by Period/Right-click user/Select Time Details This report displays all time accounting records for a user. This report can be used in Excel™ to create a pivot table to examine the time accounting records in a custom format. Select Date Range There are two ways to select dates. You can use the drop-down list next to the Use local time box to select today, this week, this month, this year, last week, last month, or last year. When you select one of these options and click Update, the dates change to reflect your selection, and the report displays. Another way to select dates is to click the date. A calendar displays. Use the arrows next to the calendar heading to scroll through the months. Or click the heading to see a list of months. Click the heading a second time to see a list of years. Use the arrows to scroll through the months or years. Report Columns The report has these columns: Column Description Activity The name of the activity. Activity names generally correspond to the page location in Ascribe™ User Name The user name of the associate Study ID The study ID Study Name The study name Question ID The question ID. If blank, this time accounting record is not associated with a specific question, or the question ID is blank. Question Label The question label. If blank, this time accounting record is not associated with a specific question, or the question label is blank. Question Type The question type. If blank, this time accounting record is not associated with a specific question. Start The time the activity started, based on the server clock and adjusted to your local time zone Tracked Hours The number of hours spent in this activity hh:mm:ss The tracked hours also display in hours:minutes:seconds format. When tracked hours exceed 24, the number of days also displays. For example, 46.50 hours would display as 1 + 22:30:00, which means 1 day, 22 hours, and 30 minutes. Responses Coded The number of responses coded during this activity Production Bands Report Navigate: Administrator/Production Bands 202 Client Access to Data The Production Bands Report allows users with administrator privilege to run productivity numbers for their account. Select Date Range There are two ways to select dates. You can use the drop-down list next to the Use local time box to select today, this week, this month, this year, last week, last month, or last year. When you select one of these options and click Update, the dates change to reflect your selection, and the report displays. Another way to select dates is to click the date. A calendar displays. Use the arrows next to the calendar heading to scroll through the months. Or click the heading to see a list of months. Click the heading a second time to see a list of years. Use the arrows to scroll through the months or years. Rows Represent a Time Accounting Session Each row in the Production Bands Report represents a time accounting session in Ascribe™ where responses were coded in a particular window (either Review Responses, Phrase Analyzer, the Coder Window, or the Multiple Response Coding Tool). The time accounting session starts when a code is applied, and it stops when the user leaves the window. The responses coded in that session are counted. The characters for each response are counted and divided by the total responses to determine the average character band for the session. Notes about Auto-Coding and ACM Coding Responses that are auto-coded using the Auto-coder script and Auto Coded on Load are not effectively reported in the production bands report. Responses that are ACM coded are reported if the suggestions are accepted using the Desktop-ACM Coding Review tool; otherwise they are not effectively reported. Sometimes, activities that have nothing to do with coding will be reported in the Production Bands Report. The auto-coding script tries to attribute coded responses to the user who ran the script and to the page the user is on while the script is running. If the user submits the auto-coding script and moves to another page, that page will be reported as the activity for the verbatim coded by the script. Columns in the Production Bands Report If you right-click the table, you have the option to choose which columns display. Here are the available columns: Column Description Account The name of the Ascribe™ account. (As a default, this column is not displayed. Right-click the table and select Choose Columns. You can choose which columns will display.) Year The year Month 1 = January 2 = February 3 = March 203 PrintDoc01 020813 4 = April 5 = May 6 = June 7 = July 8 = August 9 = September 10 = October 11 = November 12 = December Quarter 1 = January, February, and March 2 = April, May, and June 3 = July, August, and September 4 = October, November, and December User Name The user who performed the activity Band # The Band # (Number) and Band description columns are used together to describe the type of responses that were coded. The Band Numbers are 0 to 5. The Band description has the number of characters in brackets like this:[1-12]. The character band is determined by how many responses are coded divided by the number of characters. For example, a user codes 80 responses, and the responses have 2,000 characters. 2000 divided by 80 is 25 so the user's work would be reported in Band Number 3. Individual Bands were not tracked Until October 2006 0 = [0] - This band includes responses from images and voice recordings where the character band can’t be determined; it also includes numeric responses. 1 = [1-12] - This band includes responses with one or two words or perhaps a brand list. 2 = [13-21] - This band includes responses with two to five words; they are short answers or phrases. 3 = [22-64] - This band includes responses that are short paragraphs; it usually includes more than one idea and takes a bit of time to read and distinguish ideas. 4 = [65-133] - This band includes responses that are long sentences with complicated or repeated ideas. 5 = [>133] - This band includes responses that are long sentences with complicated or repeated ideas. A small percentage of responses loaded into Ascribe™ are this size or larger. Band The Band description column shows the number of characters in brackets. For example, [0], [1-12], etc. See Band # for more information. Activity Coding activities on these pages are tracked: Coding Page (Coder Window) 204 Client Access to Data Review Responses Phrase Analyzer Desktop- Code Multiple Responses Desktop - ACM Coding Review Note that other pages may be display here; see Notes about Auto-Coding and ACM Coding. Type The Type column displays the question type (closed, open, other specify, etc.) Hours The Hours column is the time spent using a coding tool where at least one response was coded. Time accounting in Ascribe™ is based on entering and leaving certain web pages (tracked pages) on the site. Ascribe™ maintains a clock for each tracked activity. The clock starts when the user enters a tracked page. The clock stops (and restarts for a new activity) when the user enters a different tracked page. The clock also stops when the user navigates to the Home page, or when the user logs off. For example, if you are working in the Coding Page for 40 minutes and then you navigate to Review Responses, your tracked time would stop when you leave the Coding Page. It would restart when you reach Review Responses. Responses Coded This column is the sum of all responses coded. It includes text, numeric, and media (images and voice recording.) Numeric responses are those which contain only the digits 0 to 9, spaces, and the hyphen character. Text Responses Coded Text responses are those which contain any character other than 0 to 9, spaces, and the hyphen character. *Numeric responses could fall into this category “if” they are coded in the same time accounting session as the text responses. *Text responses were not tracked separately until December 2006 Characters Coded This column is the determined by the number of characters in all text responses coded divided by the sum of characters included in text response count. Export the Production Bands Report Click the Excel™ icon ( ) to export the data to Excel™. 205 Ascribe Presentations™ Navigate: Supervisor/Studies/Right-click a study/Select Ascribe™ Presentations Client/Studies/Right-click a study/Select Ascribe™ Presentations Ascribe Presentations are a full featured set of tools for presenting Ascribe™ data in various graphical and tabular formats. These tools are part of Ascribe™ Web and do not require additional installation of any software. You can generate these graphs with Ascribe Presentations: bar, line, and pie charts, cooccurrence maps, correlation matrices, proportional code font views, multidimensional scalings, clusters, word clouds, and concept clouds. Also, you can create many types of cross-tabs: single-dimension list of frequencies, crossed frequencies (one or more questions on the side and one or more as banners), filtered crosstabs, stacked banners, and grids. The Cross-Tabs module has many options to customize the tables. You can choose font size, sorting, drill downs, percentages, t-tests, etc. The tables can contain "codebook codes" or the actual words extracted from the respondent's original responses. You can also create bar, line, or pie charts at the same time you create the cross-tab. The cross-tabs and graphs can be exported to Excel ®, PowerPoint, or a simple HTML file that can be saved and sent independently from Ascribe™ Presentations. Base and Professional Editions Ascribe Presentations has two versions: Base Edition and Professional Edition. Your version is dependent on your license agreement. The words "Base Edition" or "Professional Edition" will display on the Ascribe Presentations main page to let you know which version you have. The Base Edition does not allow access to all features of Ascribe Presentations. The chart below explains which features are available in the two editions: Features Base Professional Graphs X X Bar Charts X X Line Charts X X Pie Charts X X Proportional Code Font Views X X Co-Occurence Maps X X Correlation Matrices Across Codes X Multidimensional Scalings X Clusters X Multiple Charts Cross-Tabs X X X 207 PrintDoc01 020813 Sort Tables X X Crosstab Words (Rather Than Codes) X X Highlight Column Percentage Differences X X Show Ranks X X X X Ability To Filter Tables Graphs Attached To Tables X Co-Occurrence Graphs Across Questions X t-Tests Across Columns Word Clouds & Concept Clouds Ability To Sort Words (Alpha/Freq/Random) Display Frequencies And Percentages Ability To Filter Word/Concept Clouds Options X X X X X X X Allow Drill Down X Export To Word/Power Point/Excel Export To JPEG X X Export To HTML X Copy To Clipboard X Email Results X Ascribe Presentations Main Page The Ascribe Presentations main page has a toolbar across the top of the page and a list of presentation tools. To select one of the tools, click one of the image buttons. The toolbar has the following fields: Field Description Question Selection Select the type or types of questions for the reports. Options are open, closed, and other specify. You can also select to work with only questions that are completely coded. Study ID: Study Name This field has a drop-down box with study information. Studies This button navigates to the Studies page. The right-click menu also provides navigation to the Studies page or other areas of 208 Ascribe Presentations™ Ascribe™. Cross-Tabs Navigate: Supervisor/Studies/Right-click a study/Select Ascribe Presentations/Select Cross-Tabs Client/Studies/Right-click a study/Select Ascribe Presentations/Select Cross-Tabs You can create many types of cross-tabs with this module. You select the questions to be used on the Settings page, which has six tabs of options. The Settings page has a toolbar with these fields: Field Description Run This option runs the report and navigates to the Results page. Click this button after you have selected the questions and various options. Settings This option navigates from the Results page to the Settings page. (It is grayed out on the Settings page.) Study ID: Study Name This field has a drop-down box with study information. Ascribe Presentations This button navigates to the Ascribe Presentations main page. The right-click menu also provides navigation to the Ascribe Presentations main page or other areas of Ascribe™. Questions Tab The Questions Tab has three lists of questions for the Cross-Tab report. In order for a question to be listed on the Questions Tab, the question must have the Cross-Tab task selected on Task Tab of the Edit Question dialog. You must select at least one side question or one grid question in order to generate a cross-tab table. 209 PrintDoc01 020813 Choose from these options: Option Description Side Questions These questions will display on the side of the table. A cross-tab table generates for each side question. If you want to select or unselect all the side questions, click the check box next to the Side Questions heading. Options: - Use the Up and Down buttons to move a selected question higher or lower in the list. Banner Questions These questions will display along the top of the table. You can select more than one banner question. Any banner question that contains more than 256 codes will display only the first 256. Options: - You can filter the list of questions and display only closed-ended questions if you click the check box next to Closed-Ends Only. - Banners can be stacked over each other with the use of this button. If you select two questions and click the Stack Banners button, a new temporary banner question is created. You can stack up to 3 questions. A stacked question cannot be stacked over another one. It is advised to use only closed-ended questions as the number of codes may grow very quickly. - Use the Up and Down buttons to move a selected 210 Ascribe Presentations™ question higher or lower in the list. When you are stacking banner questions, you can use these buttons to define the order in which questions are stacked over each other. Grids This option compares questions that share the same codebook. The codebook ID (in bold) displays first, followed by the questions which share the codebook. You can select all questions by clicking the box next to the codebook ID or you can select individual questions. Option: - This option displays all questions that share a codebook, including those in other studies. It enables you to make comparisons across studies. Crosstab By This drop-down box displays all the closed-ended questions of the current study. When you select a side question and a question here, the Cross-Tab module creates one set of tables for each of the codes of the selected closed-end. (The difference is between using the Crosstab By option and selecting a banner question is that the banner question will have only one table. The Crosstab By option will create as many tables as there are codes in the closedend.) Each question has this format: Field Description When you click this symbol, a drop-down dialog displays codebook information and the list of codes. Use the double arrow button ( dialog. or ) to collapse the upper or lower panes of the Click the check box to select or unselect a question. Question ID.Question Label The question ID displays, followed by the question label (if the label is present.) Closed-ended questions have brackets around these fields. (Responses Coded/Total Responses) The number of responses coded in the question followed by the total number of responses in the question. (Percentage) The percentage of responses that have been coded in the question. Here is an example of a cross-tab with only a side question: Here is an example of a question where the language question is cross-tabbed by the country question: 211 PrintDoc01 020813 Here is an example of a cross-tab grid across shared questions: Codes Tab On the Codes Tab, you can select options that apply to codes and/or nets. Here are the options: Option Description Codes On The Side The tables will display codes on the side, rather than words. 212 Ascribe Presentations™ Codes As Banners The tables will display codes in the banners, rather than words. Display Select to display: Codes and Nets Codes Only Nets Only Applies to Banners - applies the Display settings to banners also. Show Code Output ID Select this option to display the code output ID in front of the code. Responses Count This option applies to nets only. But default, net counts are unduplicated counts of respondents. Unduplicated means that if a respondent was given more than one code belonging to a net, the respondent will be counted only once in the net total. If you select the Responses Count option, the respondent counts will not be unduplicated. Codes To Ignore Enter codes that you want to ignore in the side questions. You can enter the code output ID or the code description. Separate multiple entries with semicolons. Codes Not To Sort Enter codes that you do not want to be sorted. You can enter the code output ID or the code description. Separate multiple entries with semicolons. Code Selection Use the Get Codes button to see a list of questions. Check the box next to a question to see the list of codes and nets in the question. By default, each code is selected. When you unselect a code or a net, it will not display in the cross-tab table. Underneath each codebook list is a text box where you can merge codes. The codes must be separated by a semicolon. To create multiple groups, type a vertical bar (|) and enter another combination. Here is an example: 1;3|5;6;7|12;13;14 Here's an example of the Get Codes codebook list. In this example, the text box contains three codes to be merged: Here is the cross-tab using those settings: 213 PrintDoc01 020813 Words Tab The Words Tab has options related to the generation of word counts (as opposed to code counts.) Here are the options: Option Description Words On The Side The tables will display words on the side, rather than codes. Selecting this option will unselect its equivalent in the Codes Tab (either you want tables of codes or words, but not both.) Words As Banners The tables will display words in the banners, rather than codes. Selecting this option will unselect its equivalent in the Codes Tab (either you want tables of codes or words, but not both.) Response Source Select to display: Verbatim 214 Ascribe Presentations™ Translation Transcription Use Verbatim if no Translation/Transcription Min Word Length Words that are equal to or less than the number entered here will be ignored. Words To Ignore A default list of words appears in the text box. The list is optional; you can unselect it with the check box. Also, you can change the list to whatever you want (be sure to use semicolons to separate the words.) To restore the default ignore list, click the Default button at the bottom of the Options Tab. When creating word count tables, it is advised to set the Minimum Frequency field (on the Options Tab) to something meaningful like 10 or 50. Otherwise, you may end up with thousands of words. Your table could be huge and difficult to read. Here is an example of a word count table, with minimum frequency set to 50: Drill Down with Words If the Allow Drilldown box is checked on the Options Tab, you can click a word in the list to display the list of responses which contain that word. The default display centers the selected word in the list, so it is easy to see the context in which it appears. If the response before the keyword is longer than 50 characters, the response will be truncated. The truncation is noted with a leading ellipsis (...). You can choose whether the responses are centered or not. Here is an example with the keyword centered: 215 PrintDoc01 020813 Here is an example where the keyword is not centered: If you click on any line, you will see the respondent's answers to all questions to the current study. 216 Ascribe Presentations™ Here are the options in the word drill down pane: Option Description Response List You have the choice to display the responses that contain the selected word or you can choose word proximities. Click any line in the table to see the respondent's answers to all questions in the current study. You can export the table to HTML or Excel ®. Also, you can export all data for these respondents (all of their verbatim in the study) to a text file. Word Proximities Creates a table of words that are close to the selected word and will also display them in a circular graph. The graph is displayed underneath the table. You can export the table to HTML or Excel®. Graph Size Enter whatever size you would like the graph to be. Set As Filter Use this option to set the selected word as filter for the crosstab. You have to re-run the cross-tab to see the filtering. Center Keyword Use this check box to toggle on or off the centering of the selected word. In this example of a word proximities table, the word "free" appears next to the word "static" 71 times. The word "the" is next to "static" 35 times and is one word away from "static" 26 times. Here is an example of the graph that goes with the table: 217 PrintDoc01 020813 Graphs Tab The Graphs Tab is where you select the type of graph you want to generate. The left side of the page lets you choose the graphs that are associated with each table. The right side lets you generate summary graphs across questions. Here are the options: 218 Ascribe Presentations™ Option Description Display Below Each Table None Choose None if you do not want any graphs for the cross-tab. Graph Height Lets you specify the height of the graph. Descriptive Graphs Bar Charts Generates a bar chart for the cross-tab. One Series Per Banner Point - The bar charts will display each banner point as a different series. It allows for comparison of data across banners in a graphical way. Vertical - Displays the bar chart vertically rather than horizontally. Line Charts Generates a line chart for the cross-tab. Pie Charts Generates a pie chart for the cross-tab, which are most likely to be used for closed-ended questions. Analytic Graphs Co-Occurrences The purpose of a co-occurrence chart is to look at the frequency with which codes were applied to the same response. A cooccurrence map contains circles and lines. Circles are codes, and line represent the co-occurrence between two codes. The high the code frequency, the larger the circle. The higher the cooccurrence between two codes, the thicker the line. Vertical CoOccurrences Represents the co-occurrences in a vertical way rather than a circle. First n CoOccurrences Displays the first n co-occurrences in terms of frequency. Min CoOccurrence Co-occurrences below this number will not display. Hide Codes with No CoOccurrence Codes that have no-occurrence with other codes will not display. Correlation Matrices Displays the correlation matrix between each pair of codes. The bluer the color, the higher the correlation between two questions. The redder the color, the higher the negative correlation between questions. The whiter the color, the higher the non-existence of any type of correlation. Include Nets In Analytic Graphs We advise you not to use nets in co-occurrence graphs and correlation matrices as they contain redundant information. For example, all codes will have high co-occurrences with their parent nets. First N Codes Enter a number to restrict the number of codes displayed in the analytic graphs. 219 PrintDoc01 020813 Cross-Questions Graphs These graphs help you visualize the relationship between two questions or two sets of codes (or words) coming from two different questions. (Must also select Correlation Matrices, CoOccurrences, or Vertical Co-Occurrences under Analytic Graphs to use the following options.) Cross Questions Correlations Generates a correlation matrix across all the selected numeric questions (questions in which the responses are numbers.) Note that for readability, the highest 10% of positive correlations are surrounded by a red square. Vertical CoOccurrences Across Questions Generates a co-occurrence graph for each pair of side questions selected in the Questions Tab. The following options are useful if you have large tables and want to limit the amount of information exported to a chart: First n Co-Occurrences - Display the first n cooccurrences in terms of frequency. Min Co-Occurrence - Co-occurrences below this number will not display. Hide Codes with No Co-Occurrence - Codes that have no-occurrence with other codes will not display. Options Tab This tab contains a variety of options for creating cross-tabs. The options are: Option Description General Options Sort Tables Provides the ability to sort tables by decreasing order of frequency. If the table contains nets, the nets are sorted first and then each code (or subnet) within each net. This option applies only to regular tables, not to grid tables. t-tests A t-test is a statistical test which can be performed between two column percentages in order to determine if the difference of percentage observed is significant or not. If you check that box, all the banner codes are assigned numbers which display in parentheses underneath the codes. Red numbers display in each cell indicating the columns which are statistically different. Note that you can set the confidence level of the t-test (80%, 90% 95% or 99%). See the example below. Allow Drilldown Allows you to click any cell in any table or graph and look at all the responses that make up that cell. A separate dialog opens with the responses and allows export to HTML or Excel ®. Also, you can export all data for these respondents (all of their verbatim in the study) to a text file. If you click the "Set as Filter" button, it will set the current respondents as a filter for all your tables. Remember to refresh 220 Ascribe Presentations™ your tables by clicking "Tables" button after the drill down window has closed. Clicking any line of the grid will display the respondent details. If you plan to generate hundreds of tables with large banners, you may not want to allow drill down, as it may render the tables more slowly. Proportional Font Depth This is a very powerful feature that lets you see which codes have a higher frequency without looking at the numbers themselves. If checked, the highest frequency codes or nets will appear in black. The lower the frequency, the lighter the font. It may be better to display only codes or only nets here as nets may make all the codes' fonts look light. Proportional Background Color This enables you to quickly locate cells that have a high percentage. The higher the percentage, the deeper the background color. Table Of Content This is a very useful feature if you are generating many tables. It creates a link to each table at the top of the result page. It also creates a "Top" link above each table so you can easily navigate back to the top. Enable Export To Word Allows export to Word. If enabled, this icon displays at the top of Enable Export To PowerPoint Allows export to PowerPoint. If enabled, this icon displays at the Max. Rows Per Slide Use this option to specify how many lines per table you want per page in exports to Word or PowerPoint. Split Excel Cells If the user displays absolute values and percentages in each of the table cells, it is difficult to manipulate these numbers in Excel. If you select this option, you will see multiple sheets in Excel: a sheet with the absolute values and percentages in the same cell, a sheet with just the absolute values in the cell, and a sheet with just the percentages. Total These options define what number should appear in the top row of each table. The main purpose of this option is to adapt column percentages. Total Sample Displays the total sample. Total Coded Displays the number of people who gave at least one answer that was coded in the current side question. Raw Counts Allows you to define whether or not you want raw counts to appear in the tables. It may be that you just want percentages. the Results page: top of the Results page: Percentages None Displays no percentages. 221 PrintDoc01 020813 Column Total Displays percentages based on the total column (first column.) Row Total Displays percentages based on the total row (first row.) Table Total Displays percentages based on the table total. Percentage Decimals A number between 0 and 2 that lets you choose how many decimals you want to be displayed for percentages. Highlight Column Percentage Differences This tool enables you to compare banners column percentages to the total column percentages. Higher percentages are displayed in blue and lower percentages in red. Minimum Difference Displayed Enter the minimum percentage necessary to have the differences displayed. The default is 10. Display Minimum Frequency This field allows you to display only side codes that are above a certain value. First N Codes This field allows you to display the top N codes in terms of frequency. A zero in this field means this option is not used. Note that this applies to regular side tables, not to grids. Applies Only To Banners Applies First N Codes only to banners. Alternate Row Color If checked, table rows will alternate colors between white and light yellow to make them more easily readable. Show Date If checked, the current date will display in the top left corner of each table. It will also appear at the bottom left of each chart. Show Logo If checked and a logo has been provided, the account's logo displays on the chart. (Email support@goascribe.com if you would like to provide a logo for your account.) If checked and a logo has not been provided, the Ascribe™ logo displays on the chart. If the field is not checked, no logo displays on the chart. Total Coded Respondents At the bottom of each table, this option displays the number of respondents have received at least one code. Total Uncoded Respondents At the bottom of each table, this option displays the number of respondents that did not receive a code. (The total of uncoded respondents and coded respondents should always equal the sample size.) Total Coded Responses Displays at the bottom of each table. This option acts like a global unduplicated net of all the responses that were coded. Underneath that number, you will find the average number of responses coded by respondent. Show Average for Numeric Questions If checked, whenever Cross-Tab encounters a numeric question (a question in which the responses are numbers), the average and standard deviation will be displayed for each banner point 222 Ascribe Presentations™ underneath the table. (The question must be coded in order to generate the deviations.) Note that whenever you check that option, Cross-Tab also generates a summary table of averages at the bottom of the Results page. Show Ranks Whether or not the tables are sorted, by selecting that option, Cross-Tab will add a number in bold to the left of each code and net giving you its order in terms of frequency. Text Column Width Defines the column width of the code text column. Values Column Width Defines the column width of the code values column. Font Size Specifies the font size of the code labels and code values displayed in each table. Title Specifies the title that will appear in blue and bold at the top of the Results page. By default, it contains the study ID and study name. Default Resets all of the options to their default values. Save Saves all options. Here is an example of a t-test: Filters Tab Use the Filters Tab to filter the tables. Click the Get Filters button to display a list of questions. To select a filter, click the check box next to a question, and then click a code that will act as the filter. 223 PrintDoc01 020813 The filter titles will appear on top of each table and underneath the title of each graph. If you use multiple filters, "And" will be displayed as + and "Or" as |. Note that there are two other ways to filter your set of tables: By double clicking any cell of your table By clicking the "Set As Filter" button in the drill down table window. Whatever the way you set the filters, you can remove them by clicking the Remove Filter button ( ) on the Results page. This button only displays when a filter is set. Results Page Click the Run button to generate all the tables and/or graphs you have requested. The Results page displays the tables and charts you created. You have several options for exporting them. Whatever export you use, you will be prompted with a dialog box that asks if you want to open or save the file. If you want to use a cell as a filter, simply double click it. Here are the options on the Results page: Option Description Exports tables to an HTML file. If you have also generated charts, the HTML file and the related charts will be placed in a .zip file. You will have to save and unzip the file on your computer. Exports tables to Excel ®. Each table will be in a separate worksheet. Exports tables and charts to Word (only available if Enable Export to Word is selected on the Options Tab.) Exports tables and charts to PowerPoint (only available if Enable Export to PowerPoint is selected on the Options Tab.) Append Appends the current results to the previously saved files. Creates a form so that you can email the results. Fill in the form, select the type of attachments you want to send, and click the Send button. A message displays when the email has been sent. Click Cancel to return to the Results 224 Ascribe Presentations™ page or Settings to return to the Settings page. Copies a table or chart to the clipboard so you can copy it to another application. To paste bar charts, line charts, or pie charts from the clipboard to Word, first open a Word document and then click the Copy to Clipboard button. In Word, you must select Paste Special rather than Paste. After you select Paste Special, select Device Independent Bitmap from the popup menu that appears, and then click OK. The chart will be copied to Word from the clipboard. Drill Down Options with Codes If the Allow Drilldown box is selected on the Options tab, the actual responses are available when you click a cell of a table or part of a graph. When you hover over table cell, the cell changes color to let you know more information is available. When you hover over areas in a graph, a small hand displays to let you know more information is available. Here is an example of the drill down page for a code in a table: You can also see all of a respondent's answers to the questions in the study by clicking a verbatim. Here is an example of the respondent drill down page: Multi-Media Files You can use the drill-down options on a verbatim that points to a multimedia file (image, audio or video.) When you click the file name, the appropriate player will start. 225 PrintDoc01 020813 Create Bar, Line, or Pie Charts Navigate: Supervisor/Studies/Right-click a study/Select Ascribe Presentations/Select a chart type Client/Studies/Right-click a study/Select Ascribe Presentations/Select a chart type The screens for bar, line and pie charts look similar, and the process to create them is similar. To create a chart, first select a question from the drop-down box at the top of the screen. The number to the right of each question is the number of responses/coded responses it contains, followed by the percent coded. Closed-ended questions are surrounded by square brackets. The codebook for the selected question displays on the left side of the screen. Select which codes and/or nets you want to graph. You can either use the check boxes to the left of each code and net or you can select all the codes and/or all the nets from the two check boxes located above the codebook: If you want to select just the codes of a certain net, uncheck All Codes and All Nets and click the check box for Select Siblings. Then, click the check box next to a code you wish to display. All siblings of that code will be selected automatically. You can only select one set of siblings at a time. The Sort Codes by Frequency option lets you display the list of codes in order from most used to least used. The Options pane is the right side of the screen. All options are not available for each type of chart. Note that all the settings (except the filters) will remain across Ascribe™ sessions. Here is a list of the options available for each type of chart: Option Chart Description Title All The title defaults to “X Chart,” where X equals the type of chart. It is followed by the question ID and question label. You can change the title. Width All Set the width of the chart in pixels or have the width automatically set with the default of zero. With the auto size option, the width will depend on the number of codes/nets used. 226 Ascribe Presentations™ Height Bar Sets the height of the chart in pixels. Line Orientation Bar Usually bar charts are displayed horizontally, but you may change it to vertical. Note that you can not display the average on vertically-oriented bar charts. Line charts must be horizontal. Display Frequencies All If checked, the frequency of each code/net displays on top of each bar and in the chart labels. Display Code Output ID All If checked, the code output ID of each code/net displays in the chart labels. (For pie charts, the code description displays when you hover over a pie piece. Click on a pie piece to see the drill down window, which has the output code ID.) Show Date All If checked, the date the chart was created displays on the chart. Show Logo All If checked and a logo has been provided, the account's logo displays on the chart. (Email support@goascribe.com if you would like to provide a logo for your account.) If checked and a logo has not been provided, the Ascribe™ logo displays on the chart. If the field is not checked, no logo displays on the chart. Minimum Frequency All This field allows you to display only codes or nets that are above a certain value. For example, a minimum frequency of 10 means to display only those codes or nets with a count of 10 or above. This field overrides the selection of codes or nets made on the codebook side of the screen. A zero in this field means display all codes or nets selected on the codebook side. First N Codes or Nets All This field allows you to display the top N codes/nets in terms of frequency. This field overrides the minimum frequency and the selection on the codebook side of the screen. A zero in this field means this option is not used. Bar This option is used when you are charting nets. It allows you to limit the nets to a specific level. For example, it may not make sense to compare nets across all levels as level 2 nets will always have lower counts than level 1 nets. A zero in this field means this option is not used. Net Level Line Codes to Ignore Bar Line You can enter either code text or code output IDs that you want to ignore in the charts. For example, you may want to ignore no answer codes. You can enter several text or values, but they must be separated by a 227 PrintDoc01 020813 semicolon. Example: Nothing; N.A.;98;99 Display Average Bar Line For horizontal charts, this option adds a line across all the bars which displays the average. It will also display its value. Sort Codes by Frequency Bar This option displays the bars in decreasing order of frequency. Note that if nets are being charted, the subnets and codes will be sorted within each parent net. 3D Bar If checked, the bars display in three dimensions. % All This option displays the figures on the top of each bar in percentages rather than in absolute figures. Percentage Decimals All The accepted values are 0, 1, or 2. If you enter any other number, the field will default to 1. Bar This field has a drop-down list of closed-ended questions. You can choose one in order to compare codes or nets across all the values of that closed-end. The chart will display a series of bars or lines for each code of the selected question. Graph By Line Stacked Bars Bar If you selected a question in the “Graph by” list, this field displays the bars as stacked on top of each other instead of next to each other. Filters Bar You can filter the data by closed-ended questions. If you check this box, a pane displays with closed-ended questions and their codes. From that pane, you can select any combination of the codes as a filter by using the And or the Or buttons. The code(s) used as filter(s) will display underneath the chart title in the chart. Line Click the OK button to generate a chart. The chart displays, and you can save or export it. You can return to the Ascribe Presentations main page by clicking the Ascribe Presentations button. Here is an example of the Options pane for a bar chart: 228 Ascribe Presentations™ Here is an example of a filter pane (in this case we are filtering on Spanish): Bar Chart Examples Navigate: Supervisor/Studies/Right-click a study/Select Ascribe Presentations/Select Bar Charts Client/Studies/Right-click a study/Select Ascribe Presentations/Select Bar Charts Bar charts are usually used to compare frequencies across codes and or nets. This section contains examples of common bar charts. Here is a simple unsorted bar chart with the average: 229 PrintDoc01 020813 Note that code texts display a maximum of 30 characters. To display the full text, hover above any of the bars, and a tool tip with it displays. When you create a chart with codes and nets, nets are represented with diagonal lines. When bar charts contain nets or codes of different levels, each level gets its own color. Here is an example: Line Chart Examples Navigate: Supervisor/Studies/Right-click a study/Select Ascribe Presentations/Select Line Charts Client/Studies/Right-click a study/Select Ascribe Presentations/Select Line Charts Line charts are used to display the evolution of data across time. Also, they can be a replacement for bar charts when displaying series data. An example of series data would be how codes differ from one country to another. Here is an example of a basic line chart: 230 Ascribe Presentations™ Here is the same example across three countries: Pie Chart Example Navigate: Supervisor/Studies/Right-click a study/Select Ascribe Presentations/Select Pie Charts Client/Studies/Right-click a study/Select Ascribe Presentations/Select Pie Charts Pie charts are used to compare proportions. For example, you can compare proportions of codes between each other or nets between each other. You can also compare sample sizes (country, gender, region, etc…) Here is an example of a simple pie chart: 231 PrintDoc01 020813 Co-Occurrence Maps Navigate: Supervisor/Studies/Right-click a study/Select Ascribe Presentations/Select Co-Occurence Maps Client/Studies/Right-click a study/Select Ascribe Presentations/Select Co-Occurrence Maps A co-occurrence map displays how frequently codes and/or nets were applied. It also displays which codes and/or nets were applied together the most frequently. The co-occurrence map uses circles and lines to show these frequencies. The circles represent codes and/or nets, and the code/net descriptions display above the circles. The more the code was used, the larger the circle will be. The lines between the circles represent how often the two codes or nets were both applied to responses. The thickness of the lines represents the number of responses that have both codes or nets applied. The higher the co-occurrence between two codes/nets, the thicker the line will be. Here is a simple example: 232 Ascribe Presentations™ The co-occurrence screen is similar to the other chart screens, but has different options. To create a co-occurrence chart, first select a question from the drop-down box at the top. You can either use the check boxes to the left of each code and net or you can select all the codes and/or all the nets from the two check boxes located above the codebook: If you want to select just the codes of a certain net, uncheck All Codes and All Nets and click the check box for Select Siblings. Then, click the check box next to a code you wish to display. All siblings of that code will be selected automatically. You can only select one set of siblings at a time. The Sort Codes by Frequency option lets you display the list of codes in order from most used to least used. (The + check box is used to create maps across questions.) The last step is to set the options on right side of the screen. Here are the options for a co-occurrence map: Option Description Title The title defaults to “Co-Occurrence Map,” followed by the question ID and question label. You can change the title. Graph Size Sets the diameter of the circle in pixels. Margin Size Sets the margin around the map in pixels. The margins are the main containers for the code and net labels. When you change the margin size, only the side margins are affected. Largest Circle Size Sets the size of the largest circle (the one with the highest frequency.) All other circles’ sizes will be proportional to the largest 233 PrintDoc01 020813 one. Code Font Size Sets the font size for the code and net labels. Cooc Font Size Sets the font size for the numbers on the lines. Largest Line Width Sets the width of the largest line (the one with the highest cooccurrence.) All other lines’ sizes will be proportional to the largest line. Rescale Line Widths Increases the thickness of the lines. For example, we have cooccurrences ranging between 50 and 70. In this example, the thinnest co-occurrence line will be 5/7 of the largest one. The rescale line widths option tells the map to define 50 as 1 and 70 as 70. Display Frequencies If checked, the frequency of each code/net displays on top of each circle and in the map labels. Max Characters per Enter the maximum number of characters you want to display for Code each code. Display Code Output ID If checked, the code output ID of each code/net displays in the map labels. Show Date If checked, the date the chart was created displays on the chart. Show Logo If checked and a logo has been provided, the account's logo displays on the chart. (Email support@goascribe.com if you would like to provide a logo for your account.) If checked and a logo has not been provided, the Ascribe™ logo displays on the chart. If the field is not checked, no logo displays on the chart. Minimum Frequency This field allows you to display only codes or nets that are above a certain value. For example, a minimum frequency of 10 means to display only those codes or nets with a count of 10 or above. This field overrides the selection of codes or nets made on the codebook side of the screen. A zero in this field means display all codes or nets selected on the codebook side. Minimum CoOccurrence This field allows you to display only codes or nets that have cooccurrences above a certain value. A zero in this field means display all codes or nets selected on the codebook side. First N Codes or Nets This field allows you to display the top N codes/nets in terms of frequency. This field overrides the minimum frequency and the selection on the codebook side of the screen. A zero in this field means this option is not used. First N CoOccurrences Use this option to only display co-occurrences. It does not display the codes’ relative importance to each other. This option calculates all of the co-occurrences and displays the first N ones, where N equals a number that you enter. Hide Codes/Nets with No CoOccurrence This option displays only those codes/nets that have co-occurrence with other codes/nets. Net Level This option is used when you are charting nets. It allows you to limit 234 Ascribe Presentations™ the nets to a specific level. For example, it may not make sense to compare nets across all levels as level 2 nets will always have lower counts than level 1 nets. A zero in this field means this option is not used. Codes to Ignore You can enter either code text or code output IDs that you want to ignore in the charts. For example, you may want to ignore no answer codes. You can enter several text or values, but they must be separated by a semicolon. Example: Nothing; N.A.;98;99 Sort Codes by Frequency This option displays the codes in decreasing order of frequency. Note that if nets are being charted, the subnets and codes will be sorted within each parent net. Colors Set the colors of the circles and lines. When you click one of the color buttons, a color screen displays. Click any color, and it displays next to the OK button. Click OK to confirm or Cancel to ignore. Filters You can filter the data by closed-ended questions. If you check this box, a pane displays with closed-ended questions and their codes. From that pane, you can select any combination of the codes as a filter by using the And or the Or buttons. The code(s) used as filter(s) will display underneath the chart title in the chart. Click the OK button to generate the chart. The chart displays, and you can save or export it. You can return to the Ascribe Presentations main page by clicking the Ascribe Presentations button. Here is the Option pane for the co-occurrence screen: 235 PrintDoc01 020813 Here is the color screen: Correlation Matrices Navigate: Supervisor/Studies/Right-click a study/Select Ascribe Presentations/Select Correlation Matrices Client/Studies/Right-click a study/Select Ascribe Presentations/Select Correlation Matrices Correlation matrices allow you to look at the strength and direction of a relationship between two variables (codes and/or nets in our case.) While co-occurrence maps provide a simple count of variables, correlation matrices examine the positive, neutral, or negative relationship between the variables. The mathematical formula for correlation is applied, and 236 Ascribe Presentations™ a number between -1 and +1 is displayed. A 1 means the variables are totally correlated, 0 means no correlation at all, and -1 means inversely correlated. A positive correlation (1 or greater than .04) between two codes means that, in general, when one code is applied, the other code is also applied. For example, there is likely to be a positive correlation between these two questions: "Did it rain today?" and "Did you take your umbrella?" A negative correlation (-1 or close to -1) between two codes means that, in general, when one code is applied, the other code is not applied. For example, there is likely to be a negative correlation between these two questions: "Do you smoke?" and "Do you run marathons?" No correlation (0 or close to 0) between two codes means that there is no relationship between the codes. One code cannot predict the outcome of the other, and they are considered statistically independent. For example, the questions "Do you like Mozart?" and "Do you run marathons?" have nothing to do with each other. To create the matrices, first select a question from the drop-down box at the top. You can either use the check boxes to the left of each code and net or you can select all the codes and/or all the nets from the two check boxes located above the codebook: If you want to select just the codes of a certain net, uncheck All Codes and All Nets and click the check box for Select Siblings. Then, click the check box next to a code you wish to display. All siblings of that code will be selected automatically. You can only select one set of siblings at a time. The Sort Codes by Frequency option lets you display the list of codes in order from most used to least used. (The + check box is used to create maps across questions.) The last step is to set the options on right side of the screen. The Options pane has many of the same fields as the Co-Occurrence Option Pane, with one addition. The field Display Correlation gives the user the option to display the actual correlation number on each square. The colors in the matrices indicate the correlation. White means uncorrelated or that the correlation = 0. The darker the blue, the more the variables are correlated. The darker the red, the more the variables are inversely correlated. As with the other charts, you can hover above any square, and a tool tip displays the code text. Also, you can click a square and see the responses or respondents involved. Here is the Options pane for correlation matrices: 237 PrintDoc01 020813 Here is an example of a correlation matrix with the correlation numbers on each square: Co-Occurrence Maps and Correlation Matrices Across Questions Navigate: Client/Studies/Right-click a study/Select Graphs/Select Co-Occurrence 238 Ascribe Presentations™ To create a co-occurrence map or a correlation matrix between codes across questions, click the + check box located on the top of the screen: Next, select a question. Every time you select a question, that question’s codes are added to the check list below the question box. In this example, the country and gender questions are selected: The next step is to select the set of codes. Here is an example of where the three country codes and the two gender codes were selected: Proportional Code Font Views Navigate: Supervisor/Studies/Right-click a study/Select Ascribe Presentations/Select Proportional Code Font Views Client/Studies/Right-click a study/Select Ascribe Presentations/Select Proportional Code Fonts View 239 PrintDoc01 020813 Proportional code font charts enable the user to display a codebook by adjusting the font size and/or the font color depth based on each code's frequency. To create a chart, first select a question from the drop-down box at the top. You can either use the check boxes to the left of each code and net or you can select all the codes and/or all the nets from the two check boxes located above the codebook: If you want to select just the codes of a certain net, uncheck All Codes and All Nets and click the check box for Select Siblings. Then, click the check box next to a code you wish to display. All siblings of that code will be selected automatically. You can only select one set of siblings at a time. The Sort Codes by Frequency option lets you display the list of codes in order from most used to least used. The last step is to set the options on right side of the screen. Many of the options are the same as other charts. Here are the options unique to Proportional Code Font Charts: Status Description Min Font Size and Max Font Size The min font size will be applied to the code/net with the smallest frequency. The max font size will be applied to the code/net with the largest frequency. All other codes/nets will have a font size somewhere in between based on its frequency. If you do not want to use proportional font sizes (and only use the color depth), just give the same value to both the min and the max font sizes. Display First N Characters Use this field to shorten long code descriptions. If the field is 0, the entire code/net description text displays; otherwise, the first N characters display. Adjust Font Density If checked, the codes/nets with the largest frequency display in black. The smaller the frequency, the more they display towards the light gray color. In this example, only the font size is proportional to the frequency (and the list is sorted): In this example, only the font density is proportional to the frequency: 240 Ascribe Presentations™ In this example, both the font density and size are proportional to the frequency: As with the other charts, drill down information is available when you click any code. A window displays with all the responses/respondents where that code has been applied. Multidimensional Scalings Navigate: Supervisor/Studies/Right-click a study/Select Ascribe Presentations/Select Multidimensional Scalings Client/Studies/Right-click a study/Select Ascribe Presentations/Select Multidimensional Scalings Multidimensional Scaling (MDS) are a way to graphically represent a correlation matrix in a two-dimensional space. It enables the user to quickly spot which codes/nets are similar or dissimilar. Similar objects (those with a positive correlation) are represented by points that are close to each other. Dissimilar objects (those with a negative correlation) are represented by points that are far apart. To create a map, first select a question from the drop-down box at the top. You can either use the check boxes to the left of each code and net or you can select all the codes and/or all the nets from the two check boxes located above the codebook: If you want to select just the codes of a certain net, uncheck All Codes and All Nets and click the check box for Select Siblings. Then, click the check box next to a code you wish to display. All siblings of that code will be selected automatically. You can only select one set of siblings at a time. The Sort Codes by Frequency option lets you display the list of codes in order from most used to least used. The + check box is used to create maps across questions. To create a map across questions, click the + check box located on the top of the screen. Next, select a question. 241 PrintDoc01 020813 Every time you select a question, that question’s codes are added to the check list below the question box. The last step is to set the options on right side of the screen. Here is the Options pane: Most of the options are the same as other charts. The only unique option is Display Legends. If many codes/nets are mapped, their text may overlap on the map. Display Legends puts the code output IDs on the map, and the code descriptions in a column to the right of the map. If you do not select Display Legends, the code descriptions display with the code output IDs on the map. In this example, Display Legends has been selected: In this example, Display Legends has not been selected: 242 Ascribe Presentations™ Clusters Navigate: Supervisor/Studies/Right-click a study/Select Ascribe Presentations/Select Clusters Client/Studies/Right-click a study/Select Ascribe Presentations/Select Clusters Cluster analysis is a branch of statistics called "multivariate statistics," which means multiple variables are examined simultaneously. Multivariate statistics is often called 'exploratory data analysis' as its main purpose is to explore the data. When doing exploratory data analysis, sometimes we find it useful and sometimes not. In Ascribe™, we are applying cluster analysis to codes. More often, it is done on respondents. When done on respondents, cluster analysis tries to create groups of respondents that are similar by trying to maximize the similarities within a group while maximizing the dissimilarities across groups. "Similar" means that they tend to respond likely to several responses (usually metric responses such as a scale from 1 to 5.) However, here we are trying to create groups of codes that are similar. "Similar" in this case means that they have a high correlation coefficient. The correlation coefficient is a number between -1 and +1. Anything around 0 means that there is no correlation, and anything close to 1 means that two codes are highly correlated. So, the data used to create the clusters of codes is the correlation matrix between codes (all the possible pairs of codes' correlations.) The algorithm used to assign a code to a cluster is very simple: Take n codes that are highly uncorrelated at random to create n initial clusters. (You select the number of codes and number of clusters in the Options pane on the right side of the screen.) Parse all the other codes and assign them to the cluster to which their average correlation is the highest with all the members of that cluster. Here is an example of a cluster chart: 243 PrintDoc01 020813 In the example, the Cluster 1 has 2 codes (n=2), and an average correlation of 0.58 (r=0.58). It means that the average correlation between each pair of codes within that cluster is 0.58. In Cluster 2, the average correlation between each pair of codes is 0.56. The codes within this cluster are slightly less correlated than the codes in Cluster 1. The Correlation Across Clusters indicates the average correlation between each code ACROSS clusters. Because clustering is exploratory data analysis, you may need to run the cluster report more than once to get the most useful results. For example, if you create three clusters, and the average correlations are close to zero, you might run the report again with only two clusters. You may find it useful to create a correlation matrix for your data first to see if there are strong correlations. A cluster chart may not be helpful for all data if there are not strong correlations. When you create a cluster report, use either codes or nets. If you only use nets, make sure that the nets are at the same level. For example, compare nets to nets, sub-nets to subnets, etc. To create a cluster report, first select a question from the drop-down box at the top. You can either use the check boxes to the left of each code and net or you can select all the codes or all the nets from the two check boxes located above the codebook: If you want to select just the codes of a certain net, uncheck All Codes and All Nets and click the check box for Select Siblings. Then, click the check box next to a code you wish to display. All siblings of that code will be selected automatically. You can only select one set of siblings at a time. The Sort Codes by Frequency option lets you display the list of codes in order from most used to least used. The last step is to set the options on right side of the screen. 244 Ascribe Presentations™ Multiple Charts Navigate: Supervisor/Studies/Right-click a study/Select Ascribe Presentations/Select Multiple Charts Client/Studies/Right-click a study/Select Ascribe Presentations/Select Multiple Charts You can create one graph for several questions at one time through the Multiple Charts option. The Multiple Charts window is similar to the other chart screens, but has some different options. In the drop-down box, select the chart type. Then, click the box next to the questions you want to chart. (Closed-ended questions are displayed between square brackets.) In the Options pane on the right side, select options for the chart. Most options are the same as those found on the other chart screens, but there are some differences. (See Create a Bar, Line, or Pie Chart, Co-Occurrence Maps, Correlation Matrices, Proportional Code Font Views, and Multidimensional Scalings for more information about the options.) For multiple charts, you do not have access to the list of codes for each question. Therefore, in the Options pane, you need to select whether you want to chart codes and nets, codes only or nets only. Also, you have the option to generate 1 chart per net. The Filters option provides a way to create multiple charts with the Loop Filters option. Click the Filters check box, and another section displays with filtering options. Select one or more filtering questions and the desired codes and select Loop Filters. One chart will be generated for each code. See Save and Export Charts for export options and Drill Down Tools for drill down options. Word Clouds Navigate: Supervisor/Studies/Right-click a study/Select Ascribe Presentations/Select Word Clouds Client/Studies/Right-click a study/Select Ascribe Presentations/Select Word Clouds The Word Clouds feature generates tables of words arranged in a special way so that words with the largest frequencies appear in larger font. The Settings page has a toolbar with these fields: Field Description Run This option creates the word cloud and navigates to the Results page. Click this button after you have selected the questions and various options. Settings This option navigates from the Results page to the Settings page. (It is grayed out on the Settings page.) Study ID: Study Name This field has a drop-down box with study information. 245 PrintDoc01 020813 Ascribe Presentations This button navigates to the Ascribe Presentations main page. The right-click menu also provides navigation to the Ascribe Presentations main page or other areas of Ascribe™. Questions Tab The Questions Tab contains a list of study questions. You must select at least one question in order to generate a word cloud. The Word Cloud module will generate one word cloud for each question checked here. Choose from these options: Option Description Questions Select as many questions as you would like. If you want to select or unselect all side questions, click the check box next to the Questions heading. Options: - Use the Up and Down buttons to move a selected question higher or lower in the list. Word Clouds By This drop-down box displays all the closed-ended questions of the current study. By selecting a question there, you are asking to generate one word cloud filtered by each of the codes of the selected closed-end. Each question has this format: Field Description When you click this symbol, a drop-down dialog displays codebook information and the list of codes. Use the double arrow button ( dialog. or ) to collapse the upper or lower panes of the Click the check box to select or unselect a question. Question ID.Question Label The question ID displays, followed by the question label (if the label is present.) Closed-ended questions have brackets around these fields. (Number) The number of items in a codebook. 246 Ascribe Presentations™ (Percentage) The percentage of responses that have been coded in the question. Words Tab You can select word-related options on this tab. Here are the options: Option Description Response Source Select to display: Verbatim Translation Transcription Use Verbatim if no Translation/Transcription Words To Link You can link two or more words that are found next to each other in the data. Separate the set of linked words with a semicolon. For example: new york; fort worth; long beach When you link words, the individual words may display when they are used in a different context. For example, if you link new york, any other instances of the word new not used next to the word york will display separately. Words To Combine You can combine instances of similar words into the count of one word. All words will be represented by the first word in the set. Separate the set of combined words with a semicolon. For example: frizz frizzy frizziness; great wonderful The drill down in the word cloud will replace the combined words with the first word in the set. Min Word Length Words that are equal to or less than the number entered here will be ignored. Words To Ignore A default list of words appears in the text box. The list is optional; you can unselect it with the check box. Also, you can change the list to whatever you want (be sure to use semicolons to separate the words.) To restore the default ignore list, click the Default button at the bottom of the Options Tab. Options Tab This tab contains a variety of options for creating word clouds. The options are: Option Description General Options 247 PrintDoc01 020813 HTML View You have the choice to generate an HTML view or a JPEG view or both. The words can be sorted alphabetically or by frequency or you can choose no sorting (randomly.) JPEG View For the JPEG view, there are options to sort the words by frequency or to just have them in a cluster. The words can display with no shape, in a rectangle, or in a capsule. You can change the graph width if you wish. Also, you can change the graph's background color, the items' background color, and the items' color. Allow Drilldown Allows you to click word and look at all the responses with that word. A separate dialog opens with the responses and allows export to HTML or Excel®. Also, you can export all data for these respondents (all of their verbatim in the study) to a text file. If you click the "Set as Filter" button, it will set the current respondents as a filter for all your tables. Remember to refresh your tables by clicking "Tables" button after the drill down window has closed. Enable Export To Word Allows export to Word. If enabled, this icon displays at the top of the Enable Export To PowerPoint Allows export to PowerPoint. If enabled, this icon displays at the top Results page: of the Results page: Display First N Words This field allows you to display the top N words in terms of frequency. A zero in this field means this option is not used. Minimum Frequency This field allows you to display only words that are above a certain value. Use To Scale Items Use the Minimum Frequency field to display the words in a more proportional way. Show Date If checked, the current date will display at the top of each word cloud. Show Logo If checked and a logo has been provided, the account's logo displays on the chart. (Email support@goascribe.com if you would like to provide a logo for your account.) If checked and a logo has not been provided, the Ascribe™ logo displays on the chart. If the field is not checked, no logo displays on the chart. Title Specifies the title that will appear in blue and bold at the top of the Results page. By default, it contains the study ID and study name. Frequencies If checked, the number of times a word occurs in the data will display next to the word. Percentages If checked, the percentage of times a word occurs in the data will display next to the word. Percentage The accepted values are 0, 1, or 2. If you enter any other number, 248 Ascribe Presentations™ Decimals the field will default to 1. Allow Find & Replace If checked, the Find and Replace option displays on the Results Page. It is allowed for HTML views only. Default Resets all options to their default values. Save Saves all options. Filters Tab Use the Filters Tab to filter the word clouds. Click the Get Filters button to display a list of questions. To select a filter, click the check box next to a question, and then click a code that will act as the filter. Note that there are two other ways to filter your set of tables: By double clicking any word in the word cloud By clicking the "Set As Filter" button in the drill down table window. The filter titles will appear on top of each word cloud underneath the title. If you use multiple filters, "And" will be displayed as + and "Or" as |. You can remove the filters by clicking the Remove Filter button ( This button only displays when a filter is set. ) on the Results page. Results Page Click the Run button to generate the word clouds you have requested. The Results page displays the word clouds you created. You have several options for exporting them. Whatever export you use, you will be prompted with a dialog box that asks if you want to open or save the file. Here are some features of word clouds: 249 PrintDoc01 020813 If the Allow Drilldown box is selected on the Options tab, the actual responses are available when you click a word in the cloud. If you want to use a word as a filter, simply double click it. You can ignore a word in the cloud if you ctrl-click it. The cloud will be rebuilt without that word. This feature can be useful if there are extraneous words like "nothing" in the word cloud. Here are the options on the Results page: Option Description Exports tables to an HTML file. Exports tables and charts to Word (only available if Enable Export to Word is selected on the Options Tab.) Exports tables and charts to PowerPoint (only available if Enable Export to PowerPoint is selected on the Options Tab.) Append Appends the current results to the previously saved files. Creates a form so that you can email the results. Fill in the form, select the type of attachments you want to send, and click the Send button. A message displays when the email has been sent. Click Cancel to return to the Results page or Settings to return to the Settings page. Copies a word cloud to the clipboard so you can copy it to another application. Find & Replace If selected on the Options tab, the Find and Replace text boxes are available for HTML views. If you enter a search string in the Find box and no search string in the Replace box, all occurrences of the search string in the cloud will be highlighted. If you enter a search string in the Find box and a replacement string in the Replace box, the find string will be replaced. To remove the highlight from the word cloud, clear the Find box and click the Find button. Color-Coded Words The colors in the word cloud indicate its frequency. You can also hover over a word to see its frequency. Here is a guide to the colors: Color Description Red The frequency of the word is in the top 80%. Blue The frequency of the word is in the 60-80% range. Green The frequency of the word is in the 40-60% range. Black The frequency of the word is in the 20-40% range. Gray The frequency of the word is below 20%. 250 Ascribe Presentations™ Here is an example of a word cloud: Drill Down Options The Allow Drilldown box must be selected on the Options Tab in order to see more information. The drill down option for word clouds has the same features as the drill down option for cross-tabs with words. Just click a word in the cloud to see all of the responses that contain that word. For more information, see Drill Down with Words. Concept Clouds Navigate: Supervisor/Studies/Right-click a study/Select Ascribe Presentations/Select Concept Clouds Client/Studies/Right-click a study/Select Ascribe Presentations/Select Concept Clouds The Concept Clouds takes a codebook and displays it in a word cloud based on the frequency of each code. Any codebook can be used, including concept questions generated with Ascribe Inspector. The Settings page has a toolbar with these fields: Field Description Run This option creates the concept cloud and navigates to the Results page. Click this button after you have selected the questions and various options. Settings This option navigates from the Results page to the Settings page. (It is grayed out on the Settings page.) Study ID: Study Name This field has a drop-down box with study information. Ascribe Presentations This button navigates to the Ascribe Presentations main page. The right-click menu also provides navigation to the Ascribe Presentations main page or other areas of Ascribe™. 251 PrintDoc01 020813 Questions Tab The Questions Tab contains a list of study questions. You must select at least one question in order to generate a concept cloud. The Concept Cloud module will generate one concept cloud for each question checked here. Choose from these options: Option Description Questions Select as many questions as you would like. If you want to select or unselect all side questions, click the check box next to the Questions heading. Options: - Use the Up and Down buttons to move a selected question higher or lower in the list. Concept Clouds By This drop-down box displays all the closed-ended questions of the current study. By selecting a question there, you are asking to generate one concept cloud filtered by each of the codes of the selected closed-end. Each question has this format: Field Description When you click this symbol, a drop-down dialog displays codebook information and the list of codes. Use the double arrow button ( dialog. or ) to collapse the upper or lower panes of the Click the check box to select or unselect a question. Question ID.Question Label The question ID displays, followed by the question label (if the label is present.) Closed-ended questions have brackets around these fields. (Number) The number of items in a codebook. (Percentage) The percentage of responses that have been coded in the question. Options Tab This tab contains a variety of options for creating concept clouds. The options are: Option Description General Options Allow Drilldown Allows you to click a code and look at all the responses that received the code. A separate dialog opens with the responses and allows export to HTML or Excel ®. HTML View You have the choice to generate an HTML view or a JPEG view or 252 Ascribe Presentations™ both. JPEG View For the JPEG view, there are options to sort the words by frequency or to just have them in a cluster. The words can display with no shape, in a rectangle, or in a capsule. You can change the graph width if you wish. Also, you can change the graph's background color, the items' background color, and the items' color. Enable Export To Word Allows export to Word. If enabled, this icon displays at the top of Enable Export To PowerPoint Allows export to PowerPoint. If enabled, this icon displays at the Codes/Concepts To Ignore You have the option to enter codes to ignore in the concept cloud. Click the check box and enter the code description of any codes you wish to ignore. Be sure to use semicolons to separate the codes. the Results page: top of the Results page: Display First N Concepts This field allows you to display the top N concepts in terms of frequency. A zero in this field means this option is not used. Minimum Frequency This field allows you to display only codes that occur above a certain value. Use To Scale Items Use the Minimum Frequency field to display the codes in a more proportional way. Max Characters Per Concept A concept or code description may display up to 50 characters. You can select any number up to 50. Show Date If checked, the current date will display at the top of each concept cloud. Show Logo If checked and a logo has been provided, the account's logo displays on the chart. (Email support@goascribe.com if you would like to provide a logo for your account.) If checked and a logo has not been provided, the Ascribe™ logo displays on the chart. If the field is not checked, no logo displays on the chart. Title Specifies the title that will appear in blue and bold at the top of the Results page. By default, it contains the study ID and study name. Frequencies If checked, the number of times a code occurs in the data will display next to the code. Percentages If checked, the percentage of times a code occurs in the data will display next to the code. Percentage Decimals The accepted values are 0, 1, or 2. If you enter any other number, the field will default to 1. Default Resets all options to their default values. 253 PrintDoc01 020813 Save Saves all options. Filters Tab Use the Filters Tab to filter the concept clouds. Click the Get Filters button to display a list of questions. The filter titles will appear on top of each concept cloud underneath the title. If you use multiple filters, "And" will be displayed as + and "Or" as |. You can remove the filters by clicking the Remove Filter button ( This button only displays when a filter is set. ) on the Results page. Results Page Click the Run button to generate the concept clouds you have requested. The Results page displays the concept clouds you created. You have several options for exporting them. Whatever export you use, you will be prompted with a dialog box that asks if you want to open or save the file. If the Allow Drilldown box is selected on the Options tab, the actual responses are available when you click a code in the cloud. Here are the options on the Results page: Option Description Exports tables to an HTML file. Exports tables and charts to Word (only available if Enable Export to Word is selected on the Options Tab.) Exports tables and charts to PowerPoint (only available if Enable Export to PowerPoint is selected on the Options Tab.) Append Appends the current results to the previously saved files. Creates a form so that you can email the results. Fill in the form, select the type of attachments you want to send, and click the Send button. A message displays when the email has been sent. Click Cancel to return to the Results page or Settings to return to the Settings page. Color-Coded Words The colors in the word cloud indicate its frequency. You can also hover over a word to see its frequency. Here is a guide to the colors: Color Description Red The frequency of the word is in the top 80%. Blue The frequency of the word is in the 60-80% range. Green The frequency of the word is in the 40-60% range. Black The frequency of the word is in the 20-40% range. 254 Ascribe Presentations™ Gray The frequency of the word is below 20%. Drill Down Options The Allow Drilldown box must be selected on the Options Tab in order to see more information. Click a code in the cloud to see all of the responses that received that code. In that table of responses, click the Code Applied field to see all of a respondent's answers in the study. Cross-Tab (Legacy) Navigate: Supervisor/Studies/Right-click a study/Select Ascribe Presentations/Select Cross-tab (legacy) Client/Studies/Right-click a study/Select Ascribe Presentations/Select Cross-tab (legacy) Ascribe™ provides an easy-to-use cross-tabulated report for simple analysis of the study results. The report is primarily intended for account executives and clients, but is also useful for the tabulation person. The Cross-Tab Report creates cross-tabs from the collected data. You may cross-tab closed against closed, open against open, and any combination of the two. Note This is a generalized cross-tab, and you may cross-tab any question by any other question. The resulting cross-tab has some special characteristics: Drill Down. You may click any cell and see all of the underlying data, both open and close ended, that make up that cell. You may sort the display by any of the columns, and you may export the data to Excel™ if you click the Excel™ icon. Filtering. Click any row, column, cell or any combination of rows, columns and cells, and you will define the respondents for your next cross-tab. As an example, let’s say you want to run a report for just those respondents who are "Top Two Box." Just highlight the Definitely Would, and Probably Would Buy columns, press the filter button and then run your next cross-tab. It is like a cross-tab within a cross-tab. Convert to Excel™. Click the Excel™ icon to copy your cross-tab to Excel™. To create the report, click the question menu. The first question you choose will be in the row, the next question will be in the column. All subsequent questions will form additional dimensions in the rows. 255 PrintDoc01 020813 Select Cross-Tab Questions To start your analysis, select a question from the drop-down list. Questions are listed by the Question ID and Question label and also display the question type and the count of coded responses. A question will appear in the list if it satisfies both these requirements: The Cross-tab check box is set in the Tasks Tab of the question At least one response to the question is coded. The question you select will appear in a table with two columns. The codes are listed in the first column, and the second is reserved for the counts. 256 Ascribe Presentations™ You can select additional questions as well. After the first question, additional questions are listed in columns. Questions with large codebooks are not good candidates for display in the columns. These take a long time to display. Tip: You should normally select the question with the largest codebook first. It takes longer to display codebooks in columns than in rows. Fill the Cross-Tab Table with Data Once you have displayed one or more questions, you can fill the table with data by clicking the Fill button. The format of the data depends on your selection in the second drop-down list. The default is Count. Format Description Count Displays the count of responses with codes applied that match the row and column headers. For nets, the counts are unduplicated totals of all child codes. This means that if a given response had more than one subcode applied, it will appear only once in the net total (not once for each code applied). This means that the net counts can be less than the total of the counts for the sub-codes. The count at the upper left corner of the table is the table total. This is the unduplicated number of responses, which is the number of unique respondents to the selected questions. Percent of Table Total This is the Count value divided by the table total, displayed as a percentage and rounded down. Percent of Column Total This is the Count value divided by the column total, displayed as a percentage and rounded down. The column total is the count of unique respondents for the code or net in the column header. This will always total 100% across the column headers. Percent of Row Total This is the Count value divided by the row total, displayed as a percentage and rounded down. The row total is the count of unique respondents for the code or net in the row header. This will always total 100% across the row headers. Count and Percent of Table Total Displays both the Count value and the Percent of table total in each cell. Count and Percent of Column Total Displays both the Count value and the Percent of column total in each cell. Count and Percent of Row Total Displays both the Count value and the Percent of row total in each cell. Expand and Collapse Rows You can show or hide rows in the table by clicking the + and - icons next to each row. Collapsed rows are also hidden when you print the page. If you export to Excel™, all rows are exported regardless of whether they are hidden. 257 PrintDoc01 020813 Drill Down to View Responses To view the responses for a cell, click on the text in the cell. You must click on the text, not on the blank part of the cell. The drill down view shows you the responses for the selected cell, and the codes applied. All questions in the study are displayed in the drill down view. If a cell contains more than about 1,000 responses, the drill down view may not show all responses. The drill down view appears in a popup window. At the top of the window are selection boxes for what to display in the window. Filter the Cross-Tab View You can filter the view by selecting codes or cells in the table. To filter by a code, click the text of the code. This selects the entire row or column for the filter. To filter by a cell, click in the cell anywhere outside of the text displayed. This selects only the cell. You can select as many rows, columns, and cells as you like for your filter. Note that the description at the top of the table changes to describe the filter. Selections in a single filter are OR'ed together. To apply the filter, click the Filter button. The table disappears, but the filter description remains. Now construct another cross-tab. When you fill this cross-tab, it will contain only those responses that meet the filter condition. You can filter repeatedly. Each filter is added to any previous filters. Successive filters are AND'ed together. Note that if you select both a net and a sub-code of the net as a filter, only the net filter applies. Save and Export Charts When you generate a chart, the chart displays on a new page. Use the buttons below the chart title to copy, save, and export the chart. 258 Ascribe Presentations™ Buttons Description Settings Button Use this button to return to the previous page (the settings page.) Do not use the back button on the browser. Copy to Clipboard Use this button to copy the chart to the clipboard. You can save the file or open it with your default picture file viewer when you click the button. This action opens a file download dialog. To paste bar charts, line charts, or pie charts from the clipboard to Word, first open a Word document and then click the Copy to Clipboard button. In Word, you must select Paste Special rather than Paste. After you select Paste Special, select Device Independent Bitmap from the popup menu that appears, and then click OK. The chart will be copied to Word from the clipboard. Copy to Word or PowerPoint Use this button to add the chart to Word or PowerPoint. (This button directly copies the chart to Word or Powerpoint, without using the clipboard.) When you click the button, a progress circle ( ) displays for a few seconds, and then the Word and PowerPoint icons are enabled. If you click either one, a dialog box displays, and you can save or view the file in the appropriate format. Here is an example of the file in PowerPoint: 259 PrintDoc01 020813 Note: If you go back to settings, generate another chart, and click the button again, the newly generated chart will be APPENDED to the existing Word and PowerPoint documents. If you want to create a new Word document or PowerPoint presentation, click the button. Make sure you save the files first. When you use the batch charts option, you generate multiple graphs. You can export all of the graphs or select individual ones with the clipboard button and check box. Each individual chart has a clipboard button and check box, like this: Drill Down Tools Ascribe Presentations display summary information about codes and nets. You can see detailed information about the codes with the drill down tools. The drill down tools are available for each type of chart. After you generate a chart, hover above a bar, a pie slice, a line, or a circle or number in a co-occurrence chart. A tool tip displays, and the cursor changes to a pointer or hand like this: 260 Ascribe Presentations™ Click the left mouse button to display the drill down window and show the details of all the responses and respondents coded in the current area. Here is an example of the drill down window: Underneath the Drill Down Results title, the code text displays in blue and the frequency in red. When you click a number on a co-occurrence line, the drill down results screen displays two codes applied columns. Here is an example: 261 Ascribe Inspector™ Ascribe Inspector™ allows for automatic exploration of unstructured data for sentiment analysis and concept identification. Here are some highlights: Use coded or uncoded data Generate sentiment analysis with drill-down capabilities Search data for words or phrases Identify concepts, people, and products Create a codebooks and apply codes Generate charts through Inspector or Ascribe Presentations Use Review Responses to further explore data. How to use Inspector 1. Have inspector privilege. 2. Understand inspection ownership. 3. Create an inspection for a study question or create an inspection from scratch. 4. Add variables for filtering if necessary. 5. Load data to an inspection as necessary. 6. Use the Inspection page to run analysis and see results. Inspector Privilege Navigate: Administrator/Associates/Right-click associate/Select Edit In order to use Inspector, your account must have an Inspector license. Please contact support@goascribe.com if you have questions about the license. Next, the associate must have inspector privilege. An administrator of the account can assign inspector privilege by going to the Associates page. Right-click the associate, and select Edit. A dialog opens, and the administrator can select the inspector privilege. If an associate only has inspector privilege and no other privilege, the associate will have limited access to Ascribe. The associate will have access to the General menu options, Inspector, Ascribe Presentations, Accelerator (if licensed), and Illustrator (if licensed.) Inspection Ownership Some rules about inspection ownership: An inspection has an owner, which defaults to the user who created the inspection. A user with administrator privilege can change the owner of an inspection to any user who has inspector privilege. An inspection can be deleted, opened, or modified only by the owner. Only the owner can create an illustration from an inspection. When viewing the Inspections page, users see only those inspections that they own. Administrators can see all inspections in the list. Inspector/Inspections Page The Inspector/Inspections page lists all inspections owned by you. (If you have administrator privilege, all inspections on the account display.) 263 PrintDoc01 020813 Inspections Page Layout The Inspections page has these columns: Column Description Created The date the inspection was created. Inspection ID Unique identifier for the inspection. Description Optional information about the inspection. Owner This column only displays if you have administrator privilege. It lists the owner of the inspection. Study Source If an inspection was created from a study question, the study name and question are listed. If the inspection was created from scratch, the word None displays. Comments The number of comments/responses in the inspection. You can filter the list of inspections by using the filter text box in the gray bar at the top of the page. Enter a character string in the box and press Enter to only display inspection IDs that begin with those characters. Enter an * before the characters to display inspection IDs that contain those characters. To remove the filter, delete the characters and press Enter. Inspections Page Right-Click Menu The Inspections page has a number of options for working with inspections. Right-click an inspection to see these options: Option Description Open Opens to the inspection's General Information page. New Inspection Creates an inspection from scratch. Properties Opens the inspection's property dialog where changes can be made. Delete Deletes the inspection. Variables Opens the inspection's Variables page. Ascribe Presentations Navigates to Ascribe Presentations, where you can create data visualizations of the inspection. Load Data Opens the Load Data dialog where you can start the loading process. Scripted Output Navigates to the Scripted Output page where you can create reports about the inspection. (Must have supervisor privilege.) Illustrator Navigates to Ascribe Illustrator, which is another data 264 Data Management visualization tool. You must have an Illustrator license and privilege to use the product. Choose Columns Opens a dialog where you can choose what columns to display. Note that the Owner column only displays if you have administrator privilege. Create an Inspection from a Study Question Navigate: Supervisor/Studies/Right-click study/Select Questions Note that you must have supervisor privilege to create an inspection from a study question. If you only have inspector privilege, see Create an Inspection from Scratch. To create an inspection from a study question, navigate to the Questions page. Right-click the question and select Inspector/Add New Inspection. The following dialog displays: Here are the fields: Field Description ID Required, unique identifier for the inspection. Description Optional information about the inspection or question used to create the inspection. Owner The field defaults to the user ID who is creating the inspection. Administrators can change the owner through the drop-down list. If the user is not an administrator, the field is grayed out, and the owner cannot be changed. Filters You can copy coded closed-ended or other specify questions to the inspection to be used for filtering. Click a question and then click the forward arrow button to move the question to the Selection Questions section. You can 265 PrintDoc01 020813 select multiple questions by using Ctrl/Click or Ctrl/Shift and then click the forward arrow button. NOTE: If you want to use any questions for filtering, you must include them when creating the inspection. There is no way to copy them to the inspection later; you would have to re-create the inspection. Selected Questions These questions are copied to the inspection and can be used for filtering the inspection results. If you wish to remove a questions, click the question and then click the back arrow button. OK Click OK to create the inspection. Cancel Click Cancel to return to the Questions page without creating the inspection. After you create an inspection, the dialog closes, and the Questions page displays. To view the inspection that you created, right-click the question again and select Inspector/View. This action navigates to the Inspections page. When you create an inspection from a question, you are creating a new data set which is independent of the question. When you work with the inspection, you are not changing the data in the original question. You may make any number of inspections from the same question. Create an Inspection from Scratch Navigate: Inspector/Inspections To add an inspection from scratch, navigate to Inspector. The Inspections page displays. Right-click anywhere on the page and select Add new inspection. After you enter the information in the Inspection Creation dialog, the next step is to load data to the inspection. The Inspection Creation dialog displays with these fields: Field Description ID Required, unique identifier for the inspection. The Data Variable ID is for the main set of comments or data that you want to load and analyze. Data Variable ID 266 The default ID is Comments. Leave the default if your data has no question ID or column label. If you have a dataset which has a question ID or column label, change the Data Variable ID to that ID. For example, if you are loading data with the question ID of Q10_01, change the Data Variable ID to Q10_01. Data Management If you have filter data to load, add those IDs on the Variables page. Description Optional information about the inspection. OK Click OK to create the inspection. Cancel Click Cancel to return to the Inspections page without creating the inspection. Load Data to an Inspection Navigate: Inspector/Inspections Data can be loaded from standard Ascribe load file types (such as Excel) as well as from single-column text files and two-column tab delimited files. A single-column file would contain only the comments to be analyzed. A two-column file would contain the respondent IDs in the first column and the comments in the second column. If you are loading data which has filter variables/questions, add the IDs of the filter variables on the Variables page for the inspection. To load data, right-click the inspection and select Load Data. The Load Data dialog opens with these fields: Field Description Enter The Name Of The Data File Use the Browse button to find the file on your computer. OK Click OK to start the load process. A message will indicate the job has been submitted. If you want to see the progress of the job, navigate to the Jobs page (after you close the dialog, right-click and select General/Jobs.) Close Close the dialog after you submit the file to be loaded. After you load data with filter variables/questions, return to the Variables page to label the data in the variable. Inspection Page Navigate: Inspector/Inspections/Right-click inspection/Select Open The Inspection page has this layout: Inspection: Inspection ID Menu Toolbar Results Area 267 PrintDoc01 020813 The house icon ( ) navigates to the Ascribe home page. After the icon is the inspection ID. Below is the toolbar, and the options will vary according to which menu item has been selected. The Inspection page opens with the General Information page displayed. You can navigate to other areas through the menu on the left. The menu has these options: Option Description Inspections Navigate to the Inspector/Inspections page. Ascribe Illustrator Navigates to Ascribe Illustrator, which is a data visualization tool. You must have an Illustrator license and privilege to use the product. Help Opens the help information for this page. General Information The General Information section acts an overview of what has been done with the inspection. You can see how many comments were loaded and when and what analysis has been done on the inspection. Search Look for certain words, exact phrases, and words near each other. You can also create a codebook from search results and apply codes to search results. Curate Edit the data. Sentiment Ratings Sentiment analysis evaluates the emotion associated with the responses or comments. The Ratings option analyzes the data for sentiment on a five-point scale. For each rating, you can view the sentiment extracts (words that contain sentiment) and the full comments as well as a pie chart of the results. You can save the results as a separate question, with the comments coded according to the sentiment rating. Sentiment Topics View the topics (the ideas in the comments) which are described by the sentiment extracts. You can view the average rating for each topic, the comments associated with the topics, and a "heat map" of the topics. You can save the results as a separate question, with the comments coded by topic. Concepts Analyze the data for concepts or ideas without exploring sentiment. Entities Analyze the data for names of people, products, and organizations. Accelerator Create an accelerator or automated coding model from an inspection and use it to code new data. 268 Data Management Set Filters Allows you to view sentiment results and concepts through variable filters. General Information Section The General Information section provides an overview of what has been done with the inspection. You can see how many comments are in the inspection, what kind of analysis has been performed and when, and what kind of jobs have been run against the inspection. You can also reset the text analysis processors from this screen. The General Information section has this layout: Field Description Inspection Information Comment Context Inspection description. Comments Loaded Number of comments in the inspection. Load History Load Date Date that comments were loaded. Comments Added Number of comments loaded (by date). Processing Status Processor List of text analytic processors for sentiment, concepts, and entities. Comments Number of comments analyzed by the processor. If you add or remove comments, the number in this column displays as red to indicate some change has been made after analysis. Last Processed Date that the comments were analyzed. Status After you submit comments to be analyzed, you can check the status of the job from the area. You can use the refresh icon to update the status. The status indicates if analysis was successful, failed, or is still running. Process On Load Allows for automatic analysis of comments after they are loaded to the inspection. Reset Inspectors To Defaults Button Resets all processors and removes any analysis results. Inspection Job Status Job Type Indicates what type of jobs have been run for the inspection, like scripted output or data loads to the inspection. Date Indicates when the job was run. 269 PrintDoc01 020813 File Info Indicates TA Support when a text analysis job was run or the name of the data file when a data load was processed. Status Indicates whether a job was successful, failed, or is still running. You can use the refresh icon to update the status. Reset Processors There are multiple ways to reset processors for inspections. The General Information page provides two ways. The third way is to change the type of processor used in sentiment or concept analysis. To reset an individual processor on the General Information page, right-click the processor and select Reset Inspector processing. A dialog opens with these options: Option Description Reset Processing Results Choose this option if you just want to re-analyze the current comments. Reset Processing Results And Responses Choose this option if you have made changes to the current comments and want to re-analyze the comments. For example, if you used the curate option after you analyzed the data, you may want to re-analyze the data to pick up those changes. To reset all of the processors, use the Reset Inspectors to Defaults button, found at the bottom of the Processing Status section of the screen. All processors will be reset. 270 Data Management Search The Search text box offers flexibility in finding words or phrases in the comments. You can enter search criteria in the text box or in the Advanced Search Options dialog. Either enter Here are the options for the Search text box: Option Description Search Criteria Use the Search text box to look for certain words, exact phrases, and words near each other. You can also apply codes to the search results using Code this search. Recent Searches The Recent Searches option displays the search criteria you entered in the text box in case you want to use them again. The list is cleared when you leave Text Analytics. Advanced Search Options The Advanced Search Options dialog allows you to enter search criteria without entering Boolean or other operators. Search Options The Search Options dialog has options to change the highlight color in the search results, and options for using the NEAR operator. Code This Search On first use, this option copies all of the responses to a new question/variable. You can view the new variable on the Variables page. The variable ID will be TextSearch. With every use, it creates a code for the search criteria and applies that code to the appropriate responses. Prior Coded Searches This option displays the search pattern and code description used to code search results. Search Criteria There are a variety of ways to find words or phrases in the comments. Below are some suggestions: Option Description Find One Word Enter the word, and press Enter or the arrow ( ). Comments that contain the word will display, and the word will be highlighted in the results. Enter the phrase with double quotes (") surrounding the beginning and ending words, and press Enter or the arrow ( ). Comments that contain that phrase will display, and the phrase will be highlighted in the results. Find An Exact Phrase For example, you could enter "good service" in the text box to display comments that contain that phrase. If you enter multiple words without using the quotes, the search will look for either word and not as an exact phrase. (It's as though you entered OR between the 271 PrintDoc01 020813 words. See Find Multiple Words for more information.) To find more than one word, use Boolean operators such as AND and OR. The AND and OR must be in all capital letters. After you enter the criteria in the text box, press Enter or the arrow ( ). For example, if you enter server AND service in the text box, the search results display responses that have both words. Find Multiple Words With Boolean Operators If you enter server OR service, the search results display responses that contain either word. If you enter multiple words without using a Boolean operator or quotes, the search assumes that it is an OR search. For example, if you enter good service server, the search will look for good OR service OR server. You can enter various combinations with the Boolean operators. For example, if you enter "good service" OR server, the search will display responses that have either the phrase good service or the word server. You can also search for multiple phrases. For example, you can enter "good service" AND "good server". Find Multiple Terms Using Parentheses You can use parentheses to group terms for a search. For example, if you enter (soft OR gentle) AND strong, the search will find responses with either the words soft OR gentle and the word strong. You can also use phrases with parentheses. For example, ("gentle to nose" or "gentle on nose") AND thick. Advanced Search Options The Advanced Search Options dialog allows you to enter search criteria without entering Boolean or other operators. The text box will display the search criteria with the operators automatically inserted between the words. You can use the Advanced Search Options dialog to change your search criteria, or you can manually change/edit the criteria in the text box. Here are the options: Option Description Any Of These Words When you enter words in this field, the search automatically inserts the Boolean operator OR. You can enter a series of words with just a space between them. The text box will display the search criteria with the OR inserted between the words, and the search results will 272 Data Management display responses with any of the words. Each Of These Words When you enter words in this field, the search automatically inserts the Boolean operator AND. You can enter a series of words with just a space between them, and the search results will display responses that contain all of the words. This Exact Phrase When you enter a phrase in this field, the search automatically inserts the double quotes around the phrase. (Do not add the double quotes yourself as this will lead to incorrect results.) The search results will display responses that contain the phrase. This option finds words that are within a certain range of each other. The search automatically inserts double quotes around the words and adds a tilde (~) and the range number after the words. For example, if you entered chocolate within 3 words of flavors, the text box would display "chocolate flavors"~3. The results would display responses such as these: The Word [ ] Within [ ] Words Of [ ] The flavors are not only chocolate, but taste good. The chocolate and coffee combo flavors. Probably the chocolate flavors. Notice that the entered order of the words is not respected unless you have Keep Order for NEAR Search selected in the Search Options dialog. If you have the Keep Order option selected, then this comment would not be found since flavors came before chocolate in the response: The flavors are not only chocolate, but taste good. Search Options The Search Options dialog has options to change the highlight color in the search results, and options for using the NEAR operator. Here are the options: Option Description Highlight Color Click this box to change the color of the search results 273 PrintDoc01 020813 highlighting. 'NEAR' Search Options NEAR Distance To use the NEAR operator, enter the search criteria in the text box like this: word NEAR word For example, chocolate NEAR flavor The distance field defines the maximum number of words that can separate two search words. For example, if you enter rich NEAR smooth and the distance field is 5, then only responses with 5 or less words separating rich and smooth will display. If this box is checked, then the entered order of the search words will be respected. For example, if you enter chocolate NEAR flavor and have this box checked, only responses where the word chocolate comes before flavor will display; keep in mind that the distance field determines how many words can separate chocolate and flavor. Keep Order For 'NEAR' Search If this box is not checked, then the order of the words does not matter. Also, this field affects the search type "The word [ ] within [ ] words of [ ]" found on the Advanced Search Options page. If the Keep Order box is checked, the order of the words is respected (just like the NEAR search.) Code This Search This option allows you to use the search results to create a codebook and apply a code to search results. You must do a search first in order to use this option. The first time you select the option, all of the responses in the inspection are copied to a new question/variable. The new variable ID will be TextSearch and can be viewed on the Variables page. When you select the option, the Code Label dialog displays. Enter the code description in the code label box. The search term is automatically placed in the long help field of the code properties. Here is an example of the Code Label dialog where the search term was the word static, and the code description is Eliminates/controls static: 274 Data Management Press OK to apply the code to search results. You can use Review Responses to view or change the coding and the codebook. (Navigate to Review Responses from the TextSearch variable on the Variables page.) Here is what the codebook looks like in Review Responses; you can also see that the code was applied to the responses: Here is the code properties dialog for the new code: 275 PrintDoc01 020813 You can continue to search and code the responses using the Code this search option. The codes are added to the codebook for the search question/variable. You can keep track of the search terms you've used to create codes with Prior Coded Searches. Prior Coded Searches This option displays the search terms used to code search results. You can check the dialog to see what search terms have been used to code responses. A dialog displays with the following information: Field Description Description The term entered in the code label field when the code was created. Search Pattern The term that was used to get the search results. Here is an example of the Coded Searches dialog box: 276 Data Management Curate The Curate options are used to edit or "clean" the responses. For example, you could remove extraneous characters or misspellings in the responses. You can also delete entire responses or respondents. When you click Curate, the following screen displays. The Curate screen has a toolbar with these buttons: Options, Search and Replace, Update All, and Remove. The responses display in the Comments column, and changed responses display in the Edited Comment column. In addition, the screen has a right-click menu. You can sort the Comments in alphanumeric order by clicking the Comments column heading. Right-Click Menu If you right-click the responses in the Comments column or the Edited Value column, the following menu options are available: 277 PrintDoc01 020813 Option Description Select All Click this option to select all comments. Clear All Click this option to clear all comments. Undo Selected Changes Click this option to clear selected comments. Submit Selected Changes Saves the changes to selected responses; the responses are cleared from the Edited Value column, and the updated responses display in the Comments column. Edit Individual Response You can edit an individual response by double-clicking it in the Comments column and making changes to the response. When you press Enter, the changes display in the Edited Value column. To save the changes, click Update All or right-click the comment and select Submit Selected Changes. Options for Curate The curate option allows you to select how many items to display per page or to disable paging. If you use paging while you curate data, save your changes before you move to the next page. Option Description Items To Display Per Page Defaults to 1000; however, you can edit this field. Enter 0 if you do not want any paging. Search and Replace The Search and Replace option allows you to edit the responses. You have the option to find certain text before replacing the verbiage or you can find and replace at the same time. The changed responses display in the Edited Value column. When you select Search and Replace, the following dialog displays. 278 Data Management Here are the options: Option Description Enter the text you want to find or use the drop-down list of recently used text. To just find the text in the responses, click OK. To find and replace at the same time, click the Replace With check box and enter replacement text. Search Expression: You can use standard regular expressions in the searches. For example, if you enter the word as, the search results will contain all words with those characters (has, was, taste, etc.) To find only the word as, use the regular expression \bas\b. (Note that standard regular expressions are different than the "regular expressions" used in other places in Ascribe™. You can search for a phrase without enclosing it in quotes. For example, you could enter a phrase like friendly service. Replace With: Click the check box to enable the Replace With text box. Enter the replacement text or use the drop-down list of recently used text. To replace text with blanks, click the check box and do not enter anything in Replace With text box. Look In: You have the option to search either Original List (all of the responses) or Current List (just responses that were last displayed.) Use Paging If you using paging, you are only searching/replacing the data on a page. Update the page before you move to the next page. If you do not use paging, all of the data will be 279 PrintDoc01 020813 searched/replaced. Here is an example of the screen after a search and replace action. The Comments column has the original responses with the "search" criteria highlighted. The Edited Value column displays the edited responses with the replacement text highlighted. The edited action is not saved until you click Update All. Update All When you click Update All, the responses in the Edited Value column are saved. The responses are cleared from the Edited Value column, and the changed responses display in the Comments column. Here's what the screen looks like after the comments have been updated: You can continue working with this list or use the Search and Replace option to work with all responses. Remove The Remove option allows you to delete a response from this question or to delete a respondent's responses to all questions in the study. 280 Data Management To remove a response, click the response, and it will be highlighted. Click the Remove button. Select the option to remove response only for this question only or delete entire respondent. (If you choose to delete the entire respondent, all of the respondent's comments are deleted for all questions in this study.) Click OK. To remove more than one respondent at a time, you can use the Ctrl/Click or Ctrl/Shift keys to select multiple responses and then click Remove. Sentiment Ratings The Sentiment option extracts and evaluates the "feeling" associated with the response. The system looks at the entire phrase and tries to find words indicating sentiment. Here are examples of responses with sentiment: "I didn't like the price," "The cost is too high," "The service was wonderful." The system would not find sentiment in a response that only said "price" or "cost" or "service." Only responses with distinct sentiment are included in the analysis results. The Sentiment option is divided into two areas: Ratings and Topics. The Ratings page displays the rating scale, the responses, and the counts associated with rating. If sentiment is found in a response, it is evaluated on a five-point rating scale. The rating scale is: strong positive (+2), positive (+1), neutral (0), negative (-1), and strong negative (-2). Note that a neutral sentiment does not mean the absence of sentiment. A response like "it was okay" is considered to have neutral sentiment. The toolbar for the Ratings page has the following options: Option Description Choose Standard English Sentiment Processor or Modified English Sentiment Processor. Options The Modified English Sentiment Processor uses the Problems extractor in addition to the Sentiment extractor. The Problems extractor finds responses that have negative connotations such as "moldy" or "dirty"; however, these words may not have been in a phrase with negative sentiment. For example, "I hate the moldy shower" is a response with negative sentiment. "The shower was moldy" does not contain negative sentiment and would not be found by the Standard English Sentiment Processor. The Modified English Sentiment Processor will find more negative ratings than the Standard Processor and may skew your chart view to the negative. Minor problems will be in the negative rating, and major problems will be in the strong negative rating. Changing processors will discard the results - If you change processors after you run an analysis, the results of the first analysis will be lost. 281 PrintDoc01 020813 Analyze The first time you select Sentiment for a question, you have to select Analyze to start the extraction and evaluation process. List View Displays list of responses and ratings. Chart View Displays a pie chart of the rating information. This option creates a variable on the Variables page called Sentiment Rating. It also copies the responses from the inspection and labels the responses by sentiment rating. The labels are strong positive, strong negative, negative, neutral, and positive. Responses without sentiment are not given labels. If a response has more than one sentiment associated with it, the response will receive more than one label. Save to Ascribe™ Note that all responses in the data are saved whether they are labeled or not. You are charged transactions for all of the responses, not just the ones that are labeled. An inspection has to be saved in order to be used for scripted outputs, Ascribe Presentations, or Ascribe Illustrator. When you select Analyze, a dialog box displays with these options: Option Wait For Completion Description If you choose this option, you cannot do any other activities while the analysis is processing. A notice displays when the submitted job starts and completes. The results display when the process is finished. The dialog box also has an option to change the default timeout limit of 3 minutes. Submit Job And Continue This option submits the analysis as a background job and allows you to perform other tasks while the job completes. You can view the job on the Jobs page; it is listed a scripted output with the file name of TA Support. You can also view the status on the General Information page. List View for Ratings The system examines the responses for sentiment and assigns a rating to it. The List View displays after the analysis is finished and has these columns: Column Description Rating The rating scale is strong positive, positive, neutral, negative, and strong negative. Mentions The number of responses that were found to have the 282 Data Management corresponding sentiment. Click this number to display the comments on the right side of the screen. The sentiment word(s) will be highlighted in the responses. Extracts The number of sentiment items or extracts found in the responses. Sentiment items are words or phrases that indicate mood or feeling. Click this number to see the extracts on the right side of the screen. In the example below, there are 25 mentions that contain strong positive sentiment items. To see the comments, click the Mentions field. The comments display in dialog, and the sentiment items are highlighted in the comment: To display the sentiment extracts (words or phrases that contain sentiment), click the Extracts field. In this example, 8 sentiment extracts were found to be strong positive. After clicking the Extracts field, the sentiment extracts display on the right side of the screen. You can view the comments which contain the sentiment extracts by clicking the Comments field. The comments display in a dialog box, and the extracts are highlighted in the comments. Note that the number of mentions of an extract may be greater than the number of comments because an extract may be mentioned twice in a comment. 283 PrintDoc01 020813 Chart View for Ratings To see a pie chart based on the Ratings, click Chart View. If you hover over a pie piece, the count and percentage for that sentiment display. Click a piece of the pie to view the sentiment extracts. Click the Comments field to see the responses for that topic. You can sort the Sentiment column by alphanumeric order if you click the column heading. You can sort the Comments column from highest to lowest (and vice versa) if you click that column heading. Here is an example where the Strong Positive pie piece was clicked, and the Comments field for "great" was clicked to display the responses: 284 Data Management You can also click one of the sentiment labels at the top of the chart (strong positive, positive, neutral, negative, or strong negative) to see the topics associated with the sentiments. Click the Comments field to see the responses that contain the sentiment item. You can sort the Topic column by alphanumeric order if you click the column heading. You can sort the Comments column from highest to lowest (and vice versa) if you click that column heading. Here is an example where the Positive sentiment label was clicked, and the Comments field for "price" was clicked: Sentiment Topics 285 PrintDoc01 020813 Sentiment analysis has two parts: the sentiment items, which are words that indicate "feeling" or "emotion", and the topics, which are the part of the response that the sentiment describes. Generally speaking, the topics are usually (but not always) the nouns associated with the sentiment. The toolbar for the Topics page has the following options: Option Description Choose Standard English Sentiment Processor or Modified English Sentiment Processor. Options The Modified English Sentiment Processor uses the Problems extractor in addition to the Sentiment extractor. The Problems extractor finds responses that have negative connotations such as "moldy" or "dirty"; however, these words may not have been in a phrase with negative sentiment. For example, "I hate the moldy shower" is a response with negative sentiment. "The shower was moldy" does not contain negative sentiment and would not be found by the Standard English Sentiment Processor. The Modified English Sentiment Processor will find more negative ratings than the Standard Processor and may skew your chart view to the negative. Minor problems will be in the negative rating, and major problems will be in the strong negative rating. Changing processors will discard the results - If you change processors after you run an analysis, the results of the first analysis will be lost. Analyze The first time you select Sentiment for a question, you have to select Analyze to start the extraction and evaluation process. Note that if you select Analyze for Ratings, you do not have to run it again for Topics. Group Opens a dialog box and offers options for combining topics. Export Exports file to Excel (creates an XML file which can be opened in Excel.) This option is grayed out until you save the results to Ascribe™. List View Displays list of topics and average ratings. Chart View Displays a "heat map" of the topics; it shows how positive or negative the topics are. Minimum Mentions Allows you to change the display according to the number of mentions of a topic. Minimum mentions defaults to 2. The arrows on the box allows you to increase or decrease the number of mentions. This field does not affect grouping; in other words, all extracts are used when creating groups. 286 Data Management Note: If you are thinking of saving the inspection to Ascribe, change the minimum mentions to 1 to see everything that would be labeled. This option creates a variable on the Variables page called Sentiment. It also copies the responses from the inspection and labels the responses by topics. Save to Ascribe™ Note that all responses in the data are saved whether they are labeled or not. You are charged transactions for all of the responses, not just the ones that are labeled. An inspection has to be saved in order to be used for scripted outputs, Ascribe Presentations, or Ascribe Illustrator. When you select Analyze, a dialog box displays with these options: Option Wait For Completion Description If you choose this option, you cannot do any other activities while the analysis is processing. A notice displays when the submitted job starts and completes. The results display when the process is finished. The dialog box also has an option to change the default timeout limit of 3 minutes. Submit Job And Continue This option submits the analysis as a background job and allows you to perform other tasks while the job completes. You can view the job on the Jobs page; it is listed a scripted output with the file name of TA Support. You can also view the status on the General Information page. List View for Topics The system examines the responses for sentiment and assigns a rating to it. The List View for Topics displays after the analysis is finished and has these columns: Column Topics (uuu/ttt) Description The topic in the response that sentiment described. The system finds similar words and combines them into a single topic. For example, the topic "hair" could include topic phrases such as "my hair," "your hair," and "hair in picture." The numbers in parentheses stand for the number of unique respondents (uuu), followed by the total number of response mentions (ttt). In other words, if you save the extract as a question, the first number indicates how 287 PrintDoc01 020813 many respondents would be coded. The second number indicates how many codes would be applied. You can sort the Topic column by alphanumeric order if you click the column heading. Comments The number of comments which contain the corresponding topic. Click the number in the Average Rating column to display the comments. You can sort the Comments column from highest to lowest (and vice versa) if you click the column heading. Each topic and sentiment combination is given a score: strong positive is +2, positive is +1, neutral is 0, negative is -1, and strong negative is -2. The scores for each combination are added and divided by the number of combinations to get an average rating for that topic. For example, there are three topic/sentiment combinations for topic "movie." Two of the combinations were "I love the movie," which gets a strong positive score. One combination was "I hate the movie," which gets a strong negative score. The formula would be (2 + 2 + -2)/3 = .667, which would be a positive average rating. Average Rating The background color of this column reflects the category of the average rating: Bright green = strong positive (an average rating of 1.01 to 2.00) Dark green = positive (0.01 to 1.00) Gray = neutral (0.00) Orange = negative (-0.01 to -1.00) Red = strong negative (-1.01 to -2.00) You can sort the Average Rating column from highest to lowest (and vice versa) if you click the column heading. In the example below, the topic "price" was found to have sentiment in 14 comments. To view the comments, click the Average Rating field, and a dialog displays. 288 Data Management If a topic is grouped, the comments display on the right side of the screen as shown below. Note that you can change the Group name by double-clicking the name and entering a new name. You can remove extracts by right-clicking them and selecting Remove selected items from group. You can copy individual responses if you right-click the comment and select Copy selected items. To copy multiple items, use the Ctrl/Click or Ctrl/Shift keys to select multiple comments and then right-click and select Copy selected items. If you right-click the topics in the Topic column, the following options are available: 289 PrintDoc01 020813 Option Description Group Similar Click any topic to search and combine other topics that have the same characters. A dialog opens and allows you to enter a group name. The group name defaults to the topic with the highest number of mentions. Click a topic you want to combine with another topic. Ctrl/click any other topics. Right-click and select Combine selected items. A dialog opens. Enter the topic label in the text box, and click OK to combine the topics. The rating average for the combined topic will be updated and will be based on the scores of the combined topics. Group Selected Items If you combine topics with an item that has already been grouped, the single topic or topics will automatically receive the group label. You do not receive the dialog where you can enter a different label name. If you combine two groups of topics, you do have the opportunity to create a new label name or choose one of the existing group names. You may also use the Group option to combine topics. Remove Selected Items Un-group Selected Items If you would like to remove a topic, right-click it and select Remove selected items. If you want to remove multiple items, you can use the Ctrl/Click or Ctrl/Shift keys to select multiple topics before using the right-click menu. Also, you can right-click a topic, select Remove selected items, and then select Remove topics with mentions fewer than X. Enter the appropriate number, and topics with less than that number of mentions will no longer display. This option allows you to take topics that were grouped and return them to their original state. Chart View for Topics To see a "heat map" of the topics, click Chart View. The "heat map" displays the topics in one of four quadrants. The vertical axis represents positive to negative ratings, and its color goes from dark green to dark red. The horizontal axis represents weak to strong feeling, and its color goes from yellow to blue. Here is an example of a Topics chart: 290 Data Management The quadrants have the following ratings: Upper right is strong positive Upper left is positive Lower left is negative Lower right is strong negative If a topic is listed in the middle, it is neutral. The color of the topic also indicates its rating: Bright green = strong positive (an average rating of 1.01 to 2.00) Dark green = positive (0.01 to 1.00) Gray = neutral (0.00) Orange = negative (-0.01 to -1.00) Red = strong negative (-1.01 to -2.00) The font size indicates the number of mentions - the larger the font, the higher the number of mentions. You can change some display features of the chart through the Options button on the Chart toolbar. To view the responses associated with a topic, double-click a topic in the chart. Chart Options 291 PrintDoc01 020813 Click Options to change the appearance of the chart. Here are the options: Option Description Chart Background Color Click the color selection box to change the background color of the chart. Item Shape And Item Shape Attributes Use this option if you would like the words on the chart to appear in a shape. The default is none (which means the word displays without a shape around it.) Shape options are capsule, oval, and rectangle. If you choose an item shape, you can choose the item background color and whether the color should be solid or gradient (goes from light to dark.) Item Font Attributes Choose the font and font sizes for the chart. Maximum Label Length The maximum label length defaults to 20 characters. You can increase the maximum length. Note that if you increase the maximum too much, the label may become larger than the screen width. The Minimum Mentions text box also allows you to change the chart display. The field refers to the number of mentions of the word in the responses. The arrows on the box allows you to increase or decrease the number of mentions. Note that if there are a large number of items on the chart, all of the lowest mention items may not display. The chart tries to display the items with the most mentions first and tries to determine placement without overlaying too many items. Group The Group option is another way to combine topics. Note that the Group dialog is the same for both Sentiment Topics and for Concepts (with the exception of the Average Rating column.) When you click Group, a dialog opens with these fields: Fields Description Grouping Expression Enter standard regular expressions to search for topics (or concepts) with this criteria. Group Button Once you enter a grouping expression, click the Group button to start the search. Group Label Enter the name for this group of topics (or concepts.) Concepts This column lists all of the topics (or concepts) that met the search criteria. Average Rating For Topics, the average rating value displays and can be sorted in ascending or descending order. (This column does not display for Concepts.) Mentions This column lists the number of responses in each topic (or concept.) 292 Data Management Remove Left-click a topic (or concept) that you want to remove from the list, and click Remove. That topic (or concept) will not be combined with the other topics (or concepts.) If you want to remove multiple items, you can use the Ctrl/Click or Ctrl/Shift keys to select multiple topics (or concepts) before you click Remove. OK Click OK to combine the topics (or concepts.) Export The Export option creates an XML file which can be opened with Excel. This option is only available once the topics have been saved to Ascribe. When you click Export, the Save As dialog box for your computer opens. Enter a name for the file. To view the file, browse to the location where you saved the file. The file has this format: Field Description Sentiment The type of sentiment found in the comment. Expression The extracted word or phrase which shows sentiment. Topic The noun or topic described by the sentiment. Topic Group If the comment was included in a group of comments, the grouped label displays here. If the comment was not grouped, the topic group will be the same as the topic. Value The rating associated with the sentiment. 293 PrintDoc01 020813 Comment ID The ID associated with the comment or response. Extract The phrase or words that the sentiment describes Comment The full comment or response. Here is a screen shot of an exported file: Adding and Analyzing Additional Data for Sentiment If you analyze data for sentiment and then add additional data, you should analyze the question again. When you select Analyze a second time, only the new data will be analyzed. The system will try to fit the new data into any groups that have been established, based on an exact match of extracts. If you want the entire dataset to be analyzed, then use the right-click menu option on the General Information screen to reset processing results and responses and then analyze the data again. Concepts The Concepts option examines the responses for related words or ideas. It does not examine the responses for sentiment. The toolbar for the Concepts page has the following options: Option Description Choose Standard Concept Processor or Modified Concept Processor. Options The Modified Concept Processor may find more concepts. It examines the sentence structure for nouns and modifiers separated by "being" words. For example, "the dog is brown" may be found as a concept for "brown dog." It will also look for gerunds (the -ing form of a verb that functions as a noun.) Changing processors will discard the results - If you change processors after you run an analysis, the results 294 Data Management of the first analysis will be lost. Analyze The first time you select Concepts, you have to select Analyze to start the process. Group Opens a dialog box and offers options for combining concepts. See Group in the Sentiment Topics section for more information. The Group dialog is the same for both Concepts and Sentiment Topics (with the exception of the Average Rating column.) List View Displays list of concepts. Chart View Displays a word cloud of the concepts. Minimum Mentions Allows you to change the display according to the number of mentions of a topic. Minimum mentions defaults to 2. The arrows on the box allows you to increase or decrease the number of mentions. Note: If you are thinking of saving the inspection to Ascribe, change the minimum mentions to 1 to see everything that would be labeled. This option creates a variable on the Variables page called Concepts. It also copies the responses from the inspection and labels the responses by concept. Save to Ascribe™ Note that all responses in the data are saved whether they are labeled or not. You are charged transactions for all of the responses, not just the ones that are labeled. An inspection has to be saved in order to be used for scripted outputs, Ascribe Presentations, or Ascribe Illustrator. To start the process, click the Analyze button. When you select Analyze, a dialog box displays with these options: Option Description Wait For Completion You cannot do any other activities while the analysis is processing. The results will automatically display when the process is finished. Submit As A Job This option submits the analysis as a background job and allows you to perform other tasks while the job completes. You can view the job on the Jobs page; it is listed a scripted output with the file name of TA Support. You can also view the status on the General Information page. List View for Concepts The List View for Concepts will display these columns: Column Description Concepts (uuu/ttt) The set of words or ideas that appear in the responses. 295 PrintDoc01 020813 The numbers in parentheses stand for the number of unique respondents (uuu), followed by the total number of response mentions (ttt). In other words, if you save the extract as a question, the first number indicates how many respondents would be coded. The second number indicates how many codes would be applied. You can sort the Concepts column by alphanumeric order if you click the column heading. The number of comments which contain the corresponding concept. Click this number to display the comments. If the concept is not grouped, the comments display in a dialog. Comments If the concept is grouped, information displays in the Group and Extract columns on the right side of the screen. The Group column lists the group name. (You can change the Group name by double-clicking the name and entering a new name.) The Extract column lists the extracts that are included in the group. To remove an extract from the grouped concept, right-click an extract and select Remove selected items from group. You can view the combined comments by clicking the Comments field next to the Extract. You can sort the Comments column from highest to lowest (and vice versa) if you click the column heading. 296 Data Management If you right-click a concept in the Concepts column, the following options are available: Option Description Group Similar Click any concept to search and combine other concepts that have the same characters. A dialog opens and allows you to enter a group name. The group name defaults to the concept with the highest number of mentions. Click a concept you want to combine with another concept. Ctrl/click any other concepts. Right-click and select Combine selected items. A dialog opens. Enter the label in the text box, and click OK to combine the concepts. Group Selected Items If you combine concepts with an item that has already been grouped, the single concept or concepts will automatically receive the group label. You do not receive the dialog where you can enter a different label name. If you combine two groups of concepts, you do have the opportunity to create a new label name or choose one of the existing group names. You may also use the Group option to combine concepts. Remove Selected Items If you would like to remove a concept, right-click it and select Remove selected items. If you want to remove multiple items, you can use the Ctrl/Click or Ctrl/Shift keys to select multiple concepts before using the right- 297 PrintDoc01 020813 click menu. Also, you can right-click a concept, select Remove selected items, and then select Remove topics with mentions fewer than X. Enter the appropriate number, and concepts with less than that number of mentions will no longer display. Un-group Selected Items This option allows you to take concepts that were grouped and return them to their original state. Chart View for Concepts The Chart option for Concepts displays a word cloud. The font size indicates the number of mentions - the larger the font, the higher the number of mentions. Here is an example of a chart created from a Concept List: The word cloud has the same options as the Topics Chart. The Minimum Mentions text box also allows you to change the chart display. The field refers to the number of mentions of the word in the responses. The arrows on the box allows you to increase or decrease the number of mentions. Note that if there are a large number of items on the chart, all of the lowest mention items may not display. The chart tries to display the items with the most mentions first and tries to determine placement without overlaying items. To view the responses associated with a concept, double-click a concept in the word cloud. Adding and Analyzing Additional Data for Concepts If you analyze data for concepts and then add additional data, you should analyze again. Only the new data will be analyzed. The system will try to fit the new data into any groups that have been established, based on an exact match of extracts. 298 Data Management If you want the entire dataset to be analyzed, then use the right-click menu option on the General Information screen to reset processing results and responses and then analyze the data again. Entities This option analyzes the data for names of people, products, and organizations. The toolbar for the Entities page has the following options: Option Description Options Entities processing options are under development. Analyze The first time you select Entities for a question, you have to select Analyze to start the process. List View Displays list of concepts. Chart View No chart view is available at this time. When you select Analyze, a dialog box displays with these options: Option Wait For Completion Description If you choose this option, you cannot do any other activities while the analysis is processing. A notice displays when the submitted job starts and completes. The results display when the process is finished. The dialog box also has an option to change the default timeout limit of 3 minutes. Submit Job And Continue This option submits the analysis as a background job and allows you to perform other tasks while the job completes. You can view the job on the Jobs page; it is listed a scripted output with the file name of TA Support. List View for Entities When the analysis is complete, the List View displays. The list has these columns: Column Description The categories are: People - names of people (brands may also display here) Entity Products - names of products Organizations - names of companies Other Mentions - proper nouns Comments The number of comments that correspond to the category. Click this number to display the comments on the right side of the screen. Mentions A "mention" is the specific word or phrase found in the 299 PrintDoc01 020813 response that was identified as a person, product, organization, or as a proper noun. The number of mentions is listed for each category. Click the number to display the specific words or phrases on the right side of the screen. Here is an example of the Entities page after clicking the number of mentions for Organizations: Accelerator You can create an accelerator (automated coding model) from an inspection which has been saved to Ascribe. Click the Accelerator menu item, which displays the Accelerator page for the inspection. Here are the options on that page: Option Description Create Allows you to create an accelerator from an inspection. Train Starts the training job for the accelerator. Edit Opens an editor for the training examples. Analyze Create an Accelerator from an Inspection Click the Create button from the top menu. A dialog opens with these options: Option Description Accelerator ID Unique identifier for the accelerator; it defaults to ACM_inspection name. Use Saved Sentiment Topics As Training Examples Creates the accelerator from the saved Sentiment Topics. Use Saved Concepts As Training Examples Creates the accelerator from the saved Concepts. Use Existing Allows you to use an existing accelerator as training 300 Data Management Accelerator examples. Train After Creation Automatically starts the training job after the accelerator is created. Edit Accelerator Training Examples The training example editor in Inspector allows for very simple editing of examples. The top part of the screen has these fields: Field Description Accelerator ID Unique identifier for the accelerator; it defaults to ACM_inspection name. Description Optional description of the accelerator. Approved The accelerator defaults to approved. Account Access Designates this codelist as an account codebook. See Account Codebooks for more information. 301 PrintDoc01 020813 The second part of the screen displays the category (topic or concept) and the number of comments for that category. Click the number in the comments field to display the comments. The category section has these right-click options: Option Description Insert Examples Opens a dialog where you can enter more training examples. Click OK to add the examples. Delete All Examples Permanently removes all examples from the category. 302 Data Management The comment section has these right-click options: Option Description Copy Selected Items Copies selected items to clipboard so you can use them as new examples. Edit Example Opens a dialog where you can change the example text. Delete Selected Examples Permanently removes the selected examples from the category. To return to the Accelerator page, click the Close button at the top of the page. After making any changes to the examples, you will need to train the accelerator before using it to code other responses. Set Filters You can filter the results of sentiment and concept analysis by filter variables. Filtering only works for sentiment and concepts - not for Search, Curate, or Entities. The Set Filters dialog allows you to select and remove filter variables. The Filter Bar displays on the Inspection page when a filter is set. It allows you to see what filters are set. Click the Set Filters button to open the Set Filters dialog. This dialog displays: 303 PrintDoc01 020813 The top left pane of the dialog displays a list of available variables. Click the variable you wish to use as a filter, and then click the Add button to move the variable to the top right pane. Once the variable is in the top right pane, click the variable again, and the options of the filter variable display in bottom pane. 304 Data Management By default, the options are all selected. You can select and unselect the options from this screen. Click OK to save and close the dialog. You can work with more than one variable filter at a time. If you want to remove a filter question, click the appropriate question in the top right pane, and click Remove. To remove all filter questions, click Remove All. Once you select a question in the Set Filters dialog and exit that dialog, a yellow bar displays next to the sentiment or concept results pane. The initial view shows what filter has been selected. The screen looks like this: To close the filter bar, click the yellow icon and then the black icon. 305 PrintDoc01 020813 Note that if you decide not to use a filter question, you need to remove it using the Set Filters dialog. Filtering Within a Filter Variable When you select options within a filter variable, the logic is based on "or." You are selecting one option or another or some combination of the options. For example, you can select to see responses from France or England or Germany. Alternatively, you can select to see only responses from France or only responses from England or only responses from Germany. You can also select some combination of the countries. Filtering With Multiple Filter Variables When using multiple variables for filtering, the logic within the variable is still based on "or". But, the logic for selecting between the variables is based on "and". In the example below, two filters are used. The system sees the logic as 'show respondents who are females AND who live in the Southeast OR Southwest region.' When you have multiple filter questions, you have to have at least one code selected in each question or no responses will be returned. Save to Ascribe or Export to Excel While Filtering When you save concepts or topics to Ascribe while filtering, all of the data in the inspection will be copied to a new variable, but only the filtered data will be labeled. The counts at the top of the Topics or Concepts columns reflect what will be labeled. In the example below, 1442 responses will be labeled using 4744 labels. 306 Data Management When you export topics to Excel while filtering, all of the data in the inspection will be copied to Excel. Variables Page Navigate: Inspector/Inspections/Right-click an inspection/Select Variables The Variables page contains additional data associated with an inspection. The additional data may be: Filter variables copied from a study question Filter variables loaded directly to an inspection Sentiment rating, topics, or concepts that were saved to Ascribe. Page Layout The Variables page has two sections. The left side contains the list of variables associated with the inspection and has these fields: Field Description The ID can indicate the ID of a filter variable copied to an inspection or loaded to an inspection. ID It can also indicate inspection results saved to Ascribe. For example, Sentiment is the topic list, Sentiment Rating is the rating summary, and Concepts is the concept list. These variables only display if those results have been saved to Ascribe. Description The description is usually the same as the ID. Comments The number of comments or responses in a variable. Labeled The number of comments or responses which received a label. 307 PrintDoc01 020813 The right-side is the Data Labels section. It allows you to add, delete, and apply labels. To work with a variable or display its data labels, click anywhere on the variable on the left side of the screen. If a variable has labeled responses, the information displays on the right side of the screen with these fields: Field Description Value The number associated with the label; it is use for reporting purposes (such as scripted outputs, Ascribe Presentations, or Ascribe Illustrator.) Label The description of a filter variable or the description of how a response was categorized in text analysis. Comments The number of comments or responses in a variable. Labeled The number of comments or responses which received a label. Add a New Variable If you add the variables directly to the inspection via the Load Data option, you must first create the variable labels on the Variables page. Right-click anywhere on the left section and select Add new variable. The dialog opens with these fields: Field Description ID The identifier of the column in the dataset that contains the filter variables. For example, if your dataset contains a column with country information and is labeled Country, you need to add a variable with the ID of Country. Label The label can be the same as the ID. OK Click OK to create the variable. Cancel Click Cancel to exit the dialog without creating the variable. Label the Data in a Variable After you have loaded the variable data, you should label the data so you can use it for filtering. To label the data, click the variable on left section. Then click the Apply Label button. Ascribe will create and apply the labels to the appropriate data. The labels display on the right side of the screen. Refresh the screen to see how many were labeled (see column on left side of screen.) Insert or Remove Labels The Data Labels section (right side of page) has options for changing the labels. Use the Insert button to add labels to the list. Use the Remove button to delete labels from the list. Create an Account Codebook 308 Data Management You can designate a list of data labels as an account codebook. Account codebooks can be used again in other studies or in creating coding models. Click the check box next to Account Codebook (on right section of the page) to set the designation. For more information, see Account Codebooks. Navigate to Review Responses You may edit the list of data labels in a variable using Review Responses. Note that if you do not have supervisor privilege, you will not have full access to all functionality in Review Responses. You will be able to remove and add labels from responses. To navigate to Review Responses, right-click a variable on the left side of the page and choose Review Responses. Remove a Variable To remove a variable, right-click the variable in the left section of the page. Select Remove variable. Variable Properties To change the ID or label of a variable, right-click the variable in the left section of the page. Select Properties. The dialog opens, and you can make changes there. Return to Inspector/Inspections Right-click anywhere on the page and select Inspector. This action navigates to the Inspector/Inspections page. Data Management This chapter details how to manage study data, from start to finish within Ascribe™. How to Load Data There are number of ways to load data: From an Excel® file From a tab delimited file From a CfMC SURVENT generated . lst file From an XML file in the Ascribe™ format From a "mapped ASCII" file that you define From custom developed interfaces that typically deliver data in "real time." Each has different formatting requirements, and each is recognized by Ascribe™ by the file name extension. When you load the data, the system will load only "new" data. This means that you may load cumulative data sets and be assured that only those records that are new will be loaded. Note The respondent number/question ID combination must be unique to be loaded. Load Tab Delimited Data 309 PrintDoc01 020813 Three-column tab delimited data is the "native format" for Ascribe™. When Ascribe™ encounters a file with a . txt extension, it assumes that the file is in this format: question ID, respondent ID, and verbatim/closed ended data. If a fourth column is present, Ascribe™ assumes it is transcription data. Each field is delimited by a tab character. (If this format is inconvenient, or you have a file with a different delimiter, you can define your own file format. See Define Your Own Load File Types for more information.) To load a tab delimited file, prepare a file of data that contains question ID, respondent ID, and verbatim or closed ended data in a tab delimited text file. (Transcription data can follow the verbatim.) At Language Logic, we use Excel™ to prepare the file. Here are the steps to prepare the file: Read the file into Excel™ (you may have to use the text to columns feature). Delete any header rows and irrelevant columns. Check that you have the question ID in the first column, the respondent ID in the second, and the verbatim text or closed-ended data in the third column. Save the file as a tab delimited text file; use Save As from the Excel™ File menu. Name the file studyname_date.txt. If you use this naming format, the person who loads the data into the study will have an easier time finding it. We recommend that you include the date in the filename for identification purposes and to make support more efficient. Tab delimited files will look like this: The next step is to load the data to the study. Data loads in the order it is in the file. So if you sort the data before you load, it will load in that sorted order. For more information, see Load Data to a Study. Load Excel® Files Ascribe™ loads Microsoft Excel ® files in their native, Excel® format. To load an Excel ® file, you will need to have it in this format. The column containing the respondent number must have the word " RespondentID", " ResponseID", "RID", " CaseID" or " SessionID" as the first cell in the column. Note You can also specify the column header using a "Load File Type". It is advised to do this only if the files you will be loading consistently use the column header you wish to specify. See "Define Your Own Load File Types" for further information. 310 Data Management Each of the questions that you are loading must have their question ID in the top row (the first cell), with the data for that question in the rows beneath those questions. A few comments: You may "customize" your Excel ® format by defining a "Load File Type". If you have multiple spreadsheets in an Excel ® workbook, only the first one alphabetically will load. The rest will be ignored. Language Logic has created an Excel ® add-in called "llutilities2.xla". This add-in places a "Language Logic Tools" option in the Excel ® toolbar. "Language Logic Tools" provides the user with a number of tools to help the user troubleshoot an Excel ® file before loading it into Ascribe™. Usually, we recommend users convert Excel ® files to a three-column tab delimited text files to avoid a few problems Excel ® files can cause when loading. (See Excel® Add-In for further information.) Load File Types Navigate: Administrator/Load File Types Ascribe™ processes files based upon the file extension. The "extension" in a file name is the set of characters following the last period. For example, if a file is named " abc.xls" then " xls" is referred to as the extension. The Load File Types page allows you to describe how a particular file will be interpreted for loading data. This page shows load file types established for the account. The list contains the following columns: Column File Extension Example DTX Description File extension used to match the incoming load file with the correct Load File Type Name standard DTX Brief description of the Load File Type Description This is the User description of the file type standard DTX file format. Type Mapped ASCII The format of the data in the file Ascribe™ can load data and study information from the following file types. In addition, Ascribe™ can define file types that are specific to your situation (see Define Your Own Load File Types.) Extension CXML File Contents File Source Question, Codebook, and Setup Confirmit v14 Ascribe™ Load Type Comments Custom Proprietary format for Confirmit question and codebook setup 311 PrintDoc01 020813 HRD Question, Codebook, and Setup CfMC Custom Proprietary format for CfMC question and codebook setup MDD Question, Codebook, and Setup Dimensions Custom Dimensions setup file ORD Question, Codebook, and Setup CfMC Custom Same as HRD with parser modified for ORC use of _c QDI Question, Codebook, and Setup Quancept Custom Proprietary format for Quancept question and codebook setup Setup.xls Question, Codebook, and Setup Language Logic Excel™ Excel™ Add-In Horizontal Excel™file formatted with the Language Logic question template. It creates or changes Ascribe™ question information. SSS Question, Codebook, and Setup Triple-S Custom Triple-S XML representation of study metadata SAV Question, Codebook, Setup, or Response Data SPSS Custom SPSS sav file VBS Question, Codebook, Setup, or Response Data Verbastat Custom Verbastat project file XML Question, Codebook, Setup, or Response Data Various Custom Ascribe™ XML representation of study metadata; XML file with associated XLST file Response Data Various Custom This format is an ASCII representation of an Excel™ spreadsheet with the columns delimited by with a comma. Processing rules are the same Setup.xlsx CSV 312 Data Management as Excel™. CTXT Response Data Confirmit v14 Custom Confirmit tabbed delimited text file. If looped questions exist, there may be multiple files. DAU Response Data mrInterview Custom Proprietary format for mrInterview data DDF.ZIP Response Data Dimensions Custom The zip file contains both the MDD & DDF files. The DDF file must have the MDD file with it. DRS Response Data Quancept Custom Proprietary format for Quancept data DRU Response Data Quancept after dedup Custom Same as DRS DTX Response Data Quancept Mapped ASCII Quancept data file with this default field mapping: Respondent ID (position 1 for 6) Questions ID (position 13 for 8) Response (position 30 to end of row ) LST Response Data CfMC Custom Proprietary CfMC data file format OST Response Data CfMc Custom Proprietary CfMC data file format TEX Response Data Quancept after dedup Mapped ASCII Same as DTX TXT Response Data Various Delimited This format is the default data load file type for Ascribe™. It has three fields per row delimited by a tab. The fields are question ID, response ID, and response. XLS Response Data Various Excel™ Horizontal Standard Excel™ horizontal format with one column for 313 PrintDoc01 020813 respondent ID and other columns for responses to questions. Question ID is identified by the first cell in each column (after the respondent ID column.) XLSX Response Data Various Excel™ Horizontal Standard Excel™ horizontal format with one column for respondent ID and other columns for responses to questions. Question ID is identified by the first cell in each column (after the respondent ID column.) Define Your Own Load File Types Navigate: Administrator/Select Load File Types/Right-click/Select Add New File Load Type A user with administrative privileges can define load file types for an account. You define the file type once and give it a unique extension. Every time the Ascribe™ loader recognizes the extension, the defined rules will be applied to the file. You can add, edit, or delete load file types from the Load File Type Properties screen. The Load File Type Properties screen has these fields: Field Description File Extension The file extension characters are the last characters following a '.' in the file name. For example a file named "study1.dtx" has an extension of dtx. A file named studi1.itr.dtx has an extension of itr.dtx. DO NOT INCLUDE THE LEADING '.' FOR THIS PROPERTY. File Type Here are the options: 314 Mapped ASCII Single -The file contains one row of data per respondent per question. Key fields are in known positions within the row. Use Field Definitions properties to specify positions. Mapped ASCII Multi -The file contains one row of data per respondent. Each row contains data for multiple questions. The start position and length of the data field is set on the Data Tab from Edit the Question. A question must be established to represent the respondent field, and its question ID should be entered as described in the Field Definitions section. Data Management Delimited -The file contains one row per respondent per question. Key fields are delimited by a special character. Use Field Delimiter property to specify delimiting character. Excel™ Horizontal -File is an Excel™ spreadsheet. The first row contains the question ID for data or the respondent ID caption for the column containing the respondent ID. Excel™ Vertical - File is an Excel™ spreadsheet containing a column for each of the key fields. Use the Field Definitions properties to specify field/column order. CSV - The file contains data in a comma separated value file. This format is the same as that produced by Excel™. Each column of data is separated by a comma. Columns that contain commas are enclosed in double quotes ("). Double quotes within a column are represented by a preceding double quote (" becomes ""). Custom - The file contains data in a format other than row/record based. Data files of this type are processed entirely by the script designated below. Default - The file will be processed according to the default format based upon a known file extension. Any script established below will be processed against the standard three column data file produced by this default processing. For a list of the current known file type and extensions, see Load File Types. Name The name of the load file type File Purpose Ascribe™ will process files for the purpose of loading responses into our database (DATA) or for the purpose of adding studies, questions and codebooks (SETUP). Description The description of the load file type Study ID Delimiter (in Filename) This is the single character which will be used to identify the study ID portion of a file name when submitting data for loading to our FTP site. This option is only used when files are submitted to the FTP site for loading. Files submitted interactively will already have the study identified. The study ID must be the first portion of the file name and will end when the delimiting character specified here has been reached. Field Delimiter The character used to separate the fields for a delimited file type. It is typically a tab character or comma. Field Definitions The use of these fields varies based upon the file type. Mapped ASCII Single Name is used for information purposes only. Column represents the starting position in the record (first position is 1). Enter a zero to specify an unused field. Width is the length of the field in characters (NOT THE END POSITION). Use 0 to specify variable length to end of line. 315 PrintDoc01 020813 Mapped ASCII Multi (only the RID field is used) Name must agree with the question ID for the question representing the respondent ID. Column represents the starting position in the record (first position is 1). Enter a zero to specify an unused field. Width is the length of the field in characters (NOT THE END POSITION). Use 0 to specify variable length to end of line. Delimited Name is used for information purposes only. Column represents the column ordinal number in the record (first position is 1). Enter a zero to specify an unused field. Width is unused for this file type. Excel™ Horizontal Name is used to specify the row 1 caption identifying the respondent ID column and the iteration column (All other columns are assumed to be question IDs.) Column is unused for this file type. Width is unused for this file type. Excel™ Vertical Name is used for information purposes only. Column represents the column ordinal number in the spreadsheet (first column is 1). Enter a zero to specify an unused field. Width is unused for this file type. CSV - Same as Excel™ Horizontal. Custom - Unused for this format. Default - Unused for this format. Values are all known based upon the default processing for a given file extension. Preprocessor Script for initial processing of files received. An example of such processing might include extraction of image snippets from a set of TIF files. These scripts are created by Language Logic or by client personnel and submitted to Language Logic. Available scripts are included in the drop-down list. Script Script provides additional custom processing after predefined processing from File Type has been completed. These scripts are created by Language Logic or client personnel and submitted to Language Logic. Available scripts are included in the drop-down list. Post Load Script to be used to provide additional custom processing after responses have been loaded into the Ascribe™ database. Typically, these are used to modify some aspect of the response information such as the translation field, based upon other database values. These scripts are created by Language Logic or by client personnel and submitted to Language Logic. Available scripts are included in the drop-down list. 316 Data Management Options Add Questions - Selecting this option will cause Ascribe™ to automatically create any question IDs found in a data file of this load file type that do not exist in the study. This option is only in effect when files are submitted for loading via our FTP site. Interactive loads (files input in Ascribe™) will use the options selected at load time. Auto-code closed end questions - When this option is selected, Ascribe™ will attempt to "code" responses being loaded by matching the data value to a code value in the codebook for the question. Only numeric responses from question types of closed ended or other specify will have this "auto code" process attempted. This option is only in effect when files are submitted for loading via our FTP site. Interactive loads (files input in Ascribe™) will use the options selected at load time. Auto-code text responses - When this option is selected, Ascribe™ will attempt to "code" responses being loaded by matching them to previously coded responses coded with the same (shared) codebook. This option is only in effect when files are submitted for loading via our FTP site. Interactive loads (files input in Ascribe™) will use the options selected at load time. Append Iteration to Question ID - This option is used to indicate that the iteration number will be appended to the question ID to create a new question ID. For example, Q1 iteration 1 will remain Q1 as question ID. Q1 iteration 2 will become Q1_2 as the question ID. Strip leading zeros from respondent ID - By default, Ascribe™ will not remove any characters or otherwise modify a respondent ID when it is loaded. For example, the respondent ID "001" will appear as "001". Use this option if you wish the respondent ID to have leading zeros removed. So, in the example, the respondent ID would load as "1". You can perform any of the following tasks to manage your load file types: Add a load file type: right-click and select Add New Load File Type. Complete the information on the Load File Type Properties page. Modify a load file type: right-click and select Edit. Complete the information on the Load File Type Properties page. Delete a load file type: right-click and select Delete. 317 PrintDoc01 020813 Load Study Setup Information Navigate: Supervisor/Studies/Right-click study/Select Load Responses / Process File Ascribe™ loads study information from various data collection systems: Quancept (an SPSS/ MR product), Survent (a CfMC product), Dimensions (an SPSS/ MR product), and Confirmit (a Firm product). Ascribe™ also reads XML study files (Ascribe™ XML.) The study information consists of: Question ID Question text Data locations Closed-end response categories and pre-lists for other specify question types Shared codebooks for iterated questions. To load either .HRD (Survent), .QDI, MDD or other setup file: 1. From the Load Response/Process File page, click the Browse button and navigate to the file location of the setup file you want to load. 2. 318 Leave the defaults and press the OK button. Data Management Load Data to a Study Navigate: Supervisor/Studies/Right-click study/Select Load Responses / Process File Once you navigate to the Load Responses/Process File page, enter the name of the file which contains the data into the field labeled 'Enter the name of the data file.' Either enter the path and file name of the file to be loaded or use the Browse button to browse your file system (just like Excel™ or Word). Load file types are determined by their extension (the characters after the last '.' in the file name.) For more information, see Load File Types. When you load data, you have some options. These options always override the load file type properties. The options are: Option Description Load Data for Existing Questions This option loads data to questions that already exist in the study. It is useful to load cumulative data files that contain questions which will not be used in Ascribe™. Data for questions that have not been defined will be ignored. Pause Load to Prompt for Missing Questions The loader will ask the user which questions should be created and/or loaded if the questions do not already exist in Ascribe™. Create Missing Questions Automatically Ascribe™ will create questions whenever data is encountered for questions that have not been previously defined in the study. When you use this option, Ascribe™ ‘guesses’ the question type since you have not defined it yet. You can change the question type on the General Tab of the Edit Question screen. Auto-Code Closed and O/S Questions While it loads data, Ascribe™ will automatically code responses to closed-ended and other specify questions which have numeric values. Auto-Code Text Responses Ascribe™ will automatically code text responses as they are loaded. See Auto-Code Text Responses on Load for more information. After you select the options, press the OK button to load the data. The Load Details screen displays. It describes the results of your load. For more information, see Load Details. 319 PrintDoc01 020813 View and Delete Loaded Data Navigate: Supervisor/Studies/Right-click study/Select Reports/Select Loads The Loads screen displays information about data loads for a study. A given load will often result in multiple rows because there is one row for each question that was loaded. Load information is not included in a saved study. If you restore a study, the restored study will have no entries on the Loads page. Here are the fields on the Loads page: Field Description Load Date The date and time the file was loaded. This time is based on the web server clock. User Name The user name of the person who loaded the file. File Name The name of the file that was loaded. Records Processed/Still Loaded The number on the left represents the responses originally loaded. The right side number is a count of those which remain loaded (have not been deleted.) 320 Data Management Question Type The type of question that was loaded. Question ID The question ID of the question that was loaded. Question Label The label of the question that was loaded. You can delete a load from a question if you right-click the question and select Delete. You must be an administrator or supervisor to delete data. When you delete loaded data, you will delete only the data that was loaded for that question, for that date and time. All of the other data remains. If you want to delete an entire load, navigate to Jobs, right-click a job, select Loads, rightclick a load, and select Delete. This action deletes all of the data from that load. Loads by Date Navigate: Supervisor/Loads The Loads by Date table shows the result of a specific file load attempt. There is one row for each load attempt for the date ranged specified. The default view is to show load activity for the past seven days. To view a different date range, select date choices from the drop-down boxes and click the OK button. The displayed list includes the following columns: Study ID The study ID. Load Date Date and time of the load. Load File Name of the file loaded. Responses Loaded Number of records loaded in total or the word "Error" if an error occurred. Comment This column is blank for successful loads or contains the error description if the load was unsuccessful. To view details of the loaded file, right-click a table entry and select Load Details. The Load Details page will display. To delete a load, right-click a table entry and select Delete. This action will delete all responses in all questions which were loaded from the file on the date and time indicated. Load Details Navigate: Supervisor/Loads/Right-click a load/Select Load Details Jobs/Right-click a job/Right-click a load/Select Load Details The Load Details screen displays the result of a load on a question by question basis. To delete responses from a question, right-click the question and select Delete. This action deletes all responses for the selected question which were loaded from the file on the date and time indicated. The Load Details screen has these fields: 321 PrintDoc01 020813 Option Description Question ID When you load data to an existing study, the question ID of the new load must match exactly the question ID of the existing study. The match is case insensitive. Question Label The question label. Responses Processed The number of responses that have been processed from the load file. Responses Loaded The number of responses added to the study from this load. Blank Responses (Ignored) Responses are ignored if they contain no data. Loaded Previously Responses are skipped if they are already loaded. Responses are skipped if you try to load a question ID/respondent ID that has been loaded previously. Responses Combined If there are responses from the same respondent to the same question, these responses are combined into a single response. The responses are separated by the ‘|’ character. Auto-Coded in Load The number of responses that were auto-coded during this load. Coded The number of responses coded in this load. Loads by Job Navigate: Jobs/Right-click a job/Select Loads This table displays each of the loads performed by the job. In many cases, there will be only one load per job, as is normal when a single file is submitted for the job. If the file submitted for the job is a zip file, there can be many loads performed by a single job. Here are the fields on this screen: Field Description Study ID The ID of the study to which the responses were loaded. Load Date The date and time the load was performed. Load File The name of the file loaded. Responses Loaded If the load was successful, this entry shows the name of responses loaded. If unsuccessful, this entry displays an error message. Comment This entry may show additional information about the load. To see more information about the load, right-click the load and select Load Details. 322 Data Management To delete a load, right-click a load and select Delete. This action will delete all responses in all questions which were loaded from the file on the date and time indicated. Load Voice Recordings No additional software or hardware is required on the part of the user to transcribe and code the voice recordings loaded into Ascribe™. Any file that can be played by Windows Media Player can be loaded into Ascribe™. First, you'll find it valuable to have the FTP (file transfer protocol) process in place for your account. Voice recording files are large, and you may have difficulty loading them through the Ascribe™ interface. After confirming that you have the FTP process in place, we’ll need to set up a load file type for your needs. Zip files will have a name corresponding to the study ID and the load file type. For example: 12345.wav.zip. This tells Ascribe™ which study the file is to be loaded to and the type of data files that are to be loaded. The recordings inside the zip file should be named similar to RID.QID.wav. Language Logic can customize or tailor this component, but the important thing to remember is we need the two important bits of information, Respondent ID and Question ID. An example: 10001_Q2.wav or 10002_Q2.wav. See Translate/Transcribe Verbatims to learn how to transcribe and code voice recorded verbatims. Load Image Files When you load image files, Ascribe™ requires the file to be named in this format: RID_QID.TIF. Here are two examples: 00001_q23.tif 42008_gender.tif Next, create a compressed archive of the TIF files to be loaded to Ascribe™. Name the archive with this format: StudyID.TIF.ZIP. Here is an example: 376029.TIF.ZIP Create the study in Ascribe™ so that the study ID matches the study ID of the load file (i.e., 376029.) Load the compressed archive using the FTP loader. Auto FTP Setup Users can load data files automatically via an FTP site setup to receive files from a secure location. To set up the FTP site, Language Logic Support needs the IP address from the sending computer. The IP address can be one static address or a range depending on your setup. For security purposes, Language Logic will only accept data files from known IP addresses. We can define a single IP address, a list of addresses, or range(s). Once we set up the FTP site, you will send the files to: media.languagelogic.net/<account>, where <account> is your Ascribe™ account ID. A username is required for the site. The default username for existing accounts will be the prefix ftp- added to the FTP folder name. The FTP folder name is also the name of the Ascribe™ account. 323 PrintDoc01 020813 A password is optional and may be specified by the account administrator; it can be any text. Here is an example of an FTP username & folder, using training-2 as the Ascribe™ account name: Username: ftp-training-2 Password: Optional (meaning it is not required) FTP folder: training-2 This folder is a write-only folder. Once your file lands here, it is immediately sent to the site for loading. To successfully transfer files and load them into a project in Ascribe™, you must follow the file naming standards. In order to load a data file into a project/study, the file name must contain the Ascribe™ study ID as the first portion of the file name. The study ID is delimited by a client defined character and may be followed by any additional file details. By default the defined delimiter is a period "."; however, this can be changed by an account administrator on the "Account Options" page in Ascribe™ or for specific file types, in the load file type editor (see Load File Types.) An example file name might be proj123.12-31-05.txt where proj123 is the Ascribe™ study ID, 12-31-05 is general file info not used by Ascribe™, and . txt is the load file type. Auto-Apply Codes Navigate: Supervisor/Studies/Right-click study/Select Questions/Right-click question/Select Auto-Apply Codes The auto-code function is used to automatically code closed-ended questions and the closed-ended portions of other specify questions. This is also helpful in coding brand lists and the open ended portions of "other" and "specify" type questions. You may instruct Ascribe™ to automatically code the responses or you may add new codes to the codebook and have them automatically applied to the responses. There’s a special code to use for other specify questions that will concentrate all of the verbatims in a single code. Auto-Code Closed-Ended Questions Navigate: Supervisor/Studies/Right-click study/Select Questions/Right-click question/Select Auto-Apply Codes Closed-ended questions have a pre-defined response list. For example, you may have a likes rating scale from 1 to 5. Ascribe™ views the response list just like any other codebook. To auto-code a closed-ended question, set up the codebook/response list first. Make sure that the input code values match the responses’ values that were generated in the interview. Once the codebook is ready, right-click the question, and select Auto-Apply Codes. On the next screen, press the Auto-Apply Codes button. Ascribe™ will match the responses to the input code values. The Auto-Apply Codes screen has these fields: 324 Data Management Field Description Response The data values found in the responses. Respondents The count of respondents whose response matched the input value of the question’s codebook. Matched Code The code description. Codes Currently Applied The number of codes currently applied. Zeroes mean that no codes are applied. If one or more data values do not match a codebook entry, the Matched Code column contains the phrase "Enter New Code.” You can add codes on this screen. Move the cursor into the Matched Code column and type the text for the new code. Do this for as many codes as necessary. Click the Add Codes button to add the newly-entered codes to the codebook and automatically apply the codes. After you click the Auto-Apply Codes button, the screen displays the updated information to confirm the coding. Auto-Code Other Specify Questions Other specify questions are really hybrid closed-ended and open-ended questions, with the pre-list representing the closed-ended responses and the other representing the verbatims. Typically there is a labor-intensive step in the tabulation process to merge the closedended, pre-list data with the coded verbatim data. Merging the data back together is made difficult by the fact the codes in the pre-list and the codes generated from coding are the same. Ascribe™ loads closed-ended and open-ended data into an other specify question. The closed-ended data is automatically coded, the open-ended coded as normal, and the resulting data set requires no merge step. When data is loaded, Ascribe™ concatenates all of the data for a single respondent in a single load file together, so that you’ll see a record that looks like this. 325 PrintDoc01 020813 10122:1|3|Grand Union brand |Super G - Giant Food's brand name. In this case respondent 10122 chose 1 and 3 from the pre-list (1= Dixie; 3= Solo), and then had two others. The | symbol is the response separator. If it is encountered in the incoming response text, Ascribe™ converts it to a /. Auto-code will code the 1 and 3 automatically as Dixie and Solo. The two other responses can be separated into a special code. The auto-code function organizes your responses by the unique strings found in them and then provides a way for you to code them. With other specify questions, there is a special function that allows you to load the closed-ended responses as pre-coded open ends, and then consolidate all of the open ends in a special code. Here is how it works: Typically an other specify will have a "pre-list" that represents the closed-ended portion of the question. In the following example, the pre-list consists of: Dixie = 1 Hefty = 2 Solo = 3 Chinet = 4 Store Brand = 5 The pre-list data is loaded as the response code numbers (i.e. 1 = Dixie.) The codebook that you build for this question will consist of the pre-list, and the other codes that you build when you examine the responses. In the pre-list, the input code value must match the incoming data (i.e., in the Dixie code, the input value must be 1.) Build a "special" code with the | as its input value. All of the open ends will be concentrated in this code. Execute Auto-Apply code, and you’ll see a screen like this: You can enter a new code for responses that do not match the pre-list codes. When you press the Add Codes button, you will code everything that has text in the Enter new code area. After you’ve pressed the Add Codes button or the Apply Codes button, you’ll see a screen that looks like this. 326 Data Management The green numbers mean that there are uncoded responses in this group of responses. Black numbers mean that there are no uncoded responses. Next, you can use Review Responses to code the verbatim portion of the other specify question. Auto-Code Text Responses on Load This feature allows the user to elect for Ascribe™ to attempt to auto-code textual responses after loading. This function works very similar to the "Code Duplicates" option in the Coder Interface, but it looks for exact textual matches across shared codebooks. This function also works similar to Phrase Analyzer, but it compares the data across studies with shared codebooks. In essence, it allows the user to code a unique response one time, and after that, the loader compares and automatically codes the responses after they are loaded. It is helpful for a few scenarios: Code across shared codebooks within a study Code across shared codebooks between two or more studies Code incremental loads of data Some considerations: "See previous response" comments – if you refer to the previous response to code this type of answer, then you might want to leave them for coding after your last load. Positive/Negative/Neutral Nets -- Auto-coding textual responses does not work well for codebooks that have positive, negative and neutral nets. If the code applied depends on a previous question, then auto-coding textual responses should not be used. For example, you have a believability question that refers back to rating question. The response "price" could be positive, negative or neutral depending on the closed-end rating question. The data coded by the load coder will not be attributed to a user; instead, it will be attributed to "loader." This allows the user to distinguish between responses coded by a person and responses coded by Ascribe™. Use the Load Coder Set your study up and share codebooks where necessary. For tracker studies, it is usually best to create your new study by merging it with the previous study. 327 PrintDoc01 020813 After loading a file, a post-load process will compare the coding source of all previously coded responses using the same shared codebook with the responses from newest load. While comparing the responses, if there are any inconsistencies in coding of a duplicate response, load coder will skip the response rather than apply a code. You can check for inconsistencies in coding before you load more data. See the Auto-Coder Script for more information. Ways to Run Load Coder There are several ways you can run the load coder: Manual Load – Check the "auto-code text responses" box before you load your data file. Automated Load – Use Account Options to set the "auto-code text responses" box. Set up a special load file type for auto-coding textual responses (works on manual and automatic loads.) Hint If you need to auto-code textual responses sometimes but not always, you could set up a load file type called auto.txt (for text files) and auto.xls (for Excel™ files.) Then, just rename the files you wanted to auto-code with the extension .auto.txt or .auto.xls. 328 Data Management Remember, load coder compares the newly loaded data with the coding source of all previously coded data across shared codebooks (across studies, across questions). You do not have the choice to compare the coding with a certain study. The load coder identifies which codebook the question uses and compares across all shared codebook responses. Download Data Navigate: Supervisor/Studies/Right-click study/Select Download Data/Select Codebooks and Results Client/Studies/Right-click study/Select Download Data/Select Codebooks and Results Ascribe™ has several standard formats for downloading data. To download codebooks or study results, set the desired options on this page and click the OK button. You can select which questions are included in the download. The table at the bottom of the page displays the questions. To change the questions selected, check the desired question types in the grey bar, and click the Update button. Or you can select individual questions in the Select column. The data included in the download is drawn from only the questions selected in this table. Download Study Results 329 PrintDoc01 020813 Navigate: Supervisor/Studies/Right-click study/Select Download Data/Select Codebooks and Results/Select Download Results Client/Studies/Right-click study/Select Download Data/Select Codebooks and Results/Select Download Results Results may be downloaded in any of a number of formats. If you choose column binary, the codes will be interpreted as column/punch (e.g., 141 = column 14, punch 1). You may also choose to include the actual verbatim in any of the ASCII data formats. For larger files it is recommended that you write the results to a file and download the file, rather than cutting and pasting which are more convenient with smaller files. To download study results, select one of the formats from the list: Format Description Table format, one row for each response This format gives a file with one text row per response. Each row contains the question ID, the respondent ID, and the output code value of the codes applied to the response. Because a variable number of codes can be applied to each response, the number of fields in each row is variable. Fields are separated by tab characters. Table formats may be pasted directly into Word. Comma separated variable format, one row for each response This format is identical to the above option, except that fields are separated by commas, and all fields except the respondent ID are enclosed in quotation marks. Comma separated variables provide fast and easy integration into Microsoft products and other applications. Table format, one row for each code This format gives a file with one text row per code. Each row contains the question ID, the respondent ID, and the output code value of the code applied. If a given response has more than one code applied, it will be written to successive rows. The number of fields is fixed at three. Fields are separated by tab characters. Table formats may be pasted directly into Word. Comma separated variable format, one This format is identical to the above option, except that fields are separated by commas, and all fields except the respondent ID 330 Data Management row for each code are enclosed in quotation marks. Comma separated variables provide fast and easy integration into Microsoft products and other applications. Column binary format This format produces a file suitable for use by tab packages such as Quantum. The properties set in the Study Edit page and the Edit Questions page determine the layout of data in the file. CfMC format This format produces output in a format compatible with the CfMC products. User defined You can define your own format. It is commonly used to generate statements that instruct other software to insert data into specific locations. For example, you can use an emit statement in Quantum. For more information, see User Defined Study Results. The first four formats listed allow you to also include the verbatim response in the output when you click the check box for this option. User Defined Study Results Navigate: Supervisor/Studies/Right-click study/Select Download Data/Select Codebooks and Results/Select Download Results/Select User Defined/Click New or Edit Client/Studies/Right-click study/Select Download Data/Select Codebooks and Results/Select Download Results/Select User Defined/Click Edit The user defined option under download data and results allows the user to customize the data output from Ascribe™ by specifying the text to output, and include substitution values that are replaced with information from the study. Note Ascribe™ has a more powerful way to create custom output files using scripted output. If you cannot create the output format you need in this window, check with Language Logic Support about a custom-written scripted output. As Ascribe™ processes the results, it writes your output values at the start of each question, the start and end of each respondent, and the start of each code. To create your file, fill in the form with a mixture of text and substitution values. Any text that is not a substitution value is output 'as is'. Entry Description Name The name of this user defined results specification. This name will appear in the drop-down list of user defined results available to all users. Start of a Question This value is output at the beginning of each question. Start of a Respondent This value is output at the start of each respondent to the question. Code This value is output for each code applied to the response for a respondent. End of a Respondent This value is output after all codes for the response. 331 PrintDoc01 020813 End of a Question This value is output after all responses to a question. By default, only responses that have codes applied are included in the output. If you want all responses regardless of whether or not they have been coded, click the Include uncoded responses check box. To include output that has double-byte characters (i.e., Chinese and Korean), click the Unicode check box. /* Dislikes if (c(101,105).eq.10001) emit + c0420'8' + c0420'2' if (c(101,105).eq.10015) emit + c0420'8' + c0421'8' if (c(101,105).eq.10023) emit + c0420'8' + c0421'8' if (c(101,105).eq.10035) emit + c0420'8' if (c(101,105).eq.10093) emit + c0420'8' if (c(101,105).eq.10095) emit + c0420'8' if (c(101,105).eq.10096) emit + c0420'8' This code instructs a TAB package to look in columns 1 through 5, and if it matches the respondent number, to insert the relevant punches in the relevant columns. Substitution Values for User Defined Study Results Here are the substitution values for the user defined study results: 332 Data Management Value Description [ CColumn] Code column (all characters of code value except punch) This gives all characters of the code Output value except for the rightmost character. For example, if the code output value is 1234, CColumn is 123. [CColumn+] QColumn incremented by QColumns for each code applied This gives the QColumn value for the first code applied to a response, then QColumn + QColumns for the next code, and so on. The QColumns value is added after each code. This is useful for Numeric output for column binary format, to determine the column value for the code. [CEndColumn+] (QColumn + QColumns - 1) incremented by QColumns for each code applied - This is similar to [CColumn+], but gives the end column for the code instead of the start column. [CHelp] Code help - The Long help text for the code. [CHover] Code hover help - The Hover help text for the code. [CInputID] Input code value - The Input code value. [CLevel] Level of code in code book (1 = top level) - The indentation level of the code in the codebook, where the value 1 means 'top level'. [ColPunch] Same as [CColumn]'[CPunch]' - This is simply a shorthand for a common way of showing column/punch style codes. A code with Output code value 1234 will produce 123'4' [COutputID] Output code value - The Output code value of the code [CNumber] Code row number - The position of the code in the codebook, where the first code is number 1, the second 2, and so on. The indentation level of the code is ignored. It is simply the vertical position of the code in the codebook. [CPunch] Code punch (rightmost character of code value) - The code punch is meaningful if you are using punch format for column binary output. By convention, the code punch in this format is the rightmost character in the Output code value. The remaining digits to the left are the column. For example: 123 = punch 3, column 12 645x = punch x, column 645 [CRegExp] Regular expression - The regular expression for the code. [CText] Code description - The description of the code. This is the text associated with a code. [NL] Insert line break - This value ends the current line and starts a new line. You must insert [NL] values unless you want all of the output on a single line. [RID] Respondent ID - The respondent ID: the identifier for the respondent in the loaded data. 333 PrintDoc01 020813 [RTrans] Transcription - The transcription text associated with this response. [RVerbatim] Verbatim response - The text response from the respondent. [QCard] Question card - Outputs the card number of the question. [QID] Question ID - Outputs the Question ID text. [QColumns] Question columns - Outputs the question columns value. [QColumn] Question column - Outputs the question column value. [QHelp] Question help - Outputs the Question help text. [QLabel] Question label - Outputs the Question label text. [QMaxCodes] Maximum number of codes to output - Outputs the Maximum codes value for the question. [QText] Question text - Outputs the question text. [SCardCols] Columns per card/respondent - Outputs the Columns per card/respondent value for the study. [SCardNumCol] Card number column - Outputs the Card number column value for the study. [SCardNumCols] Card number columns - Outputs the Card number columns value for the study. [SDesc] Study description - Outputs the Study description text. [SID] Study ID - Outputs the Study ID text. [SName] Study name - Outputs the Study name text. [SRIDCol] Respondent ID column - Outputs the Respondent ID column value for the study. [SRIDCols] Respondent ID columns - Outputs the Respondent ID columns value for the study. [Tab] Insert a tab character - This value puts a tab character in the output. It is useful when you want to import the data to Excel™. Excel™ uses a tab character to mean 'start a new column'. Download Codebooks Navigate: Supervisor/Studies/Right-click study/Select Download Data/Select Codebooks and Results/Select Download Codebooks Client/Studies/Right-click study/Select Download Data/Select Codebooks and Results/Select Download Codebooks 334 Data Management You can download the codebooks for any or all questions in the study. Codebooks may be downloaded in several formats. To download a codebook, select one of the formats from the list: Format Description Table format, one row for each code This format gives a file with one text row per code. Each row contains the question ID, the respondent ID, and the output code value of the code applied. If a given response has more than one code applied, it will be written to successive rows. The number of fields is fixed at three. Fields are separated by tab characters. Table formats may be pasted directly into Word. Comma separated variable format, one row for each code This format is identical to the above option, except that fields are separated by commas, and all fields except the respondent ID are enclosed in quotation marks. Comma separated variables provide fast and easy integration into Microsoft products and other applications. User defined You can define your own format. It is commonly used to generate statements that instruct other software to insert data into specific locations. For example, you can use an emit statement in Quantum. For more information, see User Defined Codebook Formats. User Defined Codebook Formats Navigate: Supervisor/Studies/Right-click study/Select Download Data/Select Codebooks and Results/Select Download Codebooks/Select User Defined/Click Edit or New Client/Studies/Right-click study/Select Download Data/Select Codebooks and Results/Select Download Codebooks/Select User Defined/Click Edit The user defined codebook output format allows the user to construct a custom file of codebook information. The user can specify the text included in the output, and include substitution values that are replaced with information from codebooks. Entry Description Name The name of this user defined codebook specification. This name will appear in the drop-down list of user defined codebooks available to all users. Start of the Study This value is output first, and will appear on the first line of the file. Only study substitution values are legal. Start of a Question This value is output at the start of each question, and before any codes in the question. Study and question substitution values are legal, but code substitution values are not. Start of a Net This value is output at the start of each net (any code that has children). End of a Net This value is output at the end of each net (any code that has children). Codes that are children of this net are output between the Start of a net and the End of a net. Code This value is output for each code that has no children (is not a net). Example of User Defined Codebook: 335 PrintDoc01 020813 Static Hair Master - Punch using Column Offset q7 7. What, if anything, did you particularly LIKE about the Static Free product? (PLEASE BE AS SPECIFIC AS POSSIBLE) Efficacy (Net) Appearance to Hair ( Subnet) 1 20 1 Straightens hair 1 20 2 All other appearance to hair mentions Control ( Subnet) 1 20 3 Controls fly away hair 1 20 4 Controls frizziness 1 20 5 Controls static 1 20 6 Manageable hair 1 20 7 All other control mentions Static ( Subnet) 1 20 8 Eliminates Static 1 20 9 Eliminates winter static 1 21 0 All other static mentions Miscellaneous Efficacy 1 21 1 Works well 1 21 2 Works quickly 1 21 3 All other efficacy mentions Substitution Values for User Defined Codebook Formats Here are the substitution values for user defined codebook formats: Value Description [NL] Insert line break -This value ends the current line and starts a new line. You must insert [NL] values unless you want all of the output on a single line. [Tab] Insert a tab character - This value puts a tab character in the output. It is useful when you want to import the data to Excel™. Excel™ uses a tab character to mean 'start a new column'. 336 Data Management Codes [CChildren] Count of children (sub-codes) of this code - This is the number of codes that are directly below the current code (including nets), but it does not include codes below those codes. In other words, it is a count of all codes at the next indentation level in the codebook, but not at lower levels. If you want a count of all codes below the current code regardless of level, use the [CDescendants] value. [CColumn] Code column (all characters of code value except punch) - The code column is meaningful if you are using punch format for column binary output. By convention, the code column in this format is all digits in the Output value except for the rightmost digit. The rightmost digit is the punch. For example: 123 = punch 3, column 12 645x = punch x, column 645 [CDescendants] Count of descendants (sub-codes and their sub-codes) of this code - This is a count of all sub-codes of the current code (including nets), regardless of level. If you want a count of only the codes at the next level, use the [CChildren] value. [CGroup] Code group number (e.g. 1.3.2) - The code group number is an outline-style number for the net that contains the code. The value 1.3.2 means 'this code belongs to the second net below the third net below the first net in the codebook'. The CGroup is the same for each sub-code of a given net. [CHelp] Code help - Outputs the long help for the code. [CHover] Code hover help - Outputs the hover help for the code. [CID] Output code value - This is the same as [COutputID] [CIndent] Insert [CLevel] tab characters - This value writes the number of tab characters corresponding to the indentation level of the code in the codebook. This is useful for creating a formatted view of the codebook, when you want to put a copy of the codebook in a text document. [CInputID] Input code value - Outputs the input code value for the code. [CKey] Key number for code (guaranteed unique) - Outputs a number that is guaranteed to be unique for each code in the codebook. You should consider this to be a random number. It will not change as long as the codebook is stored in Ascribe™. It will change if you save the study and restore it. [CLevel] Level of code in code book (1 = top level) - This is the indentation level of the code in the codebook. The value 1 means this is a toplevel code. Sub-codes of top level codes have CLevel = 2, etc. [CList] List of all descendant code values (but no nets), separated by commas - Outputs the COutputID value for each sub-code of this code, at all levels below this code. Nets are not included in this list. Each COutputID is separated by a comma. [CNumber] Code row number - This is the position of the code in the code book, where the first code is number 1, the second 2, and so on. 337 PrintDoc01 020813 The indentation level of the code is ignored. This is simply the vertical position of the code in the codebook. [ColPunch] Same as [CColumn]'[CPunch]' - For example, if the Output code value is 1234, it will output 123'4' [COutputID] Output code value - Writes the Output code value for the code. [CPunch] Code punch (rightmost character of code value) The code punch is meaningful if you are using punch format for column binary output. By convention, the code punch in this format is the rightmost character in the Output code value. The remaining digits to the left are the column. For example: 123 = punch 3, column 12 645x = punch x, column 645 [CPunchList] List of all descendant code [ColPunch] values (but no nets), separated by commas - This is similar to CList, except that the values in the list are in the ColPunch format, for example 34'5' [CPunchListOr] c23'534'.or.c26'18-x'.or.c34'2579'... This is similar to CPunchList, except that a 'c' character precedes each code, and the codes are separated by .or. [CRegExp] Regular expression - Outputs the Regular Expression for the code. [CText] Code description - Outputs the Description of the code. [CCodesApplied] Codes applied - Outputs the number of times this code has been applied to responses in the current question. Questions [QCard] Question card - Outputs the card number of the question. [QCBKey] Key number of code book (guaranteed unique) - Outputs a number that is guaranteed to be unique for each codebook. You should consider this to be a random number. It will not change as long as the codebook is stored in Ascribe™. It will change if you save the study and restore it. If the codebook is shared, the same number is output for each instance of the codebook. For example, if questions A and B share the same codebook, the same value for QCBKey will be output for question A as for question B. [QCodesApplied] Codes applied - Outputs the number of codes that have been applied for this question, totaled across all responses. [QColumn] Question column - Outputs the question columns value. [QID] Question ID - Outputs the Question ID text. [QKey] Key number for question (guaranteed unique) - Outputs a number that is guaranteed to be unique for each question. You should consider this to be a random number. It will not change as long as the study is stored in Ascribe™. It will change if you save the study and restore it. [QLabel] Question label - Outputs the Question label text. [QResponsesCoded] Responses coded - Outputs the number of responses that have 338 Data Management been coded for the current question. Responses after reclassify - Outputs the current number of responses stored for this question. It is possible for this number to [QResponsesCurrent] be larger than the number of responses loaded, because a response may have been reclassified to this question from another question in the study. Responses loaded - Outputs the number of responses that were loaded for this question. It is possible for this number to be [QResponsesLoaded] smaller than the number of responses currently stored for this question, because a response may have been reclassified to this question from another question in the study. [QResponsesRef] Responses currently referred - Outputs current number of responses to this question that have been referred to supervisor for help. [QText] Question text - Outputs the question text. [QType] Question type (0: Open, 1: Closed, 2: Other specify, 3: Value) Outputs a number as listed above that specifies the type of the question. If you want the name of the question type instead, use [QTypeName] [QTypeName] Question type - Outputs the name of the question type, for example Closed. Study [SCardCols] Columns per card/respondent - Outputs the Columns per card/respondent value for the study. [SCardNumCol] Card number column - Outputs the Card number column value for the study. [SCardNumCols] Card number columns - Outputs the Card numbers column value for the study. [SClosedQuestions] Closed questions - Outputs the number of closed questions in the study. [SResponsesCoded] Responses coded - Outputs the number of responses that have been coded for all questions in the study. [SCodesApplied] Codes applied - Outputs the number of codes applied for all questions in the study. [SDesc] Study description - Outputs the Study description text. [SID] Study ID - Outputs the Study ID text. [SName] Study name - Outputs the Study name text. [SOpenQuestions] Open questions - Outputs the number of open questions in the study. [SOSQuestions] Other specify questions - Outputs the number of Other specify questions in the study. [SResponses] Responses - Outputs the total number of responses to all 339 PrintDoc01 020813 questions currently stored for the study. [SRespondents] Respondents - Outputs the total number of respondents to all questions currently stored for the study. This is the number of responses that have a unique respondent ID. [SRIDCol] Respondent ID column - Outputs the Respondent ID column value for the study. [SRIDCols] Respondent ID columns - Outputs the Respondent ID columns value for the study. [SValueQuestions] Value questions - Outputs the number of Value questions in the study. Download Data with Scripted Output Navigate: Client/Studies/Right-click a study/Select Download Data/Select Scripted Output Supervisor/Studies/Right-click a study/Select Download Data/Select Scripted Output Scripted output provides very flexible custom output files to generate data and reports that are not already available in Ascribe™. By default, every account has at least 5 outputs available. Ascribe™ has the ability to create custom output files using a scripting language. This allows Ascribe™ to create very flexible custom output in virtually many formats. Scripted output is available from the Client/Download data menu. You have the ability to run output scripts, but you cannot create or modify them. Language Logic can create custom scripts to suit your needs. Contact Language Logic to request a custom output script. The Scripted Output page lists the Study ID, study name, and number of questions in a table at the top of the page. The Study ID has a drop-down box which contains information about the study. The Questions field has a drop-down box which lists the questions in this study. There are drop-down boxes for the Question ID, Codebook ID, and Shared Codebook columns which give more information about these fields. To execute a scripted output, select the desired program from the drop-down menu and check the questions you wish to include. The following reports will be available by default: Auto-Coder Script– Auto-codes exact textual matches or creates the coding consistency report Column Binary Output Script – Tabulation Data 340 Data Management Excel™ Data Output –Data in Excel™ format Excel™ Horizontal Verbatims – Output with RID in first column and QID’s in first row HTML Codeframe Output – Marginal Report in HTML format HTML Data Layout – Data Map as laid out in Ascribe™ Marginal CSV Report (duplicated nets) – Same as HTML Codeframe only net counts are not unduplicated and the format is . csv rather than HTML Max Codes Report – reports those respondents that have more codes assigned than the max code setting allows (see max codes under question setup) SAV File Export – Creates a SAV file for loading into SPSS and/or Dimensions Translation Consistency Report – Reports inconsistencies in translations of duplicate verbatims Auto-Coder Script Navigate: Client/Studies/Right-click study/Select Download Data/Select Scripted Output/Select Auto Coder Supervisor/Studies/Right-click study/Select Download Data/Select Scripted Output/Select Auto Coder This script is used to auto-code exact textual matches or to check for consistency in coding. Before you auto-code a study, you can check to see if the coding is consistent. For example, the coding consistency report can be useful for a study with multiple data loads. In the first load, these responses were coded like this: Respondent 100 "chocolate taste" was coded with code 8: Chocolate taste Respondent 110 "chocolate taste" was coded with code 21: Chocolate Respondent 150 "chocolate taste" was coded with code 8: Chocolate taste When you process a second load, any respondents who said “chocolate taste” will not be coded automatically because of the inconsistency in the first load. If you run the coding consistency report after coding of the first load, you would find the inconsistency. In this example, you would fix the inconsistency by changing Respondent 110 from code 21 to code 8. Then, when you load any new respondents who said "chocolate taste," they will be coded automatically with code 8. To run the report, select Auto Coder from the Scripted Output screen and click the Ok button. The next screen has these fields: Field Description Study List List of study IDs you wish to compare for auto-coding; leave it blank to only check the current study. Delimiter Character that separates the study IDs in the Study List. It defaults to a comma. Report or Code Enter "report" to generate a consistency report or "code" to autocode responses. Case Insensitive Comparison To compare responses and ignore case, enter "true" otherwise enter "false." Order of Mention If the order of codes applied is required criteria, enter "true" otherwise enter "false.” 341 PrintDoc01 020813 Match Question ID or Shared Codebook To only code or report on matching questions, enter " QuestionID" otherwise enter "codebooks." If you enter “ QuestionID,” you have more control over what is to be autocoded. If you enter “codebooks,” the question ID does not matter. This option allows you to auto-code after loading data without having to reload the data. After you enter the fields, click the OK button to run the report. After the report runs, rightclick the link and select the appropriate option. Column Binary Output Script Navigate: Client/Studies/Right-click study/Select Download Data/Select Scripted Output/Select Column Binary Supervisor/Studies/Right-click study/Select Download Data/Select Scripted Output/Select Column Binary This script generates a data file from the coded responses sometimes called "Tabulation Data." Account options or study setup determines how the data will be output when you choose to run this script. On the Data Tab of the Study Edit screen, choose your desired file format (almost always ASCII) and layout formats (see Layout Types.) Layout Types There are three layout types available for questions. These layout types are used only for column binary output. You can change the layout type if you edit the study. If the layout setting for the study is use question settings, then you can specify different layout types for each question in the study if you edit the question. In column binary output, rows in the output file represent cards, and data locations on the cards are called columns. This terminology dates from the use of IBM punch cards for data. Column numbering starts at 1 and extends to the value of card number columns specified for the study. In addition to code values, the respondent ID and card number are also output for each card. Ascribe™ always outputs one row (card) in the data file for each unique combination of respondent ID and card number. The respondent ID and card number fields are always output in numeric layout, regardless of the layout setting for the study and questions. Numeric Layout Ascribe writes a card for each unique combination of respondent ID and card number. The Card field on the Data tab of the Edit Question page determines the card number for that question. While numeric layout sounds as if it can only output numbers, you may use any ASCII character for the output ID of the codes. In a numeric layout, the column binary output script contains the respondent ID followed by the code output IDs. The output is written across the card in this format: 342 Data Management Field Description Respondent ID The respondent ID starts in the column indicated by the Respondent Column on the Data tab of the Edit Study page. The length of the respondent ID is controlled by the Respondent ID Columns on the Data tab of the Edit Study page. Starting Column The first code for a question starts in the column indicated by the Column field. The Column field is on the Data tab of the Edit Question page. Code Length The length of the code output ID is set by the Columns field, which is on the Data tab of the Edit Question page. If the output ID is less than the Columns field, it will have leading zeroes. Successive Codes Codes for each question follow one another. Here are some examples: Question 1 has 20 in the Column field and 3 in the Columns field. The first code is written in columns 20, 21, and 22. The next code is written in 23, 24, and 25, and so on. Question 2 has 50 in the Column field and 3 in the Columns field. Number of Columns Used by a Question The first code for Question 2 is written in 50, 51, and 52. The next code is written in 53, 54, and 55, and so on. You can see the number of columns used by each question on the Card Layout Report. The number of columns used by a question depends on the Maximum Codes field on the Data tab of the Edit Question page. If the Maximum Codes field equals zero, an unlimited number of codes may be written to the output, and you can not determine in advance the number of columns a question will use. If the Maximum Codes field is greater than zero, the number of codes written to output is limited to that number. (The Maximum Codes field does not affect the number of codes a coder may apply; it only affects output.) The number of columns used equals Maximum Codes multiplied by the Columns field. Punch Layout Ascribe writes a card for each unique combination of respondent ID and card number. The Card field on the Data tab of the Edit Question page determines the card number for that question. In a punch layout, the column binary output contains the respondent ID followed by a single digit which represents the code output ID. The output is written across the card in this format: 343 PrintDoc01 020813 Field Description Respondent ID Punch Value of the Code Output ID A single bit or digit represents each applied code in the output file. This bit is the rightmost digit of the code output ID and is called the punch value. Punch Column The respondent ID starts in the column indicated by the Respondent Column on the Data tab of the Edit Study page. The length of the respondent ID is controlled by the Respondent ID Columns on the Data tab of the Edit Study page. For example, output ID 123 has a punch value of 3. The column where the bit is written is determined by the digits in the output ID to the left of the punch value. For example, output ID 123 would have a 3 written or punched at column 12. Punch layout also has these rules: The rightmost digit of the output ID (the punch value) must be a digit or the characters ‘x’, ‘X’, ‘y’, ‘Y’, ‘&’, or ‘-‘. The column characters in the output ID must be digits. A single digit is not allowed to be an output ID. Ascribe interprets a single digit output ID as having zeroes in front of the punch value, and there is no column 0 on a card. The Card Layout Report detects this problem, and the Column Binary Script will fail if the problem is not corrected. Duplicate code output IDs are not allowed. Duplicate output IDs would punch the same value in the same column for both codes. There would be no way to determine which code was output. The Card Layout Report detects this problem, and the Column Binary Script will fail if the problem is not corrected. Here are some examples of output IDs and how they are used: 142 – Punch 2 at column 14. 3 – Punch 3 at column 0; not allowed since there is no column 0 on a card. 12Y – Punch Y at column 12. Punch Using Column Offset Layout Punch using column offset layout is similar to the punch layout. However, the column where the code is written or punched is determined in a different way. The output is written across the card in this format: Field Respondent ID 344 Description The respondent ID starts in the column indicated by the Respondent Column on the Data tab of the Edit Study page. The length of the respondent ID is controlled by the Respondent ID Columns on the Data tab of the Edit Study page. Data Management Punch Value of the Code Output ID A single bit or digit represents each applied code in the output file. This bit is the rightmost digit of the code output ID and is called the punch value. (This part is the same as the punch layout.) Punch Column For example, output ID 123 has a punch value of 3. The column where the bit is written or punched is determined by the digits in the output ID to the left of the punch value plus the Column field. (The Column field is set on the Data tab of the Edit Question page.) Simply drop the punch or last digit of the output code value, and then add the remaining number to the Column field. For example: The output ID is 123, and the Column field is 20. The punch value is 3. The punch column is equal to the digits to the left of the punch value (12) + the Columns field (20) or 32. 3 would be written at column 32. Another difference from the punch layout is that a code output ID can be a single digit. Here is an example: The Column field is 20, and the output ID is 5. Ascribe interprets the output ID as 005. The punch value is 5. The punch column is computed as 0 + 20 = 20. Therefore, 5 would be written at column 20. Here are some examples of output IDs and how they are used when the Column field is 10: 142 – Punch 2 at column 24 (14 + 10). 3 – Punch 3 at column 10 (0 + 10). 12Y – Punch Y at column 22 (12 + 10). Note Remember, Punch using Column Offset is dependent upon the Column field on the Edit Question page. Remember to simply drop the punch or last digit of the output code value, and then add the remaining number to the Column field on the Edit Question page. For instance, in the Question Edit box your Card is 41 and the starting column is 10. The codebook is numbered starting at code 1 and incrementing by 1. Then Codes 1 through 9 will be Card 41, Column 10, punches 1 through 9 or the equivalent of 101, 102, 103, 104, 105, 106, 107, 108, 109 in a regular Punch codebook. Codes 10 through 19 will be Column 11, punches 0 through 9 or 111, 112, 113, 114, 115, 116, 117, 118, 119 in a regular Punch codebook. 345 PrintDoc01 020813 If the last code is 68, that is an 8 punch in column 16 (drop the 8 then add 6 to 10)—but if the starting column is 25, then code 68 would be an 8 punch in column 31 (drop the 8 add 6 to 25 to get 31). Reasons to Use Punch Using Column Offset Punch using column offset gives you more flexibility when you deliver a study. You can keep codebooks shared and numbered the same, but still output each question to separate columns. You share codebooks and want to maintain the share but still output each question to a different starting column (often used for trackers.) You have a tracker that maintains the same questions with the same codebooks. You want to retain their original punches, but the column locations change each month. (This used to be a tedious process that required careful renumbering of each code.) You output Quantum Edit specs and need to write across cards. This requires both the card value and the column value to be concatenated into one number and entered as the Column value on the Edit Question screen. For instance, card 10 column 08 would be 1008. Now, to write to the next card (card 11), you simply figure out what it would take to jump from 1008 to 1108, meaning increment anything writing to card 11 by 100. An example, you have a code that the data analyst needs to fall on card 11, column 08, punch 3. The rest of the codebook needs to write to card 10 starting with column 08 punch 1. All of the codebook can be numbered normally, starting with 1 by increments of 1 except for the oddball Card 11 code. This code will be numbered 1003 because 100 plus 1008 equals 1108 which deciphers to Card 11 column 08. Study and Question Setup Example Study Setup for Punch Using Column Offset: Questions Setup for Punch Using Column Offset: 346 Data Management When setting up a study as Punch Using Column Offset, one major difference takes place from the Punch setting: the Column field from the Edit Question page becomes active. (See below). When a study is set up as Punch, Column means absolutely nothing to the deliverables at the end of the study. The data will be output the same way no matter what the Column field contains. Punch Using Column Offset changes that completely. The Column for each question tells Ascribe™ where to start writing the data. It also tells Ascribe™ that each code will be offset by that column number. This allows the codebooks to be numbered starting at 1 by increments of 1. The output code value of Column Using Punch Offset codebooks represents the punch (the single digit to the rightmost of the Output value will be the punch) and the number that the Column will be offset by (any numbers to the left of the punch will offset the Column). This is really a simple math game. Q8 below will start in column 40. The codebook is numbered 1-???. Let’s look at a few codes from the Q8 codebook. 347 PrintDoc01 020813 Code 11 will be written to Column 41 Punch 1 or 41’1. Here’s the math: Drop the punch off the code value (Remember the punch is the digit farthest to the right). That leaves us with 1. Add that 1 to the column (40) and you get column 41. Now we bring the punch back up and get 41’1. Code 14 will be written to Column 41 Punch 4 or 41’4 Here’s the math: Drop the punch off the code value (Remember the punch is the digit farthest to the right). That leaves us with 1. Add that 1 to the column (40) and you get column 41. Now we bring the punch back up and get 41’4. Code 236 will be written to Column 63 Punch 6 or 63’6 Here’s the math: Drop the punch off the code value (Remember the punch is the digit farthest to the right). That leaves us with 23. Add that 23 to the column (40) and you get column 63. Now we bring the punch back up and get 63’6. Card Layout Report for Punch Using Column Offset: 348 Data Management Tabulation Data for Punch using Column Offset: More Information about 1130 Column Binary Setup and Output Use the Card Layout Report under Client/Reports to check the layout of you data before you use the Column Binary scripted output. The Card Layout Report allows the user to check all aspects of setup to make sure the data will be written out as desired. It also tells the user where problems with the layout might exist. When downloading the data using the binary option (as opposed to the ASCII option under Study Setup), it will be written in what is referred to as 1130 column binary. This refers to the bit configuration in the resulting two-byte words. To read the file with a utility called MTR, use this syntax: mtr -1 –r192 –b192 –i<filename> -o<filename Where r = the record length b = blocking factor 349 PrintDoc01 020813 i = input filename o = output filename The record length will be 2 times the number of columns that are in your output file plus 2. In the example above, there are 80 columns of data with a record terminator of two bytes (hence the plus 2). The blocking factor should be exactly the same as the record length. Excel® Data Output Navigate: Client/Studies/Right-click study/Select Download Data/Select Scripted Output/Select Excel Data Supervisor/Studies/Right-click study/Select Download Data/Select Scripted Output/Select Excel Data This script produces a CSV file with respondent IDs down column A and question IDs across the top rows. When more than one code is applied, each code has a column. In the report, the column heading has this format: the question ID in Ascribe™, followed by the "_", and then the numeric order of the code applied. For instance, Q11_1 represents the first code applied to Q11 for each respondent. Q11_2 represents the second code applied to Q11 for each respondent. In the case below, there were not more than two codes applied to Q11. 350 Data Management Excel® Horizontal Verbatims Navigate: Client/Studies/Right-click study/Select Download Data/Select Scripted Output/Select Excel Horizontal Verbatims Supervisor/Studies/Right-click study/Select Download Data/Select Scripted Output/Select Excel Horizontal Verbatims This script creates a standard Excel ® data file that can be loaded into Ascribe™. The respondent IDs are in the first column, and the question IDs are across the top row with the verbatims below them. Example of Excel® Horizontal Verbatims: 351 PrintDoc01 020813 HTML Codeframe Output Navigate: Client/Studies/Right-click study/Select Download Data/Select Scripted Output/Select HTML Codeframe Supervisor/Studies/Right-click study/Select Download Data/Select Scripted Output/Select HTML Codeframe This script displays the code input and output values and textual descriptions as well as counts and percentages. The count of respondents represents an unduplicated net count. Example of HTML Codeframe output: 352 Data Management HTML Data Layout Navigate: Client/Studies/Right-click study/Select Download Data/Select Scripted Output/Select HTML Data Layout Supervisor/Studies/Right-click study/Select Download Data/Select Scripted Output/Select HTML Data Layout This script displays the data layout information of the study. Example of HTML Data Layout: Marginal CSV Report Navigate: Client/Studies/Right-click study/Select Download Data/Select Scripted Output/Select Marginal CSV Report Supervisor/Studies/Right-click study/Select Download Data/Select Scripted Output/Select Marginal CSV Report This script produces a CSV file similar to the HTML Codeframe script, but the net counts are not unduplicated. The “Dropped” column represents instances where more codes were applied than allowed by the Max Code setting. These are the codes that were dropped. 353 PrintDoc01 020813 Example of Marginal CSV Report: Max Codes Report Navigate: Client/Studies/Right-click study/Select Download Data/Select Scripted Output/Select Max Codes Report Supervisor/Studies/Right-click study/Select Download Data/Select Scripted Output/Select Max Codes Report This script creates and HTML document that details any respondent that has more codes applied than allowed by the Max Codes property for each question. Example of Max Codes Report: 354 Data Management SAV File Export Navigate: Client/Studies/Right-click study/Select Download Data/Select Scripted Output/Select Save File Export Supervisor/Studies/Right-click study/Select Download Data/Select Scripted Output/Select Save File Export This script creates a standard SAV file to load into SPSS. Translation Consistency Report Navigate: Client/Studies/Right-click study/Select Download Data/Select Scripted Output/Select Translation Consistency Supervisor/Studies/Right-click study/Select Download Data/Select Scripted Output/Select Translation Consistency This scripted output should be used in conjunction with translating textual responses. This report lists responses that have identical verbatims but different translations. Example of Translation Consistency Report: 11 unique FROM text, 12 unique TO text 1: (2 instances) elefante|vaca 355 PrintDoc01 020813 1: (0001, Exercise7a, 10002, 1 instances) Elephant 2: (0001, Exercise7a, 10034, 1 instances) elephant Manage Jobs Navigate: General/Jobs Ascribe™ is able to perform certain jobs for you while you are working on other things. This report shows you these jobs, their status (such as running), and the result. Jobs are processed in the order they are submitted. With the exception of scripted output jobs, only one job will run at a time in a given Ascribe™ account. This is because a job may depend on the completion of a previous job for it to execute properly. For example, you may restore a study as a template for a new study (as job 1), then load data to the new study (as job 2). Clearly job 2 cannot run at the same time as job 1. By ensuring that jobs run in the same order they are submitted, Ascribe™ allows you to start multiple jobs in a logical sequence without waiting for the completion of the jobs started first. Scripted output jobs will not start while another type of job is running. However, once a scripted output job is running, it will not prevent other jobs from starting (including other scripted output jobs). To see only jobs that you started, click the My Jobs Only box. You can view jobs by their creation date. To change the range of dates, use the drop-down list or edit the dates. The Jobs screen has these fields: Field Description Job Type Save study - Indicates a study was saved from the Ascribe™ software. Right-click "save study" to download the saved study. Restore study - Indicates a study was restored from a zipped .XML file (usually a file that originally was saved from Ascribe™.) Check the file name to see the original study ID and name ( studyid_name). Load from Ascribe™ - Indicates a file has been manually loaded into Ascribe™ from the "Load data/process file" option on the Supervisor/Studies page. Load from FTP - A data file (or setup file) has been loaded via an FTP site. The user should be "_loader" in these instances. Merge Study - A study was merged with a previous study. The previous study is indicated by the file name. Scripted Output - A scripted output has been run on a study. To see which study, right-click and select parameters. The type of scripted output ran is indicated in the "file" field (user defined is considered a scripted output here.) Job Status Queued - The job has been created, but execution of the job has not yet begun. Only one job can be active at a given time. Running - The job is running. 356 Data Management Paused - The job is paused awaiting action from you. Completed - The job has completed. If the job completed with an error, the row will be red. File Name The name of the file submitted to the job. For scripted output jobs, this is the name of the script. User Name The user who started the job. Time Created The time the job was created. This is not necessarily the time that the job began execution since a job may need to wait for other jobs to complete. Seconds The number of seconds that the job has run (or took to run if the job is complete.) View Detailed Information about a Job Right-click the job and select Properties from the menu, or double-click the job. The dialog box displayed contains information about the execution of the job. The Properties screen has these fields: Field Description Job Type Save study - Indicates a study was saved from the Ascribe™ software. Right-click "save study" to download the saved study. Restore study - Indicates a study was restored from a zipped .XML file (usually a file that originally was saved from Ascribe™.) Check the file name to see the original study ID and name ( studyid_name). Load from Ascribe™ - Indicates a file has been manually loaded into Ascribe™ from the "Load data/process file" option on the Supervisor/Studies page. Load from FTP - A data file (or setup file) has been loaded via an FTP site. The user should be "_loader" in these instances. Merge Study - A study was merged with a previous study. The previous study is indicated by the file name. Scripted Output - A scripted output has been run on a study. To see which study, right-click and select parameters. The type of scripted output ran is indicated in the "file" field (user defined is considered a scripted output here.) Job Status Queued - The job has been created, but execution of the job has not yet begun. Only one job can be active at a given time. Running - The job is running. Paused - The job is paused awaiting action from you. Completed - The job has completed. If the job completed with an error, the row will be red. Job Schedule Displays the name of the job if it has been added to the job schedule. User Name The user who started the job. 357 PrintDoc01 020813 Time Created Actual time a job was submitted. Time Started Time a job began running. Time Completed Time a job finished running. Total Time (Seconds) Time from submission of job to completion (includes queued time.) Run Time (Seconds) Time the job took to run. File Name Name of the file loaded or the script that was run. Message Log field for reporting information. Error Displays the error that caused the job to fail (see Error Messages or call Support for help in troubleshooting failed jobs.) View Parameters for Scripted Output Jobs Right-click the job, and then select Parameters from the menu. This menu item is available for scripted output jobs only. The dialog box displays the information passed to the job when it was started. The Parameters screen has these fields: Field Description Script Script name Study Study ID and study name Questions Question IDs and labels for those questions selected for output Parameters Some custom scripts are based on information entered by the user. (See Auto-Coder Script for an example of a script that requires parameters.) Download the Results of a Job If a job created an output file (such as a scripted output job, or a save study job), you can retrieve the file. Right-click the job, and select Download or Save Study As from the menu. The files created by jobs are not saved indefinitely. The files are available for about a week from the time the job was run. If the Download menu item is grayed out, the file is no longer available for download. Resume a Paused Job Certain jobs can pause while awaiting user input. If a job has paused, you can cause it to resume execution. Right-click the paused job. Then select Resume from the menu. This menu item is available only to users with supervisor privilege. 358 Data Management Paused jobs will not remain paused indefinitely. If you have not resumed execution of a paused job after about a day, the job will abort. Schedule Jobs Navigate: General/Jobs/Right-click job/Select Schedule Job Users with supervisor privilege can schedule certain types of jobs to run again automatically on a job schedule. To create a schedule for a job, navigate to the Jobs page, right-click the job and select Schedule Job from the menu. Click the OK button, and the Job Schedules dialog displays. The fields on the Job Schedules dialog control when the job will run. The dialog has these fields: Field Description ID The ID defaults to the name of the script, but you can change it. Enable Check the Enabled box if you want this Job to run on schedule. If the Enabled box is not checked the schedule is not active, and jobs will not run automatically based on this schedule. You can, however, run jobs that are not enabled by using the Run now command from the Job schedules window. Description Describe the job for future reference; this field is optional. The description will appear as hover help for the job schedule. UTC By default, the date and times shown in the dialog box are in Universal Coordinated Time (UTC), also known as Greenwich Mean Time. You can clear the UTC check box to change the start and end date/time values to your local time zone. If you are working in a team environment across time zones, we recommend that you use the UTC view of the schedule. This view will be the same for any team member who looks at the schedule, regardless of the time zone that person is in. Start Time Enter the start time for the job to run. End Time Enter the end time of the job. Occurrence Use the drop-down arrow to set the occurrence of your job: Daily - The job runs every N days beginning at the start date and 359 PrintDoc01 020813 time, where N is the value specified in the Every ... day(s) box. Weekly - The job runs every N weeks beginning at the start date and time, where N is the value specified in the Every ... week(s) box. The job runs only on the days of the week selected by the check boxes at the right of the dialog. Monthly - The job runs monthly, in the months specified in the check boxes at the right of the dialog. The job runs on the day of the month specified by the start date, and at the time specified by the start time. Calendars Use the calendars to choose the start and end date. Every …Day(s) Set the number of days between jobs. For example, to run the job every other day, enter the value 2. The minimum allowed value is 1, and the maximum is 365. Repeat & Every…Minutes For…Minutes If you want the job to repeat on a given day, check the Repeat box and set the number of minutes between jobs, and the length of time you want to continue the repetitions. The minimum value allowed in each of these boxes is 10, and the maximum 1440. The value in the For ... minutes box must not be less than that in the Every ... minutes box. A scheduled job will run only if all of the following conditions are true: The schedule is enabled The current date and time is not earlier than the specified start date and time No end date and time is specified, or the current date and time is not later than the specified end date and time. Note The server that runs scheduled jobs operates with a UTC clock. This is important for Weekly and Monthly schedules when specifying the day of the week and the month of the year to run a job. For example, 11PM Saturday in the Eastern time zone is Sunday on the UTC clock. When working with Weekly or Monthly 360 Data Management schedules we recommend that you set up your schedule using UTC. Set Up a Daily Schedule To set up a daily schedule for the job, select Daily in the drop-down list at the left of the dialog. If you want the job to end at a particular time, click the End box. Specify the starting time in the format HH:MM, using a 24 hour clock (for example, 4:12PM is entered as 16:12). Set the desired start date in the calendar. If the End box is checked, set the ending time and date in a similar fashion. In the Every ... Day(s) box, set the number of days between jobs. For example, to run the job every other day, enter the value 2. The minimum allowed value is 1, and the maximum is 365. Set Up a Weekly Schedule To set up a weekly schedule for the job, select Weekly in the drop-down list at the left of the dialog. If you want the job to end at a particular time, click the End box. Specify the starting time in the format HH:MM, using a 24 hour clock (for example, 4:12PM is entered as 16:12). Set the desired start date in the calendar. If the End box is checked, set the ending time and date in a similar fashion. In the Every ... Weeks(s) box, set the number of weeks between scheduled weeks. For example, to run the job every other week, enter the value 2. The minimum allowed value is 1, and the maximum is 52. In the check boxes at the right of the dialog, specify the days of the week you want the job to run. If you do not specify a day of the week, the schedule will run on the same day of the week as the start date. Set Up a Monthly Schedule To set up a monthly schedule for the job, select Monthly in the drop-down list at the left of the dialog. If you want the job to end at a particular time, click the End box. Specify the starting time in the format HH:MM, using a 24 hour clock (for example, 4:12PM is entered as 16:12). Set the desired start date in the calendar. If the End box is checked, set the ending time and date in a similar fashion. In the check boxes at the right of the dialog, specify the months of the year you want the job to run. If you do not specify a month, the job will run on the same month as the start date. The job will run on the day of the month specified in the Start date. Repeat a Job If you want the job to repeat on a given day, check the Repeat box and set the number of minutes between jobs, and the length of time you want to continue the repetitions. The minimum value allowed in each of these boxes is 10, and the maximum 1440. The value in the For ... minutes box must not be less than that in the Every ... minutes box. For example, if you want a job to run on the hour every day from 8AM to 4PM, create a Daily schedule with a start time of 8AM. Check the Repeat box, and enter 60 in the Every ... 361 PrintDoc01 020813 minutes box, and 480 in the For ... minutes box. It is 8 hours from 8AM to 4PM, and 480 minutes is 8 hours. Change a Schedule Navigate: General/Jobs/Right-click job/Select Job Schedules To change a schedule, right-click the schedule and select Edit from the menu. If you have administrator privilege, you can edit any schedule. Otherwise, you can edit only those schedules that list you in the User Name column (the schedules you created.) Delete a Schedule Navigate: General/Jobs/Right-click job/Select Job Schedules To delete a schedule, right-click the schedule and select Delete from the menu. If you have administrator privilege, you can delete any schedule. Otherwise, you can delete only those schedules that list you in the User Name column (the schedules you created.) View the Parameters for a Scheduled Job Navigate: General/Jobs/Right-click job/Select Job Schedules To view the parameters for a scheduled job, right-click the schedule and select Parameters from the menu. The dialog box displays the information passed to the job when it runs. 362 Data Management Run a Job Now Navigate: General/Jobs/Right-click job/Select Job Schedules If a schedule is not enabled, you can schedule the job to run at the current time. Right-click the schedule, and select Run Now from the menu. This menu item is grayed out if the schedule is enabled. Enabled schedules can run automatically only, based on the specified schedule. When you use the Run Now command, the job will start soon, but not immediately. Ascribe™ Study Shipping Ascribe™ can ship archived studies automatically to your company’s FTP site. Those studies are automatically deleted from Ascribe™ after they are shipped. (The studies are deleted from the site when the next weekly maintenance runs.) This feature improves the usability of Ascribe™ by removing studies that are no longer active. It also automates your data archiving procedures. The Study Shipping feature of Ascribe™ operates only after specific configuration as described in this document. Until the feature is configured, your studies will not be 363 PrintDoc01 020813 automatically shipped and deleted, and archived studies will remain in the archived state until you delete them. Study Shipping Features Study shipping can help in these ways: Weekly shipments - Study shipping occurs every Saturday. Any studies with archived status are shipped at that time and then deleted from Ascribe™. Automatic save to zip file - Every study to be shipped is saved to a compressed (zip) file. This operation is identical to the interactive Save operation for a study from the Ascribe™ website. The entire study is always saved (questions, codebooks, codes applied.) Optional encryption password - If desired, the zip file can be compressed with an encryption password of your choosing. Automatic FTP transfer of the compressed study file - Each compressed study (one study per compressed file) is transferred to your FTP site. Automatic deletion of shipped studies - Upon successful completion of the save and transfer operations, the study is deleted from Ascribe™. Automatic failure recovery - If any part of the shipping process fails for a given study, it is not deleted from the Ascribe™ website. Shipped Studies Page Navigate: Supervisor/Shipped Studies You can view a list of shipped studies for your account from the Shipped Studies page. The page defaults to show studies shipped within the last week. However, you can change the default date to view a range of dates. There are two ways to select dates. You can use the drop-down list next to the Use local time box to select today, this week, this month, this year, last week, last month, or last year. When you select one of these options and click Update, the dates change to reflect your selection, and the report displays. Another way to select dates is to click the date. A calendar displays. Use the arrows next to the calendar heading to scroll through the months. Or click the heading to see a list of months. Click the heading a second time to see a list of years. Use the arrows to scroll through the months or years. The Shipped Studies page has these columns (you can choose the columns displayed by selecting Choose Columns from the right-click menu): Column Description Ship Date The date the study was shipped. Created The date the study was created. Study Name The name of the study, from the General Tab of the Study Edit dialog. File Name The file name of the shipped study. Supervisor 364 The supervisor of the study, from the General Tab of the Study Data Management Edit dialog. Client The client of the study, from the General Tab of the Study Edit dialog. End Customer The end customer of the study, from the General Tab of the Study Edit dialog. Study Description The description of the study, from the General Tab of the Study Edit dialog. Configure Study Shipping To configure study shipping, you need to provide this information to Language Logic: Data Description FTP IP address The IP address of the FTP site to which studies are shipped. May be in dot notation or a name that can be resolved to an IP address by DNS. If no value is specified, study shipping is disabled. Examples: homeworkforce.net 216.68.104.125 FTP port The port number of the FTP site. The default value is 21. FTP ship folder The name of the folder to which study files should be shipped. Must be relative to the root folder of the FTP site. The string provided here is submitted to the CD command of the FTP client used to ship the study. It is best to include no blank spaces anywhere in the string. Examples: ShippedStudies ShippedStudies/ FilesIn FTP username The username used by the FTP client to log into your FTP site. If none is provided, the username ‘anonymous’ is used. May not have leading or trailing blanks. FTP password The password used by the FTP client to log into your FTP site. If none is provided, the password ‘ LanguageLogic’ is used. May not have leading or trailing blanks. Encryption password The password used to encrypt the zipped study file. If none is provided, the study file will not be encrypted. May not have leading or trailing blanks. Example: $56hgBG# Set Up Your FTP Site 365 PrintDoc01 020813 Your FTP site should be configured according to your own best practices. The following suggestions may help you in this configuration. These suggestions are intended to help prevent malicious use of your FTP site. Required Permissions The user account for Study Shipping must have read and write privileges in the folder specified by the ‘FTP ship folder’ configuration setting. The account must also have permission to execute the CD command from the root folder to the FTP ship folder. Study Shipping does not create folders on your FTP site, and does not need permission to do so. The folder specified by the FTP ship folder setting must exist. Suggested Settings We suggest you restrict access to your FTP site by IP address, allowing only the IP address specified by the DNS name media.languagelogic.net In addition, you may wish to restrict permissions at the file system level for the account used for study shipping. Study Shipping requires only read and write access and the ability to list folder contents. Removal of the ability to create folders can prevent many common hacker attacks Maintain Your FTP Site You must move the shipped files on a regular basis to a secure location (not on the FTP site). This improves security and performance of Study Shipping. Study Shipping will eventually fail if the FTP folder contains too many files (caused by timeouts on a request of folder contents). We recommend that you move the shipped files by a scheduled task running on Sunday. Use Encryption The encryption feature is actually separate from Study Shipping. It applies to all studies saved or restored on the Ascribe™ web site. If no encryption password is provided, Ascribe™ will not encrypt the study files it creates, and will accept only unencrypted files for restoring. If an encryption password is provided, Ascribe™ will encrypt all study files it creates (using the provided password). It will accept either encrypted or unencrypted files for restoring. If you elect to use an encryption password, you will want to give some thought to how widely you disseminate knowledge of the password. Most users of Ascribe™ should have no need for the password. The password is needed only if the XML contents of the file need to be used. Changing the password is possible, but studies submitted to Ascribe™ for later restore operations will need to be either unencrypted, or encrypted with the new password. We recommend that you keep knowledge of the encryption password to a small group, and that you do not plan to change it. It is Language Logic policy not to provide the encryption password to anyone by phone. Please make a careful record of your encryption password. File Management 366 Data Management Shipped studies are placed in compressed files, and can be decompressed using a standard "zip" program such as WinZip™. The WinZip program is available at http:// www.winzip.com. If you wish to create an automated system for management of study files, you may want a programmable utility to work with zip files. At Language Logic we use the Xceed Zip compression library, available at http:// www.xceedzip.com. File names are constructed for shipped studies as follows: <Study ID>_<Study name><version>.zip The fields in this name are described in the following table: Field Description <Study ID> The Study ID, as assigned to the study in Ascribe™. The length of this field varies, and is a zero length string if no Study ID is assigned. <Study name> The Study name, as assigned to the study in Ascribe™. The length of this field varies, and is a zero length string if no Study name is assigned. <version> A zero length string, unless the file name as constructed above already exists in the FTP site. In that case a <version> string is constructed by creating a suffix as: _N where N is an integer that is incremented until the resulting file name does not exist on the FTP site. Repository Navigate: Supervisor/Repository The repository stores the copies of all of your studies and saves them for later retrieval. The repository is updated daily. At the time of the update, every study in your Ascribe™ account is saved to disk, just as if you had saved it manually. The saved study file contains the complete study, including all responses and codes applied. The saved study does not contain media files. If the study does not yet exist in the repository, it is added. If the study is already in the repository, the saved study file is then compared to the most recently saved version of the study in the repository. If the two versions are identical, the new copy is discarded. Otherwise, the new copy is stored in the repository. The repository retains all versions of a study that were inserted in the repository within the last 30 days. Only the most recently saved copy of a study older than 30 days is retained. Hence the repository contains daily copies of all studies that have changed within the past 30 days, but only the most recent copy of studies that have not changed within 30 days. For studies that are not changed, over time the repository will contain a single copy of each study (the one most recently saved). Ascribe™ does not delete the study from the repository. You can retrieve a study from the repository at any time. The file can be restored to Ascribe™ using the Restore Study page. 367 PrintDoc01 020813 Note The studies saved in the repository do not contain media files. Do not rely on the repository to retain media files for multi-media studies. If you need to retain copies of the media files in a study, you should save the study manually. Display Studies Navigate: Supervisor/Repository The Repository page displays the studies in the repository. The list of the studies is sorted with the most recently saved studies at the top of the list. Black lines group copies of the same study, with the most recently saved copy at the top of the group. Find Box Navigate: Supervisor/Repository Use the Find box to select the studies to display. Type the text you want to search for in the Find box, and then click the Update button. (If the Find box is blank, no studies will be displayed.) Ascribe™ will display studies that contain this text in the study ID or study name. The Find box also accepts a regular expression, so you can use standard regular expressions to search for your studies. To display all studies, enter a period (.) in the Find box. The period matches any character, so all studies will be displayed that have at least one character in either the ID or name. Show Details Navigate: Supervisor/Repository There are two views available for the repository. If the Show Details box is checked, the list contains detailed information about the studies in the repository. If this box is not checked, 368 Data Management the list shows only the status and identification information for the studies. To change the view, click the Show Details box and click the Update button. The study details screen has these fields: Field Description Row Number Click this column header to restore the sort order to the original view. Saved The date this copy of the study was saved in the repository. Status The status shows the study’s current state of progress. Study ID The study ID, which is important for automatically sending data to Ascribe™ with the FTP site. Study Name The name of the study. Description Description of the study. Created The date the study originally was added to Ascribe™. Due Date The due date of the study. Client The name of the client company. End Customer The name of the end customer. Supervisor The username of the supervisor for this study. Questions The number of questions in the study. Respondents The number of unique respondents. Responses The number of responses to all questions. Responses Coded The number of coded responses. Loads The number of loads. Hours The number of work hours spent on this study, taken from the time accounting records. 369 PrintDoc01 020813 All of your studies currently in Ascribe™ are saved and placed in the repository each weekend. If the study is already in the repository and has changed, the new copy of the study is also stored. This gives you a weekly "snapshot" of all of your studies. Download a Study from the Repository Navigate: Supervisor/Repository Right-click a study and select "Download file..." from the menu, or click the link in the "Saved" column. View the Questions in the Repository Navigate: Supervisor/Repository Right-click a study and select Questions from the menu. Study Archive (Save and Restore) XML Specification Ascribe™ uses XML to save and restore studies. This section describes the contents of the XML files. It will help you create software that will read and write Ascribe™ studies, so that you can integrate Ascribe™ with other software components in your company. Structure The XML schema used by Ascribe™ is "tag-centric", meaning that tags are generally used for data elements, as opposed to attributes. This results in XML files that are lengthy but highly compressible. Normally the XML document is stored in a zip file to reduce storage space and network transfer times. The use of a zip file is optional unless media files accompany the XML document. 370 Data Management This section does not contain a full example of an Ascribe™ XML file. You may easily obtain examples by saving one or more studies in Ascribe™ and inspecting the resulting XML file. Ascribe™ does check the syntax of the XML document when restoring a study, but does not perform extensive semantic checks. For example, it is possible to create codebooks with scrambled indentation levels. You must ensure that the documents you generate are semantically correct. XML Document Outline This section describes the general form of the XML document, listing those tags that serve as containers for other tags. Each document begins with the tag: <?XML version="1.0" encoding="utf-8" standalone="yes" ?> This indicates that the document uses XML version 1.0, and utf-8 encoding. You may also use utf-16 encoding. If you plan to support multiple languages, we recommend that you write your files in utf-16 encoding. The following tags serve as containers for other tags, and define the overall structure of the XML document: Tag Description <Studies> Container for <Study> tags <Study> Defines a study < CodeBooks> Container for < CodeBook> tags < CodeBook> Defines a codebook < CodeBookCode> <Questions> Defines a code in a codebook Container for <Question> tags <Question> Defines a question <Responses> <Response> < ResponseCodes> < ResponseCode> Container for <Response> tags Defines a response Container for < ResponseCode> tags Defines a code applied to a response < ResponseQualityCodes> Container for < ResponseQualityCode> tags < Defines a quality code applied to a ResponseQualityCode> response < QuestionRelationships> < QuestionRelationship> Container for < QuestionRelationship> tags Defines a relationship between two questions 371 PrintDoc01 020813 Tag Descriptions The table below lists each of the tags required for reading and writing Ascribe™ studies. For the Text data type, a maximum length in characters may be specified in parentheses. Textual data must of course conform to standard XML representations (e.g. & quot; for a double quotation character). Integer data types allow values for 32 bit signed integers. Remember that XML tag names are case sensitive. Tag Parent tag Data type Description Studies None Container tag Required. This is the root tag of the document. Study Studies Container tag Required. The Study tag defines a study. Exactly one Study tag must be present in the document. The schema is designed to allow multiple Study tags, but this usage is not presently supported by Ascribe™. StudyIDClient Study Text(20) Required. This is the requested short identifier for the study. This text appears in Ascribe™ as the Study ID. When an XML document is sent to Ascribe™ via the FTP site a study restore will be performed if a study with this ID does not exist. If a study with this ID does exist, the file sent will be merged with the existing study. When restoring or merging a study via the Ascribe™ web site, this tag is ignored except when the study is restored as Original. In other cases the user provides the desired Study ID when submitting the file. StudyName Study Text(50) Optional. The name of the study. StudyDescription Study Text Optional. A description of the study. StudyCreateDate Study DateTime Optional. The date and time that the study was created. 372 Data Management StudyHelp Study Text Optional. Help text to provide to coders of this study. StudyFileFormat Study Integer Optional. Specifies the file format for column binary output of the file. If present, must be one of the following values: 1: Binary 2: ASCII StudyFileEncoding Study Integer Optional. Specifies the internal encoding used in column binary output. Corresponds to the Layout setting in the Ascribe™ study. If present, must be one of the following values: 1: Punch 2: Numeric 3: Use question setting 4: Punch using question column offset StudyCardColumns Study Integer Optional. Specifies the number of columns used in column binary output. StudyCardNumberColumn Study Integer Optional. Specifies the column where the card number starts in column binary output. StudyCardNumberColumns Study Integer Optional. Specifies the number of columns occupied by the card number in column binary output. StudyRespondentIDColum n Study Integer Optional. Specifies the column where the respondent ID starts in column binary output. StudyRespondentIDColum ns Study Integer Optional. Specifies the number of columns occupied by the respondent ID in column binary output. CodeBooks Study Container tag Required. Container for codebooks for the enclosing study. CodeBook CodeBooks Container tag The CodeBook tag defines a codebook for the study. 373 PrintDoc01 020813 One CodeBook tag must be present for each distinct codebook in the study. Codebooks may be shared among questions, so the total number of codebooks may be less than the number of questions. CodeBookKey CodeBook Integer Required. The value contained by this tag is used within a Question tag to reference the codebook for the question. The value of the integer is arbitrary, but it must be unique among all codebooks in the XML document. Note that there is no requirement for the value to be unique across XML documents. The simplest scheme for generating CodeBookKey values is to use a counter starting at 1 and incrementing for each codebook. CodeBookDescription CodeBook Text(200) Optional. A description of the codebook. CodeBookNoDuplicates CodeBook Boolean Optional. Specifies the setting of the No Duplicate OutputID's value for the codebook. If present, the value must be one of the following: True: The No Duplicate OutputID's option is set False: The No Duplicate OutputID's option is not set If omitted, the default value is False. CodeBookIncrement 374 CodeBook Integer Optional. Specifies the value of the Increment setting for the codebook. This setting is used to generate suggested OutputID values when a new code is added to the codebook. Data Management CodeBookGuid CodeBook GUID Optional for study restore operations. Required for study merge operations. This GUID uniquely identifies the codebook. This is a critical value if you are creating XML files that you want to use for study merge operations. Understand that codebooks in Ascribe™ are not owned by specific studies. Codebooks can be shared across studies. This GUID provides a way for you to specify a codebook in a study on which you intend to perform later merge operations. The value of this GUID must be the same for a given codebook in each file you submit for a study merge. CodeBookCodes CodeBook Container tag Optional. Contains the codes in the codebook. While each question must have a codebook, the codebook need not contain any codes. CodeBookCode CodeBookCodes Container tag Optional. Defines a code in the codebook. 375 PrintDoc01 020813 CBCKey CodeBookCode Integer Required. The value contained by this tag is used within a ResponseCode tag to reference the code applied to the response. The value of the integer is arbitrary, but it must be unique among all CodeBookCode tags in the XML document (not just among codes in the containing codebook). Note that there is no requirement for the value to be unique across XML documents. The simplest scheme for generating CBCKey values is to use a counter starting at 1 and incrementing for each code. CBCOrdinal CodeBookCode Integer Required. The ordinal number of the code, which defines its sort order in the codebook. You should use the value 1 for the first code in the codebook, and increment this value for each successive code. Note that the order of codes in the codebook is determined by this value, not by the location of the CodeBookCode tag in the XML document. CBCDepth CodeBookCode Integer Required. The indentation depth of the code. This value should be 1 for a top level code, 2 for a net, 3 for a sub-net, and so on. 376 Data Management CBCChildren CodeBookCode Integer Required. The number of direct children of this code or net. The value should be 0 for codes, and non-zero for nets. Note that this is the number of direct children of the net, not the total number of descendents, i.e. children, but not grandchildren. It is very important that all nets have a non-zero value for CBCChildren, and that all leaf codes have a zero value. CBCOutputID CodeBookCode Text(24) Optional. The output value of the code. This is the value that should be written to a data file for this code when downloading coding results. CBCInputID CodeBookCode Text(24) Optional. The input value of this code. This is the value used for this code when uploading responses to Ascribe™. It is typically used only for closed end data. CBCDescription CodeBookCode Text(1000 Required. The description ) of the code. This is the text that appears for the code in the codebook. CBCRegExp CodeBookCode Text(800) Optional. The regular expression used for pattern matching during coding. CBCHover CodeBookCode Text(400) Optional. Help to display to coder when hovering over the code with the mouse. CBCHelp CodeBookCode Text(1000 Optional. Help to display to ) coder in popup window. 377 PrintDoc01 020813 CBCTextColor CodeBookCode Text(7) Optional. Foreground text color of the code to display to coders. This string is of the form: #RRGGBB where RR, GG, and BB are hex values for the red, green, and blue values to display. If not present, the default is #000000 (black). Questions Study Container tag Optional. Container for the questions in the study. Question Questions Container tag Defines a question. QuestionKey Question Integer Required. The value contained by this tag is used within <QuestionRelationshipFrom > and <QuestionRelationshipTo> tags to reference the question. The value of the integer is arbitrary, but it must be unique among all Question tags in the XML document. Note that there is no requirement for the value to be unique across XML documents. The simplest scheme for generating QuestionKey values is to use a counter starting at 1 and incrementing for each question. QuestionID Question Text(50) Optional. The short identifier for the question, typically the question number. QuestionText Question Text Required. The text of the question. QuestionHelp Question Text Optional. Help about this question to display to coders. QuestionCard Question Integer Optional. The card number of the question. Used for column binary output. 378 Data Management QuestionColumn Question Integer Optional. The starting column position for codes. Used for column binary output. QuestionColumns Question Integer Optional. The number of columns used for each code written in column binary output when using Numeric format ( StudyFileEncoding = 2). QuestionLabel Question Text(25) Optional. Short identifier for question, used to identify the question in reports. QuestionType Question Integer Optional. Specifies the type of the question. When present, must be one of the following values: 0: Open 1: Closed 2: Other specify 3: Value QuestionCodeBookKey Question Integer Required. Specifies the codebook for this question. Must match one of the values of a CodeBookKey tag in the document. More than one question may reference the same codebook. This indicates that the codebook is shared by these questions. QuestionFileEncoding Question Integer Optional. Specifies the internal encoding used in column binary output. Corresponds to the Layout setting in the Ascribe™ study. If present, must be one of the following values (note that the value 3 is illegal): 1: Punch 2: Numeric 4: Punch using column offset 379 PrintDoc01 020813 QuestionMaxCodes Question Integer Optional. Specifies the maximum number of codes to be output in result files for this question. If not present, or if the value is 0, the number of codes is unlimited. QuestionCodingSource Question Integer Optional. Specifies the text used in coding operations. This is the text that coders look at to code. Allowed values are: 0: Verbatim (default if not present) 1: Transcription 2: Translation QuestionOrdinal Question Integer Optional. Used for sorting questions when displayed to the user in a list. Questions in lists are sorted first by this value, then by the QuestionID. If not present the default value is zero. QuestionCode Question Boolean Optional. Specifies whether this question is to be coded. If false, this question will not be present in lists of questions to code presented to coders. If not present the default value is true. QuestionTranscribe Question Boolean Optional. Specifies whether this question is to be transcribed. If false, this question will not be present in lists of questions to be transcribed, nor will it be present in lists of responses to transcribe for the question. If not present the default value is false. 380 Data Management QuestionTranslate Question Boolean Optional. Specifies whether this question is to be translated. If false, this question will not be present in lists of questions to be translated, nor will it be present in lists of responses to translate for the question. If not present the default value is false. QuestionCrosstab Question Boolean Optional. Specifies whether this question should appear in the crosstab report. If not present the default value is false. Responses Question Container tag Optional. Container for responses to this question. Response Responses Container tag Optional. Defines a response. DROVerbatim Response Text(3000 Required. The verbatim ) text. DRORespondent Response Text(20) Required. The identifier for this respondent. Must be unique among responses for the containing Question. DROTranscription Response Text Optional. A transcription of the verbatim text. DROTranslation Response Text Optional. A translation of the verbatim or transcription text. DRONotes Response Text Optional. Notes for this response. DROMedia Response Integer Optional. Specifies the media type of the response. See Media types below. ResponseCodes Response Container tag Optional. Container for the codes applied to this response. ResponseCode ResponseCodes Container tag Optional. Defines a code applied to the containing Response. 381 PrintDoc01 020813 DCCBCKey ResponseCode Integer Required. Specifies the code applied. Must match one of the values of a CBCKey tag in the document. ResponseQualityCodes Response Container tag Optional. Container for the quality codes applied to this response. ResponseQualityCode ResponseQualityCode Container s tag Optional. Defines a quality code applied to the containing Response. DQCCBCKey ResponseQualityCode Integer Required. Specifies the quality code applied. Must match one of the values of a CBCKey tag in the document. The referenced code must be in the codebook for this question. QuestionRelationships Study Container tag Optional. Container for relationships between questions. QuestionRelationship QuestionRelationship s Container tag Optional. Defines a relationship between questions. QuestionRelationshipFrom QuestionRelationship Integer Required. References question A in the relationship "Question A is related to Question B". Must match one of the values of a QuestionKey tag in the document. QuestionRelationshipTo QuestionRelationship Integer Required. References question B in the relationship "Question A is related to Question B". Must match one of the values of a QuestionKey tag in the document. Data Types The following data types are used in the Ascribe™ study archive document: Data Type Description Container Tag This is not actually a data type, but rather a type of tag used in the document. A container tag contains other tags, and do not themselves contain values. 382 Data Management Integer A 32 bit signed integer value. The value in the document is a text representation of the integer in base 10. The range of this data type is -2147483648 to 2147483647, although some fields restrict the allowed value to certain ranges. Text(max characters) A field containing UNICODE characters, up to max characters in length. Zero length strings are permitted unless otherwise specified. Omitting the tag is equivalent to specifying a zero length string value. The value must conform to XML requirements for text values (i.e. XML escaping of reserved characters is required). Text A field containing UNICODE characters. Any number of characters may be specified. The value must conform to XML requirements for text values (i.e. XML escaping of reserved characters is required). Boolean The text True or False DateTime A datetime string conforming to the XML standard: YYYY-MM-DDTHH:MM:SS The length of the string must be exactly 19 characters. YYYY: year (e.g. 2004) MM: month (e.g. 02 = February) DD: day HH: hour (24 hour clock, 02 = 2AM, 18 = 6PM) MM: minute SS: second Example: 2004-05-14T03:09:45 Media Types The < DROMedia> tag specifies the media type of the response, for example image or sound. The allowed values are: DROMedia value Description 0 Numeric text - The contents of the DROVerbatim tag is the response, containing only the digits 0-9, hyphen (-), space, and vertical bar (|) characters. 1 (default) Text - The contents of the DROVerbatim tag is the response, containing any text characters. 2 Image - The contents of the DROVerbatim tag is a file path, relative to the Media folder in the zip file containing the XML document. When rendered in Ascribe™, the file will be displayed in an < IMG> tag. 3 Sound - The contents of the DROVerbatim tag is a file path, relative to the Media folder in the zip file containing the XML document. When rendered in Ascribe™, the file will played using Windows Media Player. When constructing XML documents, you need not make a distinction between DROMedia types 0 and 1. When either of these values is supplied, Ascribe™ inspects the contents of 383 PrintDoc01 020813 DROVerbatim and assigns a value of 0 or 1 based on the verbatim text. If your verbatims are all text, you can simply omit the DROMedia tag. When media files accompany the XML document, there are several additional requirements: The XML document must be placed in a zip file. The XML document must be at the root level of the zip file. The zip file must contain a folder named "Media", and that folder must contain the media files referenced in each response. The DROVerbatim value in the XML document is a file path name, relative to the Media folder in the zip file. The Media folder may contain sub-folders. In this case, the sub-folder name must be included in DROVerbatim. The media files provided must be suitable for direct display in the corresponding rendering format. For image files, the file must render properly in an < IMG> tag. For sound files, the files must be directly playable by Windows Media Player. Write XML Documents Creating XML documents containing Ascribe™ studies is fairly straightforward. You can test the results of your work by restoring test files to Ascribe™. You will find during test efforts that the error messages returned by Ascribe™ for improperly formed documents are not as informational as you could like. We recommend that you build your documents gradually during testing, starting with very simple documents as shown in the examples that follow, and working up to more complete documents. The file name you assign to your document files is of no consequence to Ascribe™. If you wish to zip your study XML documents, you should place only one XML document in each zip file. You may encrypt the zip file if you wish. You must use the encryption password assigned to your Ascribe™ account. Ascribe™ always writes study archive XML documents in utf-8 encoding. You may use utf-16 encoding for your documents if you find it more convenient. Parse XML Documents Reading the XML files created by Ascribe™ requires the use of an XML parser. Ascribe™ studies can create very lengthy XML files, causing DOM to require a large amount of memory. If you are working in the Microsoft framework, we highly recommend the use of the SAX parser, as opposed to the DOM object. In the .NET environment, use the XmlTextReader object directly to scan the file. Study files written by Ascribe™ may contain tags that are not described in this document. Your parser should ignore these tags. Also note that additional tags may be added in future releases of Ascribe™. Your parser should not be written to assume that it will not encounter unknown tags. Study files written from Ascribe™ are always zipped. If you wish to operate directly on the zipped file, you will need to incorporate an unzipping component in your application. Example of a Study with No Questions This first example shows the minimal XML document. It defines a study with no questions. <? xml version="1.0" encoding="utf-8" standalone= "yes" ?> <Studies> 384 Data Management <Study> < StudyIDClient>XML #1</ StudyIDClient> <StudyName>XML example</StudyName> <StudyDescription>This is an example for study loading from XML files</StudyDescription> <StudyCreateDate>2004-06-04T09:32:11</StudyCreateDate> <StudyHelp>This example is described in the " Ascribe™ study archive XML specification" document.</StudyHelp> <StudyFileFormat>1</StudyFileFormat> <StudyFileEncoding>1</StudyFileEncoding> <StudyCardColumns>80</StudyCardColumns> <StudyCardNumberColumn>6</StudyCardNumberColumn> <StudyCardNumberColumns>2</StudyCardNumberColumns> <StudyRespondentIDColumn>1</StudyRespondentIDColumn> <StudyRespondentIDColumns>5</StudyRespondentIDColumns> <CodeBooks /> <Questions /> <QuestionRelationships /> </Study> </Studies> Example of a Study with Three Questions and Coded Responses This example shows a partially coded study, and also shows the use of quality codes. The Questions Q1 and Q2 share a codebook. <?xml version= "1.0" encoding= "utf-8" standalone= "yes"?> <Studies> <Study> <StudyGuid>0DE31A2C-54D7-4B6F-8B05-B22DF9C3B6F9</StudyGuid> <StudyIDClient>XML #2</StudyIDClient> <StudyName>XML example</StudyName> <StudyDescription>This is an example for study loading from XML files</StudyDescription> <StudyCreateDate>2004-06-04T09:32:11</StudyCreateDate> <StudyHelp>This example is described in the "Ascribe™ study archive XML specification" document.</StudyHelp> <StudyFileFormat>1</StudyFileFormat> <StudyFileEncoding>1</StudyFileEncoding> <StudyCardColumns>80</StudyCardColumns> <StudyCardNumberColumn>6</StudyCardNumberColumn> 385 PrintDoc01 020813 <StudyCardNumberColumns>2</StudyCardNumberColumns> <StudyRespondentIDColumn>1</StudyRespondentIDColumn> <StudyRespondentIDColumns>5</StudyRespondentIDColumns> <StudyQuestionDateStyle>mdy</StudyQuestionDateStyle> <CodeBooks> <CodeBook> <CodeBookKey>529380</CodeBookKey> <CodeBookNoDuplicates>True</CodeBookNoDuplicates> <CodeBookIncrement>10</CodeBookIncrement> <CodeBookGuid>DDAD0237-9924-46A5-B4C6-62BA83B41D6D</CodeBookGuid> <CodeBookCodes> <CodeBookCode> <CBCKey>42137310</CBCKey> <CBCOrdinal>1</CBCOrdinal> <CBCDepth>1</CBCDepth> <CBCChildren>0</CBCChildren> <CBCOutputID>100</CBCOutputID> <CBCInputID>10</CBCInputID> <CBCDescription>Code1</CBCDescription> </CodeBookCode> <CodeBookCode> <CBCKey>42137311</CBCKey> <CBCOrdinal>2</CBCOrdinal> <CBCDepth>1</CBCDepth> <CBCChildren>2</CBCChildren> <CBCOutputID>20</CBCOutputID> <CBCInputID>20</CBCInputID> <CBCDescription>Net1</CBCDescription> </CodeBookCode> <CodeBookCode> <CBCKey>42137313</CBCKey> <CBCOrdinal>3</CBCOrdinal> <CBCDepth>2</CBCDepth> <CBCChildren>0</CBCChildren> <CBCOutputID>110</CBCOutputID> <CBCInputID>40</CBCInputID> <CBCDescription>SubCode1</CBCDescription> 386 Data Management </CodeBookCode> <CodeBookCode> <CBCKey>42137314</CBCKey> <CBCOrdinal>4</CBCOrdinal> <CBCDepth>2</CBCDepth> <CBCChildren>0</CBCChildren> <CBCOutputID>120</CBCOutputID> <CBCInputID>50</CBCInputID> <CBCDescription>SubCode2</CBCDescription> </CodeBookCode> <CodeBookCode> <CBCKey>42137312</CBCKey> <CBCOrdinal>5</CBCOrdinal> <CBCDepth>1</CBCDepth> <CBCChildren>0</CBCChildren> <CBCOutputID>130</CBCOutputID> <CBCInputID>30</CBCInputID> <CBCDescription>Code2</CBCDescription> </CodeBookCode> </CodeBookCodes> </CodeBook> <CodeBook> <CodeBookKey>529382</CodeBookKey> <CodeBookNoDuplicates>True</CodeBookNoDuplicates> <CodeBookIncrement>10</CodeBookIncrement> <CodeBookGuid>FC563F0F-F5F9-48E9-9D6E-7D4577E68CC6</CodeBookGuid> <CodeBookCodes> <CodeBookCode> <CBCKey>42137315</CBCKey> <CBCOrdinal>1</CBCOrdinal> <CBCDepth>1</CBCDepth> <CBCChildren>0</CBCChildren> <CBCOutputID>10</CBCOutputID> <CBCInputID>10</CBCInputID> <CBCDescription>Code1</CBCDescription> </CodeBookCode> <CodeBookCode> 387 PrintDoc01 020813 <CBCKey>42137316</CBCKey> <CBCOrdinal>2</CBCOrdinal> <CBCDepth>1</CBCDepth> <CBCChildren>0</CBCChildren> <CBCOutputID>20</CBCOutputID> <CBCInputID>20</CBCInputID> <CBCDescription>Code2</CBCDescription> </CodeBookCode> </CodeBookCodes> </CodeBook> </CodeBooks> <Questions> <Question> <QuestionKey>730443</QuestionKey> <QuestionID>Q1</QuestionID> <QuestionText>This is the question text</QuestionText> <QuestionHelp>This is coder help for the question</QuestionHelp> <QuestionCard>0</QuestionCard> <QuestionColumn>20</QuestionColumn> <QuestionColumns>5</QuestionColumns> <QuestionLabel>Question1</QuestionLabel> <QuestionType>0</QuestionType> <QuestionCodeBookKey>529380</QuestionCodeBookKey> <QuestionFileEncoding>2</QuestionFileEncoding> <QuestionMaxCodes>0</QuestionMaxCodes> <Responses> <Response> <DROVerbatim>Response 1</DROVerbatim> <DRORespondent>100</DRORespondent> <ResponseCodes> <ResponseCode> <DCCBCKey>42137310</DCCBCKey> </ResponseCode> </ResponseCodes> </Response> <Response> <DROVerbatim>Response 2</DROVerbatim> 388 Data Management <DRORespondent>101</DRORespondent> <ResponseCodes> <ResponseCode> <DCCBCKey>42137314</DCCBCKey> </ResponseCode> <ResponseCode> <DCCBCKey>42137312</DCCBCKey> </ResponseCode> </ResponseCodes> <ResponseQualityCodes> <ResponseQualityCode> <DQCCBCKey>42137312</DQCCBCKey> </ResponseQualityCode> <ResponseQualityCode> <DQCCBCKey>42137314</DQCCBCKey> </ResponseQualityCode> </ResponseQualityCodes> </Response> <Response> <DROVerbatim>Response 3</DROVerbatim> <DRORespondent>102</DRORespondent> <ResponseCodes> <ResponseCode> <DCCBCKey>42137310</DCCBCKey> </ResponseCode> </ResponseCodes> <ResponseQualityCodes> <ResponseQualityCode> <DQCCBCKey>42137310</DQCCBCKey> </ResponseQualityCode> </ResponseQualityCodes> </Response> </Responses> </Question> <Question> <QuestionKey>730444</QuestionKey> <QuestionID>Q2</QuestionID> 389 PrintDoc01 020813 <QuestionText>This is the text for question 2</QuestionText> <QuestionHelp>This is help for question 2</QuestionHelp> <QuestionCard>0</QuestionCard> <QuestionColumn>20</QuestionColumn> <QuestionColumns>5</QuestionColumns> <QuestionLabel>Question2</QuestionLabel> <QuestionType>0</QuestionType> <QuestionCodeBookKey>529380</QuestionCodeBookKey> <QuestionFileEncoding>2</QuestionFileEncoding> <QuestionMaxCodes>0</QuestionMaxCodes> <Responses /> </Question> <Question> <QuestionKey>730445</QuestionKey> <QuestionID>Q3</QuestionID> <QuestionText>Text for question 3</QuestionText> <QuestionCard>0</QuestionCard> <QuestionColumn>20</QuestionColumn> <QuestionColumns>5</QuestionColumns> <QuestionLabel>Question3</QuestionLabel> <QuestionType>0</QuestionType> <QuestionCodeBookKey>529382</QuestionCodeBookKey> <QuestionFileEncoding>2</QuestionFileEncoding> <QuestionMaxCodes>0</QuestionMaxCodes> <Responses /> </Question> </Questions> <QuestionRelationships> <QuestionRelationship> <QuestionRelationshipFrom>730444</QuestionRelationshipFrom> <QuestionRelationshipTo>730443</QuestionRelationshipTo> </QuestionRelationship> </QuestionRelationships> </Study> </Studies> 390 Accelerator Coding Model The Accelerator Coding Model (ACM) can code open-end responses in far less time than human coders. The fundamental concept of the ACM is that it learns how to code by looking at examples of properly coded responses. It uses these examples to learn how to code new responses. To use the ACM, you first train it about how to code using a specific codebook. Once the ACM has been trained, you can use it to code new responses for questions using the trained codebook. The ACM provides several significant advantages over alternative machine coding technologies: It required no special expertise. You train the ACM by properly coding a set of training examples, just as you would ordinarily code them. It codes quickly. While the coding speed depends on many factors, the ACM can typically code thousands of responses per hour. It supports multiple languages, even intermixed in a single question. It is integrated into Ascribe™, so you can use all of the tools available in Ascribe™ for reviewing and output of the coding results. ACM Definitions Here are some definitions for commonly-used terms for automated coding models: Term Definition ACM ACM is another term for accelerator coding model. A coding model is a group of codes that have been trained to code open-ended verbatim. A coding model is made up of codes, training examples, and classifications patterns (if the model has been trained.) To Train "I need to train my coding model." Each code in the model is given "training examples," which are analyzed and used to create patterns, rules, and classifiers for that code. Training Examples Training examples are phrases, words, or responses that represent a code in the model codebook. Classification Patterns Classification patterns are rules that are created when training examples are examined during the training process. These rules determine whether a code will be suggested or note when using the model to code. ACM Theory of Operation The fundamental concept behind the ACM is that it you train a codebook using any number of properly coded examples from any number of questions. Once you have a trained codebook or model, you can use it to code any question with a similar codebook. Let's consider this in more detail. Note that we train the ACM on a codebook, not on a question. We are training the ACM about how to properly apply codes from that codebook. When trained, the ACM can be used for coding any question that shares that codebook. As 391 PrintDoc01 020813 we will see, it can also be used to code questions that do not share the codebook, but use a similar codebook. A given codebook can have no more than one trained model. Either the codebook has a model or it does not, but it cannot have multiple models. To train the codebook, we need a set of properly coded example responses (also referred to as training examples.) This is the heart of the work of creating a model. Armed with the codebook to train, and the set of training examples, we can train the ACM on how to code using that codebook. Note here that we must train the ACM with properly coded example responses. This cannot be over-emphasized. The VCS™, which is the engine at the core of the ACM, uses the training examples to construct rules so that it can code responses not encountered in the training examples. Clearly, using improperly coded examples will cause the VCS™ to attempt to mimic that improper coding! This we surely want to avoid. It is preferable to have fewer well coded examples than more examples coded with lower quality. That said, the accuracy of coding will increase with the number of properly coded examples. The ACM supports multiple languages. You can create an ACM training that will code in any number of the supported languages. The training examples are stored in a Language in the model. Each model must have at least one Language, but may have more. Training examples can come from any number of sources, such as questions, manual entry, and load files. After you have set up a model, it is submitted to the VCS™ to be trained. This is called training the model. When the training job completes successfully, we say that the model has been trained. After successful completion of the training job, you may mark the model as approved. Only approved models can be used for coding. To assist you in deciding whether to approve a model, the VCS™ can generate metrics that estimate the quality of the training results. You can review these and the training examples, before approving the training. Once you have an approved model, you may use it to code questions. To do so, you need to associate the question to be coded with the model to be used. We call this a coding question of the model. A given question can be a coding question of no more than one model. Like a codebook, a question either has a model used to code it or it does not, but a question cannot have more than one model used to code it. Once you have associated a question with an approved model, you can code that question in Ascribe™ with a single menu selection. Code Matching It is possible to code questions that do not share the model's codebook. In this case, you must specify some rule to use for matching codes between codebooks. The ACM provides several options for how to do this. For example, you may use the rule that the description of the code should be used for matching. In this case, two codes in different codebooks are considered to be the same code if their descriptions match. Other matching rules are also available. This provides additional flexibility, but of course requires a certain discipline in codebook management. Users must be aware that seemingly inconsequential changes to a codebook may cause the matching to fail, which will decrease the accuracy of coding. ACM Architecture 392 Ascribe™ Desktop The setup information for model is contained in the Ascribe™ database. When you run a training job for the ACM, this information is passed to a dedicated server running the VCS™ software. The VCS™, which is the analytic engine for ACM, constructs a mathematical model using the training examples you supply. This model remains on the VCS™ server for later use in coding. The VCS™ server returns to Ascribe™ information about the training results. You can review this information in Ascribe™, to determine whether the model is ready to be approved for coding. At coding time, Ascribe™ sends to the VCS™ server the set of responses to be coded, and specifies which model to use. The VCS™ uses the model constructed at training time to assign a figure of merit, called the coding confidence, to each code and for each response. This information is returned to Ascribe™ by the VCS™ server. In Ascribe™, you can review and optionally accept the codes suggested by the VCS™ in the ACM Coding Review window of the Desktop application. Alternatively, you can simply instruct the ACM to apply those codes with a specified minimum confidence value to the responses. VCS™ - Verbatim Coding System The Verbatim Coding System (VCS™) is the analytic engine of the ACM. Unlike other automated coding systems, VCS™ requires no special knowledge of linguistic analysis. VCS™ was developed at the Institute of Information Science and Technologies (ISTI) in Pisa, Italy. Dr. Fabrizio Sebastiani led the research team that created VCS™. Language Logic has licensed this technology for use with Ascribe™. Dr. Sebastiani has written several papers on the VCS™ technology. Some of them are available here: Machines that learn how to code open-ended survey data Cracking the Code: What Customers Say in Their Own Words How the Verbatim Coding System Works The machinery inside of the Verbatim Coding System (VCS™) consists of binary classifiers. A binary classifier makes decisions about a single specific code. Given a textual response (verbatim), the binary classifier says either "Yes, apply the code", or "No, do not apply the code". This is why it is called binary. It always gives one of two answers: yes or no. When you train a codebook, you create binary classifiers for each code trained, one such classifier per code. Binary classifiers are completely independent. There is no direct dependency or interaction one to the next, save that all classifiers for a given question were created from the same set of training examples. Because the principle of operation of each binary classifier is the same, we will discuss only a single classifier. You understand that in practice this same story is repeated for each code in the codeframe. Creation of Binary Classifiers A binary classifier for a given code is created during the VCS™ training process. Each training example is considered in turn. A given training example either has or does not have 393 PrintDoc01 020813 the code applied. In the former case we call the verbatim a positive training example, in the latter we call it a negative training example. In general there are of course far more negative than positive training examples for any given code. As each training example is considered, VCS™ breaks it down into patterns. There are many types of patterns, each type essentially a way of looking at the example. Some patterns are based on the internal structure of the words (which are in turn based on the morphology or patterns of word formation of the specific language). Examples of such patterns may include words, syllables, and prefixes or suffixes such as word stems. Other patterns are based on the overall structure of the example, and are detected using a generative grammar of the language. Generative grammar means a set of rules that will correctly predict which combinations of words will form grammatical sentences. Several dimensions of language are taken into account: morphological, lexical, syntactic, semantic, pragmatic, each giving rise to different types of patterns. It is also instructive to understand what VCS™ does not use in its analysis. VCS™ does not use dictionaries, thesauri, or any other kind of domain-dependent resource, for the target language. VCS™ does not attempt to deeply "understand" the example as an artificial intelligence system might do. Instead VCS™ tries to understand which patterns are good predictors of whether a given example should be classified with a given code. VCS™ generates a very large number of patterns from each of the training examples it is presented. For each pattern, VCS™ records whether the code for the binary classifier is applied to the example. As you might expect, many patterns appear infrequently or only once in the training examples. Other patterns, however, are encountered in multiple training responses. Patterns that appear with low frequency are relatively poor predictors of whether the code should be applied to the example. Patterns that appear often may be better predictors, depending on whether the code is applied consistently in the training examples for this pattern. If, for example, a given pattern appears 50 times in the training examples, and for each such example the code for the binary classifier is applied, then we can say that this pattern is a very good positive predictor that the code should be applied. Similarly, if the code is never applied for this pattern in the training examples, then the pattern is a very good negative predictor that the code should not be applied. In practice, a given pattern that appears multiple times will be found in some training examples with the code applied and in some examples with the code not applied. VCS™ assigns to each pattern a figure of merit specifying its predictive ability as to whether to apply or not apply the code. At the end of the training process, VCS™ has compiled a list of patterns, each of which has a related figure of merit specifying the predictive ability of the pattern. Positive values indicate that the pattern is a positive predictor; negative values indicate that the pattern is a negative predictor. The magnitude of the value is the weight of the prediction. Coding Using Binary Classifiers After the training process, VCS™ may be used to suggest coding for previously unseen verbatim responses. For each such verbatim response, VCS™ determines the patterns in the 394 Ascribe™ Desktop response, just as during the training process. The patterns are compared against the list of patterns generated during the training process. When matches are found, VCS™ combines the predictive strength (positive or negative) of the matching patterns, according to a sophisticated combination rule. The resulting number is the score assigned to the verbatim for the code. If the score is positive, it indicates that the code should be applied to the verbatim response. If the score is negative, it indicates that the code should not be applied to the verbatim response. The magnitude of the score, called the confidence, indicates the strength of this decision. For example, large positive values indicate that the code should be applied, and that VCS™ is highly confident in this. Small positive values indicate that the code should be applied, but that VCS™ is not terribly confident in this. Large negative values indicate that the code should not be applied, and that VCS™ is highly confident in this. Lessons Learned From this internal knowledge of VCS™, we can infer some important rules about how to properly apply it to our work. Sufficient training examples are required for each code. VCS™ creates a binary classifier for each code based solely on the training examples provided. In general, increasing the number of properly coded responses will increase the accuracy of the classifiers. Note that, since the classifiers for all of the codes in the codeframe are generated from the same training examples, the classifiers for the more frequently occurring codes will typically be more accurate than those for the less frequently occurring codes. Positive examples must be provided if the code is to ever be applied. Training a model with no positive examples for a given code will clearly cause VCS™ never to apply the code. No patterns generated during the training process will have a positive predictive weight. Negative examples are significant. VCS™ learns the application of a code both from positive examples and from negative examples. Clearly errors of omission of the application of a code during the training process may cause that code not to be applied during coding. Hence, it is not sufficient for the training examples to merely have each code assigned be correct. They must also be complete in the sense that all codes that should be applied are applied. Multiple identical training examples are significant. If the same verbatim response is used multiple times during training, it reinforces the predictive strength of the generated patterns, provided it is coded consistently for each example! On the other hand, multiple identical verbatim responses that are coded inconsistently will tend to nullify the predictive value of the generated patterns. Predicting the action of VCS™ is not entirely possible. As we have seen, the operation of VCS™ is complex. Why VCS™ has coded a verbatim a given way depends both on its internal machinery and on the training verbatims it has been trained with. The only sure way to determine how VCS™ will code a verbatim response is to actually try it! On the other hand, experience with VCS™ will help you gain an appreciation of the way VCS™ will tend to code based on the training examples provided. VCS™ is largely language-independent. VCS™ can be customized for new target languages by rewriting some modules only, namely, those for the generation of 395 PrintDoc01 020813 (morphological, lexical, etc.) patterns. The remainder of VCS™ is languageindependent. This allows VCS™ to be customized for additional languages rapidly as compared to systems based on other linguistic analysis technologies. Frequently Asked Questions About VCS™ Is the same verbatim always coded the same way? Yes, as it may be expected, two identical copies of the same verbatim are attributed exactly the same codes. Unless, of course, the classifiers have been retrained after validation, or with a different (e.g., augmented) set of training examples, in which case a verbatim that had been coded one way before retraining could be coded another way after retraining. If two inconsistently coded copies of the same verbatim are provided for the training process, does the system fall apart? Luckily not. VCS™ implements a "fail soft" policy, such that the system will simply behave as it had been provided with neither training example: i.e., contradictory pieces of training information neutralize each other. If an uncoded verbatim is the same as a training example, does it receive the same codes during the coding phase? Not necessarily. The training example can have a set of codes, and the uncoded verbatim may be automatically attributed a different set of codes, since the coding behavior of the system is determined by all the training it has received, and not by a single training example. Were it not like this, of course, the system could not "fail soft" as discussed in the previous point. Does VCS™ always attribute at least one code to each verbatim? No, VCS™ may even attribute no code at all to a given verbatim, since each code in the codeframe is treated in isolation of the others. This is akin to giving the verbatim the special code "Others", which is indeed what VCS™ did in a previous implementation. How much time and money am I going to save on a project by using VCS™? The answer depends very much on how many training examples are provided, how many verbatims are then coded automatically, how long is the average verbatim, and many other variables. A study on practical benefits found in operation was run by the UK-based Egg banking group, who had acquired an older implementation of VCS™; the results of this study can be found at p. 15 of a paper published at the 50th Annual Conference of the Market Research Society. Who developed VCS™? VCS™ was invented and developed by Fabrizio Sebastiani and his research team at ISTICNR, Italy. ISTI-CNR is the Institute for the Science and Technologies of Information of the Italian National Council of Research. Fabrizio and his team have been working on technologies for large-scale, robust classification of text for the last ten years, and have developed VCS™ as a result of their findings in the areas of machine learning, automated coding of verbatim text, automatic opinion analysis, and survey computing. Archimede Informatica, another Italian company, was also involved in engineering VCS™ and developing the communication layer with Ascribe™. VCS™ Training Metrics The ACM reports certain metrics at training time. These metrics are actually generated by the VCS™, the analytic engine of the ACM. An understanding of these can help you determine whether a given training is suitable for use in coding. 396 Ascribe™ Desktop The VCS™ provides metrics to assess the accuracy of the training results at the individual level (F1), and at the aggregate level ( PD). This help page provides a brief overview of the metrics. For a more detailed discussion, see Machines that Learn how to Code Open-Ended Survey Data, Part II: Experiments on Real Respondent Data. Accuracy at the Individual Level (F1) Suppose we want to compare the effectiveness of two coders. We would like to have a number that ranges from 0 to 100, with 0 meaning the two coders never agree, and 100 meaning that the two coders are in perfect agreement. The F1 accuracy metric does this. Considering a given code, let us imagine that we have a question that we code twice, once by a human coder (the "coder"), and then by the VCS™. Considering a given code, we have four possible outcomes: True positive: The coder and the VCS™ both apply the code False positive: The VCS™ applies the code, but the coder does not False negative: The VCS™ does not apply the code, but the coder does True negative: Neither the VCS™ nor the coder apply the code We can put these outcomes in a contingency table like this: Coder says yes VCS™ Says yes true positive no no false positive false negativetrue negative We sum the occurrences of these outcomes for all to obtain TP, FP, FN, and TN. TP is the number of false positives, FP the number of false positives, FN the number of false negatives, and TN the number of true negatives. We can then define Precision = P = TP / ( TP + FP) Recall = R = TP / ( TP + FN) Precision is a measure of overcoding. Its value decreases from 1 as the VCS™ applies codes when it should not (false positives). Recall is a measure of undercoding. Its value decreases from 1 as the VCS™ does not apply codes when it should (false negatives). We combine these two values using their harmonic mean to obtain our metric, F1: F1 = 2 P R / (P + R) There are several advantages to the F1 metric: The value of F1 ranges from 0 to 1, with the value 1 meaning perfect agreement. 397 PrintDoc01 020813 It partially rewards partial success. If codes A, B, C, and D should be applied to a response, applying A, B, and C results in a better F1 score than applying A alone. It rewards systems that balance precision and recall. It is symmetric. Applying the measurement between the VCS™ and the coder results in the same number as between the coder and the VCS™. When the F1 values are reported by the VCS™ they are multiplied by 100. Hence the F1 values reported range between 0 and 100, rather than between 0 and 1. We can also define a second metric, the error rate ER, as: ER = ( FP + FN) / ( FP + FN + TP + TN) This metric is the number of incorrect coding decisions divided by the total number of decisions. While this metric is conceptually simple to understand, it is not as reliable as F1. The VCS™ reports this metric based on its sole advantage of being conceptually simple. You can see that if the code is suggested by the VCS™ very infrequently (even if always incorrectly!) the error rate will be very low simply because the total number of decisions is large. For F1, a greater value means higher predicted quality. For ER (and PD), a lower value means higher predicted quality. Estimating Accuracy How does the VCS™ arrive at an estimate of its accuracy during the training process? It does so by conducting a set of experiments. Suppose we give the VCS™ 1000 example responses. The VCS™ takes some large percentage of the responses (say 900) and trains itself using those responses. It then codes the remaining 100 responses, and compares itself against the actual coding of the responses by computing the F1 value. It conducts this experiment several times, dividing the example responses in different ways. It compiles the results of these experiments to produce the overall F1 value for each code. Because the computation of expected accuracy causes the VCS™ to actually perform multiple training and coding operations, the time required to train the VCS™ with accuracy measurements is several times longer than when accuracy is not computed. Accuracy at the Aggregate Level ( PD) Accuracy at the aggregate level is estimated using the Percentile Discrepancy (PD). PD is defined as the absolute value of the difference between the true percentage and the predicted percentage of verbatims with code C. In other words, suppose that 42% of the verbatims should have code C, and suppose that VCS™ applies instead code C to 40.5%, or to 43.5%, of the verbatims; we say that VCS™ has obtained 1.5% PD on code C. A hypothetical ideal binary coder always obtains PD = 0. Multi-Lingual Models The ACM is capable of coding in multiple languages. It can even code multiple languages interspersed in a single question. 398 Ascribe™ Desktop If you need only to target a language other than English for a given model, the setup is simple. You need merely to specify that language as the single language in the model, and of course ensure that all of the training examples are in that language! You may then use that model to code questions that contain only responses in that language. If you need to target multiple languages in a single model, the setup is a bit more complex. Fundamentally, you need to tell the ACM how to determine the language of a given response. The ACM makes these assumptions: The language of a response can be determined by that respondent's response to another question The question ID which holds the language response must be the same for all studies being coded by a specific model The response that identifies the language must be the same for all studies being coded by a specific model. When constructing a model, you may specify a Language Question ID. A question with this ID must be found in all studies that have a coding question associated with the model. For each language in the model, you specify the response to the language question that identifies that language. Of course, this response must be different for each language in the model. You can see that these assumptions require a certain discipline at study setup time. If your company uses a rule such as "the question that identifies the language used by a respondent is always called LANGUAGEID", then you are off to a good start. If, in addition, you have rules such as "The (closed-end) response for English is always 4, French is always 2, and German is always 5," then everything will work beautifully. If your rules are not so rigid, then more work will be required to prepare a study for multi-lingual coding with ACM. When to Use ACM There is no single best approach to this. Factors that will affect your planning include the type of study you plan to use the model to code, your company's conventions for how you manage your projects, the complexity of the codebook being trained, and your procedures for quality assurance. Trackers Tracker studies are a natural application for the ACM, because coding work done on previous waves can be used as training examples for the next wave. In the simplest case, you would simply create a model for each codebook for open-ended questions in the tracker. For a given model, you would add the coded responses from the completed waves as training examples for that codebook. You would code the next wave with the model. While simple in concept, you should take some care to ensure that the training examples you provide from completed waves are indeed of sufficient quality to act as training examples. You will also want to consider whether you share codebooks across waves of the tracker. Use of the ACM is simplified if you do share codebooks across waves. If you find this impractical for your own work, you should consider carefully how you intend to have the ACM match codebooks across waves. The Match codes by property of the model governs this behavior. You will need to adhere strictly to some rule to match codes from one codebook to another, such as using the output code value (output ID) of the code. 399 PrintDoc01 020813 Large One-Time Studies Using the ACM can increase productivity on large studies even if the codebooks are completely custom. If the number of respondents is sufficiently high, it can be more productive to train the model than to manually code the entire job. When considering the use of the ACM for such work, you may also elect to use the ACM for targeted portions of the codebooks. This may be particularly effective if you have certain commonly applied "single-use" codes, in other words codes that cannot be applied with other codes. For example, you may have codes for "Nothing" and "Don't know" with a rule that these codes, when applied, must be the only code applied to the response. In this case, you may wish to train the ACM for those codes alone, and take a first pass at coding the study for those codes only. Remembering that the ACM will never apply a code unless it has at least one training example, you can construct a model that will code only the "Nothing" and "Don't know" codes to accomplish this task. Codes Common to Multiple Codebooks Building on the previous scenario for large one-time studies, you may get significant productivity boosts by using the ACM to do first-pass coding of certain codes that are used in multiple codebooks. Most organizations have rules for such codes as "Nothing" and "Don't know". If your company has such rules, the ACM can provide a significant productivity increase by coding these responses before manual coding begins. In this scenario, the model will likely have very few codes and contain only those that are common to multiple codebooks. The trained codebook will not be the same as the codebook used by the question being coded by the ACM, so you must enforce a rule such as "the Don't Know code always has an OutputID of 999" to allow the codebook of the model to be matched to that of the question being coded. Questions Common to Multiple Studies Your company may have commonly asked questions that appear in otherwise unrelated studies. These can be good candidates for ACM coding, subject to certain considerations. The primary consideration is in which it is asked. In other from any study in which this likely have a good candidate whether the question is conceptually independent of the study words, you should ask yourself whether training examples question appears would be suitable for other studies. If so, you for ACM coding. In this scenario, you might choose to create a study for all such questions. This study becomes the repository of the training examples for commonly asked questions. Multi-Lingual Projects The costs of coding multi-lingual studies make this an exciting area for the use of the ACM. Working with multi-lingual projects is simple with the ACM, provided you adhere to certain rules for setting up your questionnaires. See Multi-Lingual Models for more information. ACM Fees Fees for coding using the ACM are assessed in transactions. A response coded using the ACM costs 3 transactions. See How ACM Fees are Applied for detailed information. 400 Ascribe™ Desktop The transaction reports available to users with administrator privilege detail the number of ACM transactions used. How ACM Fees are Applied When a response is coded using ACM, it uses 3 Ascribe™ transactions. The transaction fee occurs at the time the code is actually applied to the response. The fee will be assessed when the coding job completes, and only for those responses to which the ACM applies a code. Once an ACM fee has been applied to a given response, it will never be applied again. If you were to code a question using the ACM, delete all of the codes applied by the ACM, and code it again identically using the ACM, no fees would be applied for the second coding job. Note that the transaction fees for ACM coding are not sensitive to whether the response is numeric. If a response is coded with the ACM, the transaction fee is always the same. After an ACM code has been applied to a response, the transaction fee for that response is set at 3 and will never change. If you should uncode it, recode it manually, uncode it again, and recode it with ACM, the fee does not change. It remains a 3. An interesting case arises when you code a response manually (let's say a non-numeric response that costs one transaction), and then code it later with the ACM. How is this handled? When the response is coded with the ACM in this case, a net charge of 2 transactions should be applied. This is the 3 transactions for the ACM, minus the one transaction that was already charged. In this case Ascribe™ "refunds" the original one transaction, and then charges 3 ACM transactions. This is done to provide an accurate picture of the number of ACM transactions your company is using. In the transaction reports available in Ascribe™, the column for ACM transactions will always show 3 transactions for each response coded with the ACM, regardless of whether the response had been previously coded manually. It is therefore possible to see negative values for transactions charged for text and numeric responses. Here is an example: Suppose question has 100 text responses and 100 numeric responses. You code them all in June. A transaction report for June will show 100 transactions for text responses, and 10 transactions (100 * .1) for numeric responses. Now let's suppose that in July we delete all of the codes and code the same question with the ACM, such that all responses are coded. The transaction report for June will not change. The transaction report for July will show 600 ACM transactions (200 responses *3 transactions per response). In the Text transactions column we will find -100 transactions. These are the transactions "refunded" because they had already been charged for manual coding. In the Numeric responses column we will find -10, the refund for the originally coded numeric responses. ACM Transaction Reporting Several reports display the number of ACM transactions used. When reviewing these reports, it is important to understand how ACM fees are applied. For totals of ACM transactions used by month, see the Transactions by Month report. For transaction usage within a time period, see the Transactions by Period report. For transaction usage for a specific study, see the Transactions by Study report. 401 PrintDoc01 020813 ACM Process on the Web Navigate: Accelerator (ACM)/Automated Coding Models The ACM process has the following steps: create and train a model, use the model for coding, improve the model, and maintain the model. Create and Train a Model Building a model requires supervisor privilege. Here are the steps for creating a model: Prepare your question/codebook: Use Ascribe™ coding tools to create a codebook and code a sample of the data. Mark responses that will become training examples as "Quality Reviewed." Make sure each code has proper representation (enough responses coded for each code.) If the codebook has positive and negative nets, be sure that the codes are equally represented by examples/coded responses. Create the model from the Questions page - Right-click the question that contains the codebook you would like to use create a model and choose Accelerator (ACM)/Create an ACM. Enter information in the Create ACM dialog box - After you enter the information, click the OK button. Train the model (from the Accelerator Coding Models page or the Accelerator Coding Model page), right-click and select Train - A training job is submitted, and you can view it on the Jobs page. You can also view the job status on the Accelerator Coding Models page in the Job Status column. During the training process, the training examples for each code are analyzed to create patterns, rules, and classifiers for that code. View training results - From the Accelerator Coding Models page or the Accelerator Coding Model page, right-click and select Training Results. The Training Results page displays the trend and accuracy of the model. Examine this report to see which codes may need more or different training examples to receive better accuracy or trend scores. Approve the model - Once you are satisfied with the training results, you need to approve the model. After a model is approved, it can be used to code a question. From the Accelerator Coding Models page or the Accelerator Coding Model page, right-click and select Approve. Use the Model for Coding Share the codebook of the coding question with codebook used by the model - The codebook of the coding question must be shared with the codebook used by the model. If you need to share the codebooks, use the Copy/Share Manager to share codebooks. Add a coding question or questions to the model - Navigate to the Accelerator Coding Model page (right-click menu option Edit ACM.) To add one question at a time, right-click an entry in the Coding Questions table and select Add/Coding question. To add all questions which share the ACM codebook, right-click an entry in the Coding Questions table and select Add/All questions sharing the ACM codebook. 402 Ascribe™ Desktop Start the ACM coding job - Navigate to the Accelerator Coding Model page. Right-click a coding question and select Start ACM coding job. Improve a Model Quality check responses coded by the model - Use Review Responses to review the ACM coding. Import quality-checked responses as new training examples - The model must be unapproved before you can add or change the training examples. Use the right-click menu item in Review Responses to add training examples. Re-train the model, approve it, and use it for coding - Continue with the process until you are satisfied with the coding. Maintain the Model Use Review Examples to change the coding for training examples. Use the Training Example Editor to create your own training examples. Use the Tester Tab in the Training Example Editor to check how a response would be coded. Create a Model from a Question Navigate: Supervisor/Studies/Right-click a study/Select Questions/Right-click a question/Select Accelerator (ACM)/Select Create an ACM When you select a question and choose to create a model from the question, a dialog opens. Enter the following information for the model: Field Description ID The unique identifier for the model. No two models can have the same identifier. The ID must be 30 characters or less. The ID defaults to the studyID_questionID. Description An optional description of the model. Copy Options Gives the option to import training examples: Do not copy responses - No training examples are imported. Copy all responses as examples - All responses from the question are imported. Select a sample of responses to copy - Enter a number or a percentage of the responses to be imported via random selection. Quality Reviewed responses only - Only responses that have been marked as Quality Reviewed are imported. Leave the check box blank if you want to import responses regardless of their Quality Reviewed status. Coded only - Only coded responses are imported. Leave the check box blank if you want to import both 403 PrintDoc01 020813 coded and uncoded responses. The question is automatically added as a coding question for the model. The model is created with these default properties: Property Description Match Codes By None All coding questions must share the model's codebook. To change this field, navigate to the Edit ACM page and use the right-click menu to select Properties. Language QuestionID Is Blank When the Language QuestionID is blank, it means the model does not have a language other than English. The language question EN is automatically added to the model. To add another language questionID, navigate to the Edit ACM page and use the right-click menu to select Properties. To add a language question, use the right-click menu on the Edit ACM page and select Add/Language. Noise Filter Is Blank No noise filters are used. To add a noise filter, To add a noise filter, navigate to the Edit ACM page and use the right-click menu to select Properties. See Noise Filter for more information. Compute Accuracy Is On The accuracy of the model will be computed when the training job is run. To deselect Compute Accuracy, navigate to the Edit ACM page and use the right-click menu to select Properties. Compute Trend Is On The trend of the accuracy will be computed when the training job is run. To deselect Compute Trend, navigate to the Edit ACM page and use the right-click menu to select Properties. Use Lemmatisation Is On Lemmatisation involves grouping words with related meanings. Hence lemmatisation might group "soft" and "softness" as closely related. The ACM normally uses lemmatisation to allow it to generalize a given example as being applicable to closely related responses. If you need to consider such related words as distinct concepts, you can turn off lemmatisation for an ACM. This will generally cause the ACM to code fewer responses than it would if lemmatisation were enabled, but allows you to tune the ACM by providing more examples so that it will distinguish closely related words from one another. To deselect Use Lemmatisation, navigate to the Edit ACM page and use the right-click menu to select Properties. Hierarchical Codebook Is On A hierarchical codebook means that the codes below each net are related conceptually; in other words, the codes below a net are refinements of the concept represented by the net. Similarly, sub-nets are refinements of the concept represented by their parent net. Specifying that a codebook is hierarchical will result in faster coding, and may also provide higher accuracy. To deselect Hierarchical Codebook, navigate to the Edit ACM page and use the right-click menu to select Properties. Minimum Confidence Is 1 When a coding job is run, those codes with a confidence value not less than this number will be applied to the response. To change the Minimum Confidence level, navigate to the Edit ACM page and use the right-click menu to select Properties. 404 Ascribe™ Desktop Adds Default Language As Training Example Language The default language for creating models is found on the Account Options page. You can add other languages on the Edit ACM page. Accelerator Coding Models Page Navigate: Accelerator (ACM)/Accelerator Coding Models This page lists all of the accelerator coding models. The filter bar at the top of the page allows you to filter the list of models by the model ID. Here are the fields on the page (you can choose which columns display by selecting Choose Columns from the right-click menu): Field Description ID The unique identifier for the model. Click to view the model details. Description An optional description of the model. Date Created Date the model was created. Match Codes By Specifies how codes are matched between codebooks. Allow Auto Coding If this option was selected in the ACM Properties dialog, the user running a coding job may request that codes be applied automatically. Use Lemmatisation If this is option was selected in the ACM Properties dialog, the user requested to use lemmatisation. Lemmatisation involves grouping words with related meanings. Hence, lemmatisation might group "soft" and "softness" as closely related. If not selected, the user did not use lemmatisation, and the ACM will consider such closely related words as distinct concepts. Hierarchical Codebook If this is option was selected in the ACM Properties dialog the codebook uses nets, and the codes below each net and sub-net are related conceptually. If not selected, the codebook is not constructed this way. Minimum Confidence This option applies only when Allow Auto Coding is selected. When codes are automatically applied, only those codes with a confidence value equal or greater than this value will be applied. Noise Filter The regular expression used to remove noise characters from the response for training and coding. Codebook ID The ID of the codebook being trained. Click to view the codebook. Shared Questions The number of questions that share the codebook. Click to view a list of the questions. Codebook Description The description of the codebook. Approved If a dot is present in this column, the model is approved. Approved By The name of the user who approved the model, or blank if the model is not approved. Trained If a dot is present in this column, the training job has been run 405 PrintDoc01 020813 successfully, and the results may be reviewed. The model can be approved only after it has been trained. Trained By The username of the person who ran the training job, or blank if the training job has not been run. Trained On The date when the training job was run. Click for details about the time. Job Status The status of the training job or blank if the training job has not been run. Click to see the job details. Compute Accuracy If this is option was selected in the ACM Properties dialog, accuracy is computed when the training job is run. Compute Trend If this is option was selected in the ACM Properties dialog, the trend is computed when the training job is run. Languages The number of languages in this model. Click to see a list of the languages. Training Examples The number of training examples. Coded Examples The number of examples that have at least one code applied. Language Question ID The question ID that identifies the language of the responses. This field is used if the responses in the data have more than one language. There is usually a closed-ended question which distinguishes the languages. Coding Questions The number of questions that are assigned to this model for coding. Click to see a list of the questions. Right-Click Menu on the Accelerator Coding Models Page Here are the right-click menu options for this page: Option Description New Opens the ACM Properties dialog box so you can create a new model. Edit ACM Navigates to the Automated Coding Model page where you can add languages and coding questions for a model. Properties Opens the ACM Properties dialog box, if the model has not been approved. Noise Filter Navigates to the Noise Filter page for this model, where you can enter noise characters which will be removed from the response at coding time. Validate Examines the model to see if it was set up correctly. Train Starts the training job, only if the model is not approved. You can check the status of the training job by clicking on the Job Status column in the table. The job status is updated each time you click on it. To update all of the information in the row, click the Update button in the filter bar at the top of the page. 406 Ascribe™ Desktop Training Results Navigates to the Training Results page, where you can view the accuracy and trend for each code. Training Results Editor Navigates to the Training Results Editor (Web) where you can add and modify training examples. Desktop - Training Navigates to the Desktop - Training Results Editor where you can Results Editor add and modify training examples. Approve ACM When you approve a model, you make it available for use in coding. You can approve a training only after a model has been trained. You should approve a model only after you have reviewed the training results and have decided the training is of sufficient quality. After a model is approved, you cannot change the model properties or add training examples. Unapprove ACM If you want to modify or retrain a model after it has been approved, you must unapprove it. When you unapprove a model, it becomes unavailable for coding. It also discards any training results. It does not delete any codes that have been applied. Save ACM This option saves the model to a file. You can retrieve the file from the Jobs page when the operation completes. Delete You can delete a model only if it is not approved. If you wish to delete an approved model, first unapprove it. Choose Columns Opens a dialog box where you can choose which columns display. Noise Filters Navigate: Accelerator (ACM)/Accelerator Coding Models/Right-click a model/Select Noise Filter The noise filter for a model is used to remove unwanted "noise" characters from the response. If present, the noise filter is used to clean responses of noise characters before they are presented to the VCS™ at coding time. The noise filter does not change the response itself; it merely hides the noise characters from the VCS™. For example, suppose that your data collection system inserts the characters < nl> in the response whenever a new line is present. This would confuse the VCS™, because the VCS™ would attempt to use this when it considers whether a code should or should not be applied to the response. Because < nl> is not really a part of what the respondent said, it should be removed from consideration by the VCS™. Other noise characters can be introduced by Ascribe™ itself. If you copy a response to another question in the coding window, Ascribe™ will add characters like {-- copied to q2 -} to the response. These noise characters should also be removed. To add or modify a noise filter to a model, the model must not be approved. When the model is approved, you can use this page to inspect the way the filter works, but you cannot modify the filter. If the model is not approved, you can change the noise filter by opening the ACM Properties dialog. Click the Properties button on the Noise filter page to edit the filter. The current filter is displayed next to the Properties button, but cannot be changed there. 407 PrintDoc01 020813 This page lists the Language questions in the model in a table at the top of the page. The Coding Questions table lists all of the coding questions for the model. Right-click a question and select show responses for this question to see how the filter works on the responses for that question. In the table at the bottom of the page, the first 1000 responses will be displayed. In the left column, the actual text of the coding source field of the response is displayed. In the center column is the same text with the characters removed by the filter highlighted in red. The right column shows the response after the noise characters have been removed. This is the text that is submitted to the VCS™ at coding time. If the filter does not remove any characters from the response, the center and right columns are blank. Constructing a Noise Filter The noise filter is a regular expression. Note that the noise filter regular expression does not use the extended regular expression syntax supported elsewhere in Ascribe™. The noise filter is a "plain old regular expression". In most cases, the regular expressions you will use for filters are quite simple to construct. If you want to explore the full power of regular expressions for filtering, see .Net Framework Regular Expressions on the Microsoft site. The noise filter works like this: any characters that are matched by the regular expression are removed. In our example, we would like to remove < nl>, and also any blocks of text that start with {-- and end with --}. Here is the regular expression we can use for the job: < nl>|{-- .* --} Notice that we use the vertical bar | to separate the two things we want to remove. The vertical bar means match the expressions on either side. The characters "*" and "." are special in regular expressions (as is the vertical bar!) The dot character "." means "match any character", and the "*" character means "match any number of the preceding expression". So, .* means "match anything at this location". Special Characters for a Noise Filter There are several special characters used in regular expressions, and you should know what these do before you create your own regular expressions. Remember this important fact: if a character does not appear in the list below, it is not special, and matches itself in regular expressions. Here is the list of special characters: .$^{[(|)*+?\ The table below describes what each of these characters mean in a regular expression. Characters Description \ Indicates that the following character is not a special regular expression character. For example, \* matches an asterisk. . The period character matches any character except \n (the newline character). [ chargroup] Characters in square brackets match any character in the chargroup. For example, [ aeiou] matches any lowercase vowel. [^ chargroup] When the ^ character appears as the first character inside square 408 Ascribe™ Desktop brackets, it matches any character not in the chargroup. For example [^ aeiouAEIOU] matches any character that is not a vowel (upper or lower case). [firstchar-lastchar] Matches any character from firstchar to lastchar. For example, [09] matches any digit. [a-fA-F] matches any character from A through F (both upper and lower case). ^ Specifies that the match must occur at the start of a line. For example, ^< matches < characters that occur at the start of a line. $ Specifies that the match must occur at the end of a line (or the end of the response). For example >$ matches > characters at the end of a line. (group) Parentheses group expressions together, so that you can use a special character to perform matching on the entire group. For example (abc)+ matches the characters abc one or more times. * Matches the preceding character or group zero or more times. For example, <* matches the character < zero or more times. + Matches the preceding character or group one or more times. For example, <+ matches the character < one or more times. ? Matches the preceding character or group zero or one time. For example, <? matches the character < zero or one time. {n} Matches exactly n times. For example, a{3} matches aaa. The expression (pizza){2} matches pizzapizza. {n,} Matches at least n times. {n,m} Matches at least n, but no more than m times. | Matches any one of the terms separated by the | (vertical bar) character; for example, cat|dog|tiger. The leftmost successful match wins. ACM Properties Dialog Navigate: Accelerator (ACM)/Accelerator Coding Models/Right-click/Select New Accelerator (ACM)/Accelerator Coding Models/Right-click a model/Select Properties The properties dialog is displayed when you create a new model or modify an existing model. This dialog is not available if the model is approved. If you want to modify the model properties, the model must not be approved. The dialog box contains these entries: Field Description ID The unique identifier for the model. No two models can have the same identifier. The ID must be 30 characters or less. Description An optional description of the model. Match Codes By Specifies how codes are matched between the model's codebook and the coding questions' codebook. The options are: None - All coding questions must share the model's 409 PrintDoc01 020813 codebook. InputID - Codes are matched by case insensitive textual comparison of the InputID. When this rule is used, 001 does not match 01. OutputID - Codes are matched by case insensitive textual comparison of the OutputID. Description - Codes are matched by case insensitive textual comparison of the text of the code. Numeric InputID - Codes are matched by attempting to convert the InputID to an integer. If this conversion is successful, the codes are compared by this number. If unsuccessful, codes are compared by case insensitive textual comparison. When this option is used, 001 matches 01. Numeric OutputID – Same as for numeric InputID, but using the OutputID to match. Search for Codebook By You can locate a codebook for the model either by entering the ID of the codebook, or by entering a study and question ID. Codebook ID This box displays if you are searching for the codebook by codebook ID. Begin typing in this box to see a list of codebooks with ID's that start with the characters you typed. We recommend that you label the codebook before you create a model to simplify the lookup of the desired codebook. Study ID This box displays if you are searching for the codebook by question ID. Begin typing the ID of the study that contains the desired question. Choose the study from the list of studies displayed. Question ID This box displays if you are searching for the codebook by question ID. Begin typing the ID of the desired question. Choose the question from the list of questions displayed. The codebook of this question is the one that the model will use. Language Question ID This entry should be used only if the model has more than one language, and in that case, it is required. This entry specifies the ID of the question (usually a closed-ended question) that distinguishes the language of the response. See Multi-Lingual Models for more information. Noise Filter The regular expression used to remove noise from the response at coding time. See Noise Filter. Compute Accuracy If checked, accuracy of the model will be computed when the training job is run. This increases the time required to run the training job. Compute Trend If checked, the trend of the accuracy will be computed when the training job is run. This increases the time required to run the training job. Use Lemmatisation You can look up lemmatisation in Wikipedia for a more detailed description, but basically lemmatisation involves grouping words with related meanings. Hence lemmatisation might group "soft" and "softness" as closely related. The ACM normally uses 410 Ascribe™ Desktop lemmatisation to allow it to generalize a given example as being applicable to closely related responses. If you need to consider such related words as distinct concepts, you can turn off lemmatisation for an ACM. This will generally cause the ACM to code fewer responses than it would if lemmatisation were enabled, but allows you to tune the ACM by providing more examples so that it will distinguish closely related words from one another. Hierarchical Codebook You may specify that the codebook for an ACM is hierarchical. This means that the codes below each net are related conceptually; in other words, the codes below a net are refinements of the concept represented by the net. Similarly, sub-nets are refinements of the concept represented by their parent net. If the entire codebook is constructed in this fashion, it is hierarchical. Specifying that a codebook is hierarchical will result in faster coding, and may also provide higher accuracy. Minimum Confidence When a coding job is run, those codes with a confidence value not less than this number will be applied to the response. Training Example Editor Navigate: Accelerator (ACM)/Accelerator Coding Models/Right-click a model/Select Training Example Editor Accelerator (ACM)/Accelerator Coding Models/Right-click a model/Select Training Results/Right-click a code or a language/Select Training Example Editor Use the Training Example Editor to enter uncoded examples, copy uncoded examples, code examples, or copy already coded data to the model. You can also delete examples from the model. You can only add or edit examples if the model is not approved. The Training Example Editor is divided into three sections: the filter bar, Training Examples table, and Training Results section. Filter Bar The filter bar has these fields: Field Description Update If you make changes to Language or Sort Order, you must click Update for the changes to take affect. Language A drop-down list of available Languages for the model. If you want to change the Language, select the language from the drop-down box and click the Update button. This change is not in effect until you click Update, even though the filter box displays the newly-selected Language. Sort Order Sort the examples by example text, example length, number of codes applied, or code description. This change is not in effect until you click Update. Descending Check Box Allows you to sort in descending order. This change is not in effect until you click Update. Train Starts the training job for the model. The model must be unapproved in order to train it. View the job on the General/Jobs 411 PrintDoc01 020813 page. Add Examples Add 1-column text files as training examples with this function. Training Results Section The Training Results section has several tabs to provide information about the model: Training Results, ACM, Word List, and Tester. The ACM and Tester tabs display only after a model has been trained. Training Results Tab The Training Results Tab has these columns (you can change your display through the rightclick menu option of Choose Columns): Columns Description Code List of codes that were trained with examples. Note that the -Uncoded Examples-- field contains examples that have not been coded. To display them, right-click and select Show Examples. Examples Count of examples that have this code applied. Trained Count of examples that were trained; this number will differ from the Examples column if examples are added after the model has been trained. ACM Applied After the model has been used to code a question, this column contains the count of responses in the coding questions that received this code. Accuracy Average The average of precision and recall. Precision The precision score indicates if the model over-applied the code. The lower the precision score, the more the model over-applied the code. Recall The recall score indicates if the model under-applied the code. The lower the recall score, the more the model under-applied the code. Aggregate The ration of the number of times ACM applied the code vs. the number of times a human applied the code. The right-click menu for the Training Results Tab has these options: Option Description Show Examples Display the training examples for this code in the Training Examples table on the left side of the screen. Create New Example Allows you to enter a new training example for this code on the left side of the screen. The Example Text defaults to 'new example' and can be changed as appropriate. Double-click the new example to edit it. A dialog opens where you can make changes. 412 Ascribe™ Desktop Copy Examples Opens dialog and displays training examples which have that code applied. To copy examples, check Select All or click the check boxes next to the examples you want to copy. To copy only quality reviewed examples, click the Quality Reviewed check box. To apply codes to the copied examples, click the Apply Codes check box. Choose Columns Choose which columns to display in the tab. ACM Tab The ACM Tab displays the following information about a model, after it has been trained: Field Description ACM Description Description from the ACM Properties dialog. Last Trained Date the model was last trained. Avg Accuracy Average accuracy score of all of the codes. Avg Precision Average precision score of all of the codes. Avg Recall Average recall score of all of the codes. Word List The Word List displays the count of all words found in the training examples. To see the examples which contain those words, right-click a word and select Show Examples. The examples display in the Training Example section of the page. Tester Tab After a model has been trained, you can use the Tester to see how a response would be coded. It can be a quick way to check the model's accuracy in coding responses. Enter the response in the text box and click Test It. The coding results display below the text box. Training Examples Table The Training Example Table has two columns: Example Text and Codes. Functionality is controlled through the right-click menu or by double-clicking in one of the columns. See Add Examples and Add,Change, or Delete Codes for information on the double-click options for each column. The right-click menu for the table has these options: Option Description Accelerator (ACM) Navigate to Accelerator Coding Models page. Edit ACM Navigate to Accelerator Coding Model page. Review Examples Navigate to Review Examples, where you can use the basics of Review Responses to view, change, or add codes for training 413 PrintDoc01 020813 examples. Copy Examples Allows you to copy responses from questions to the model and use them as training examples. See Copy Examples from Shared Questions. Training Results Navigate to the Training Results page. Select All Selects all training examples and codes which are in the current result set. Clear All Clears the highlight and unselects all of the training examples and codes which are in the current result set. Remove Selected Examples Removes selected training examples from the current result set. Change Column Widths Allows you to adjust the column widths of the columns in this table. Add Examples If a model is not approved, you can enter examples in the Example Text Table. You have to show some examples from the codelist first (right-click a code in the Training Results pane and select Show Examples.) Double-click the left column of a training example (the column under the Example Text heading.) A dialog displays with the current example. You can type new examples under the current example. Press Enter after each example. You can also copy/paste examples into the dialog box. Here are the options for the dialog: Option Description Cancel Closes the dialog and does not save any changes. OK Saves changes and closes the dialog. 414 Ascribe™ Desktop Auto Scroll The background screen automatically scrolls if the list of examples is long enough. Apply Codes Applies the same code or codes as the current training example (the training example you double-clicked to open this dialog.) Uncheck this option if you do not want to apply the same code or codes. Note: You can add, change, or delete codes by using the double-click option for the Codes column. Add, Change, or Delete Codes If a model is not approved, you can add, change, or delete codes for the examples. You have to show some examples from the codelist first (right-click a code in the Training Results pan and select Show Examples.) Once you have the examples with which you want to work, double-click the right column of a training example (the column under the Codes heading.) A dialog displays with the codebook. Click the desired code or codes to apply them. The code or codes move to the white space above the codebook. You can remove a code from the white space (which means it will be not applied) by clicking it. Click OK to close the box. The codebook box has an option to auto scroll for the background screen, which makes it easier to scroll through a long codebook. Use the auto close option if you want to close the box after selecting one code. Copy Examples from Shared Questions Navigate: Accelerator (ACM)/Accelerator Coding Models/Right-click a model/Select Training Example Editor/Rightclick/Select Copy Examples The model must be unapproved in order to use the Copy Examples function. 415 PrintDoc01 020813 Note there are two types of Copy Examples. If you want to copy examples for a specific code, right-click on the code in the Training Results tab. See Copy Examples in the Training Results tab for more information. To copy examples from questions that are shared with the model, right-click in the Training Examples table or anywhere on the page (with the exception of on a code in the Training Results tab.) and select Copy Examples. The Copy Examples dialog displays with these fields: Field Description Available Questions A list of all questions that are shared with the model (the question must contain at least one response.) The study ID, question ID, question label, number of responses, number of coded responses, and number of reviewed responses are listed. Click the question from which you want to copy examples. Use Ctrl/Click to select multiple questions. Copy Options Gives these options for copying training examples: Copy all responses as examples - All responses from the question are copied to the model. Select a sample of responses to copy - Enter a number or a percentage of the responses to be copied via random selection. Quality Reviewed responses only - Only responses that have been marked as Quality Reviewed are copied. Leave the check box blank if you want to copy responses regardless of their Quality Reviewed status. Coded only - Only coded responses are copied. Leave the check box blank if you want to copy both coded and uncoded responses. Select the question or questions from which you want to copy examples, select the copy options, and click OK. When copy process completes, a message displays the number of examples copied. The example text is selected from the coding source of the selected question. If the coding source is verbatim, the text is selected only if it is an alphabetic response. Media responses and those that are numeric only are not selected. The selected text is then cleaned by collapsing all successive white space (space, tab, end of line characters) to a single space. Leading and trailing white space is removed. If the resulting text is an empty string, it is not selected. Finally, if the text is selected, it is inserted as a new example. Codes applied to the example are also copied. No coding fee (transaction) is applied to these examples. If you later change the coding of an example, a transaction fee is applied. Considerations Before Approving a Model 416 Ascribe™ Desktop Trained models may be used for coding only after they have been approved. The intention is to control coder access to the model, so that you are assured that only models of sufficient quality are used for coding. After running a training job, you should assess the quality of the training results. If you deem the quality to be acceptable, you can approve the model to make it available to coders. To again make the model unavailable to coders, you can unapprove the model. Assess the Quality of Training Results Before you approve a model and thereby make it available for coding, you should assess the quality of the training results. This helps you decide whether the quality of the model is expected to be sufficient to use it for coding. The Training Results page is your primary tool for examining the quality of training results. When you are first building a new model, we recommend that you check the Compute Accuracy, and optionally the Compute Accuracy and Trend boxes in the ACM Properties dialog. This will increase the time required to run the training job, but will provide information about the quality of the training. This information is in the form of metrics calculated by the VCS™. In the Training Results page, you should consider these factors: Are there training examples for each code? If there are no training examples for a code, it will never be applied by the ACM. This may be appropriate and intentional in some cases, but generally you will want to ensure that each code has some training examples. Is the number of training examples sufficient for each code? There is no "correct number" of training examples for a code, and you will have to be guided by your own experience. Codes such as "mention of cool or cold" will require fewer examples than subtle concepts such as "dissatisfied with service". Is the F1 accuracy number low for a certain code relative to the others? A low F1 score may mean that the examples provided are contradictory. Examine the training examples for the code to search for conflicting examples. Once you have determined that the training results have no obvious problems, you may want to approve the model and try coding a sample question without automatically applying the codes. This will let you examine the codes suggested by the ACM without incurring transaction fees (provided you do not accept the suggested codes). Approve a Model When you approve a model, you make it available for use in coding. You can approve a model only after a training job has been run successfully. You should approve a training only after you have reviewed the training results, and have decided the training results are of sufficient quality. You can approve a training in the Automated Coding Models page or the Automated Coding Model page. After a model is approved, you cannot change the properties of the model. If you want to modify the model, you must first unapprove the training. Unapprove a Model 417 PrintDoc01 020813 When you unapprove a model, it becomes unavailable to use for coding. Unapprove a model if you no longer want it to be used for coding. You must also unapprove a model if you want to modify it, or if you want to retrain it by running a training job again. When you unapprove a model, you also discard any training results. To use the model for coding again, you must run the training job, approve the model, and then run any desired coding jobs. You can unapprove a training in the Automated Coding Models page or the Automated Coding Model page. Training Results Page Navigate: Supervisor/Automated Coding Models/Right-click a model/Select Training Results Automated Coding Model/Right-click/Select Training Results The Training Results page displays the results of a training job. This page is intended to allow you to decide whether to approve the model or whether you need to modify the training examples and retrain the model. The page contains three tables: ACM Properties, Languages, and Codebook. ACM Properties This table displays summary information about the model. Click the ID for more detailed information. Languages Table This table lists the Languages in the model. If the Compute accuracy property of the model was selected at training time, overall accuracy metrics are shown for the language. If the Compute trend property of the model was selected at the training time, overall trend metrics are shown for the language. The Codebook Table shows the training results by code for one of the Languages. To switch languages, right-click on the language in this table and select Show codebook for this language. In this table, the aggregate training metrics for each language are displayed. These numbers give you an overall estimate of the accuracy of the training for that language. You can also navigate to the Training Example Editor if you right-click a row in this table. Codebook Table This table shows the training results by code for one of the training languages (the language displayed is indicated in the title above the table). To switch languages, right-click on the desired language in the Languages Table and select Show codebook for this language. For each code in the table, the training metrics for the code are displayed. These metrics provide an estimate of the training accuracy for that specific code. The Training Consistency Score is intended to give users insight into the effectiveness of the training examples provided for each code. The data is reported in both the Training Results Page and the Training Example Editor in Ascribe™ Web. 418 Ascribe™ Desktop You can navigate to the Training Example Editor or the Training Code Cleaner if you rightclick a row in this table. View Training Examples To view the training examples used for a code, right-click the Training Examples column for that code. A drop-down list of the training examples displays. The Training Consistency Score for each example also displays. It is very important to understand that the examples shown here are those submitted at training time. After the training process is performed, the examples may be changed using various features of Ascribe™. Changing a example after you run the training job will not affect the examples displayed in this window. The ACM remembers the examples submitted at training time, and those are the ones displayed in this window. Training Code Cleaner Navigate: Supervisor/Automated Coding Models/Right-click a model/Select Training Results/Right-click a code in the Codebook table/Select Training Code Cleaner The Training Code Cleaner page facilitates the tuning of ACM training examples by providing view into the consistency which with each code is trained. It displays a list of training examples with those identified by the ACM as most likely to be inconsistent at the top, whether the inconsistency is that the code is applied, or not applied. The code can be added to or removed from an example via the check box at the left of each one. The capability is intended to be used to quickly correct coding errors in a training example set that is producing undesirable results. The Training Example Editor is still available for more significant changes to the examples. The page contains two tables: Code table and Training Examples table. The Code table has these fields Field Description ACM The name of the model. Codebook ID The codebook ID if it exists. Language ID The language that the model uses. Code Description The code which is being cleaned. The Training Examples table has these fields: Field Description Apply The check box is grayed out if the code has been applied. If you want to apply the code, leave the boxed checked. If you do not want to apply the code, remove the check. Consistency Score The Training Consistency Score is intended to give users insight into the effectiveness of the training examples provided for each code. 419 PrintDoc01 020813 Example Text This field displays the training example. Code Questions using a Model To code a question using the ACM, you must associate the question with the model. You can then code the question by running a coding job for the question. Run an ACM Coding Job To code a question with the ACM, you must first associate the question with an ACM training to use for coding the question. To code the question using the ACM, navigate to the question in the Supervisor/Studies/Questions page. Right-click on the question and select ACM/Start ACM coding job... from the menu. In the dialog displayed, specify whether you want the ACM to code all responses or only uncoded responses. Normally you will specify the default, which is Uncoded only. If you want the ACM to also apply additional codes it selects to previously coded responses, select All responses. Assess the Quality of ACM Coding Once you have coded a question with the ACM, you will want to review the quality of the coding it performed. You can use the same tools and techniques you would use to evaluate the quality of a human coder. Ascribe™ does provide a tool that is specially designed for assessing the quality of ACM coding, the Review ACM Coding window. Using this tool, you can filter and sort the responses to help locate coding issues. You can also review and approve codes suggested by the VCS™ if they were not automatically applied. To review ACM coding using the Review ACM coding window, right-click on the question and select Review ACM coding. Improve ACM Coding Quality To improve the quality of coding using the ACM, you can tune the set of training examples used at training time. There are several factors affecting ACM coding quality. Understanding these will help you improve your training examples. A key advantage of the ACM is that you do not need special linguistic analysis skills to hone the quality of ACM coding. All that you need to do is provide a proper set of coded examples. This section lists several considerations to aid you in developing a good set of training examples for your model. Number of Training Examples per Code The first consideration as you create your set of training examples is how many examples you should supply for each code. In general, the more properly coded examples you provide for a given code, the better the result. The number of examples necessary will, however, depend on the nature of the code being trained. As a simple example, suppose that you have a code that should be applied for specific word or phrase mentions, such as "close account" or "discontinue service". In this case a very few training examples (being the set of words or phrases to look for) may be all that is needed to train that code. 420 Ascribe™ Desktop On the other hand, if the code involves more subtle concepts such as emotions, or if there are many ways to express the general concept, then many training examples may be necessary. Such codes may require a hundred or more examples to achieve the required quality. You can use this idea to good purpose as you review the training results. When looking at the number of examples provided for a given code, ask yourself whether this number seems appropriate given the conceptual complexity of the code. Quality of Training Examples Although it may seem obvious, we must emphasize that the training examples you provide must be properly coded. Unlike a human, the ACM is not able to say "that example is obviously ridiculous, I'll ignore it." While the VCS™ engine does employ sophisticated techniques to minimize the effect of a few improper examples, such examples will still lower the overall quality of the results. It is better to have fewer well coded examples than more examples of lower quality. Conceptual Clarity of Codes The construction of your codebook will have a strong effect on the quality of coding the ACM can produce. If two human coders would have difficulty selecting between two codes, you can expect that the ACM will have similar difficulties. Codebooks with subtle distinctions between codes, or with codes that have overlapping meaning, will make the use of the ACM much more difficult. Specificity of Coding The ACM will perform best if the codes provided in training examples apply to the entire coded response. Longer responses there may contain multiple concepts. If these concepts are not interrelated, you can improve the ACM training by breaking the response into shorter segments, and applying the proper codes to each segment. See Dividing Responses into Concepts for more information. Extraneous Text in Responses Removing text from a response where that text has no bearing on the concept coded will improve the quality of your model. For example, consider this response to a survey on amusement parks rides: "My daughter enjoys the Log Flume when she visits with her Girl Scout troop." This is coded "Log Flume" in the likes category. Note that the information about the Girl Scout troop contributes nothing to the concept of liking the log flume. This example could be made more meaningful as a training example if it were reduced to: "My daughter enjoys the Log Flume." Negative Examples The ACM pays as much attention to codes that are not applied to the training examples as to codes that are applied. This has two important implications: Errors of omission will degrade the quality of the training results. 421 PrintDoc01 020813 To train the ACM not to apply a code in a certain circumstance, provide appropriate training examples where that code is not applied. To illustrate the first point, suppose the response: "My daughter likes the Ferris Wheel, the airplane ride, and is really keen on the roller coaster." Suppose this response is coded with Airplane Ride and Roller Coaster, but not with Ferris Wheel. This is telling the ACM that the Ferris Wheel code should not be applied to responses of this type. This will degrade the quality of the training result for the Ferris Wheel code. If you find that the ACM is not applying codes to responses when it should, you should look for training examples where that code is not applied. You probably will find examples where the code should have been applied. Fixing these will improve the quality of the training. For the second point, if you find that the ACM is applying a code when it should not do so, you should certainly check the training examples where that code is applied, and verify that it has not been applied mistakenly to any of the training examples. If this does not resolve the problem, you can construct examples to illustrate the point and ensure that the problem code is not applied. This will aid the ACM in determining when that code should not be applied. Divide Responses into Concepts When multiple concepts are present in a training response, you can improve the quality of the training by dividing the response into multiple responses. You can divide the response so that each part expresses a single concept. You can use the Training Example Editor to make the type of changes suggested in this section. Suppose in a study on amusement parks we find the following response: "My husband and I take the children to the park twice a summer. He thinks going more often would spoil the children. My son really likes the auto race. I don't care for the auto race myself. It's too loud. We really like the park, but the parking lot needs work. There are potholes everywhere." We will assume we have a codebook divided into Likes and Dislikes nets, with codes for Auto Race and Parking Lot. Now, how can we improve our training using this example? Let's tackle the Parking Lot code first. The Dislikes/Parking Lot code should be applied to this example. Because only the last two sentences relate to this concept, we can improve the training by making a training response: "We really like the park, but the parking lot needs work. There are potholes everywhere." To which we apply only the Dislikes/Parking Lot code. Now, what about the auto race? We first observe that only these sentences relate to the auto race: "My son really likes the auto race. I don't care for the auto race myself. It's too loud." Should we break this into two more responses, one for the Likes and another for the Dislikes? This will depend on your coding conventions. If your coding conventions allow a 422 Ascribe™ Desktop single response to have both the Likes/Auto Race and Dislikes/Auto Race code, then you should divide this into two responses: "My son really likes the auto race." "I don't care for the auto race myself. It's too loud." You would apply the Likes/Auto Race to the first and Dislikes/Auto Race to the second. You are instructing the ACM to consider the concepts independently. If, however, your coding conventions dictate that the Dislikes code should take precedence when there is conflicting content, you should use the full segment: "My son really likes the auto race. I don't care for the auto race myself. It's too loud." as the training example, and apply only the Dislikes/Auto Race code to it. This instructs in your coding conventions for this case. This leaves us with this portion of the response: "My husband and I take the children to the park twice a summer. He thinks going more often would spoil the children." If this concept has nothing to do with the intent of the survey (no codes exist to capture this meaning), it can be removed from the training example. It provides no helpful information to the ACM for the training. Accelerator Coding Model Page Navigate: Accelerator (ACM)/Automated Coding Models/Right-click a model/Select Edit ACM This page displays the details of a model and allows you to modify some aspects of the model. If the model is approved, you can only add or remove coding questions. To modify the training in other ways, you must first unapprove the model. The Accelerator Coding Model page has this information: ACM Properties Table, Languages Table, and Coding Questions Table. Options for this page are controlled by the right-click menu for the Accelerator Coding Model and the right-click menu for the Coding Questions Table. ACM Properties Table The first table on the page displays the ACM properties. Several of the fields have dropdown boxes that display additional information about the item. If you want to modify the properties, the model must not be approved. To modify the properties, right-click in the table and select Properties. The ACM Properties dialog box displays. Languages Table Each Language in the ACM holds the training examples for the model. For example, a model might have three Languages: French, German, and Italian. Each of these Languages holds the training examples for the ACM in that language: French examples in the French Language, German examples in the German Language, and so on. A model must have at least one Language (and often has only one Language). The Languages table lists the languages for this model. 423 PrintDoc01 020813 It is very important that the Languages specified in this list match the languages found in the responses used as training examples and coding questions. Training or coding a German response using the English language setting will not yield good results! Add a Language To add a language, right-click and select Add/Language. A dialog box displays with a dropdown list of available languages. Select the desired language, and click the OK button. Multiple Languages If the Language Question ID property of the model is not blank, then you may specify more than one Language in the Languages table. When the Language Question ID property of the ACM Training is not blank, certain rules apply both at training and coding time: For each question being coded by the model, a closed-ended language question with the specified ID must be present in the study. The Language response column in the Languages table may not be blank. For each response considered at training and coding time, the respondent must have also given a response to the question specified by the Language Question ID. In addition, this response must match one of the specified Language response values in the Languages table. Otherwise, the response is ignored. Coding Questions Table This table lists all questions that may be coded with this model. A given question may appear in the Coding Questions table of no more than one model. You can add or remove questions from a model at any time, whether or not the model is approved. To add one question at a time, right-click and select Add/Coding Question. The Question Selection Dialog displays. Enter the study ID and question ID of the coding question. To add all questions which share the ACM codebook, right-click and select Add/All questions sharing the ACM codebook. This action will not add questions that are already assigned to a model for coding. To remove a question, right-click on the question and select Remove. Right-Click Menu on the Accelerator Coding Model Page Option Description Accelerator (ACM) Navigates to the list of all models on this site. Properties Opens the ACM Properties dialog box; option is available only if the model has not been approved. Validate Examines the model to see if it was set up correctly. Option is available only if the model has not been approved. Approve ACM When you approve a model, you make it available for use in coding. You can approve a training only after a model has been 424 Ascribe™ Desktop trained. You should approve a model only after you have reviewed the training results and have decided the training is of sufficient quality. After a model is approved, you cannot change the model properties or add training examples. Unapprove ACM If you want to modify or retrain a model after it has been approved, you must unapprove it. When you unapprove a model, it becomes unavailable for coding. It also discards any training results. It does not delete any codes that have been applied. Train Starts the training job; option is available only if the model is not approved. Add Options are: Language - If the model is not approved, you can add another Language to the Languages Table. Coding Question - The Question Selection dialog opens so you can add a coding question. Questions Sharing the ACM Codebook - Adds all questions which share the ACM codebook. This action will not add questions that are already assigned to a model for coding. Training Results Navigates to the Training Results page, where you can view the accuracy and trend for each code. Option is available only if the model has been trained. Right-Click Menu for the Coding Questions Table When you right-click a question in the Questions Table, the following options are available: Option Description Start ACM Coding Job The option is available only if a model has been approved. A dialog box displays. Select All Responses to code all of the responses, including those that already had codes applied or select Uncoded Only. You also have the choice to apply suggested codes to uncoded responses. If you do not select to apply the suggested codes, you must use the Desktop - Review ACM Coding page to apply the codes. Desktop - Review ACM Coding Job Add This option is available only if the question has been coded. Navigates to the Desktop - Review ACM Coding page. Options are: Coding Question - The Question Selection dialog opens so you can add a coding question. Questions Sharing the ACM Codebook - Adds all questions which share the ACM codebook. This action will not add questions that are already assigned to a model for coding. Remove Removes this question from the model. It does not delete the question itself. Questions Navigates to the Study Questions page for this question. Choose Columns Opens a dialog box where you can choose which columns display 425 PrintDoc01 020813 for this table. Question Selection Dialog Navigate: Supervisor/Automated Coding Models/right-click a model/Select Automated Coding Model/Rightclick/Select Add/Coding Questions This dialog box allows you to select a coding question for an automated coding model. Field Description Study ID The ID of the study that contains the desired question. To pick a study, start typing the study ID or name. A drop-down list appears, listing the studies with ID or name beginning with the characters you typed. Use the mouse or up-down arrow keys to select the desired study, and then press the tab key to move to the Question ID field. Question ID The ID of the desired question. To pick a question, first pick the desired study. Then, in the question ID box, start typing the question ID. A drop-down list appears, listing the questions with ID beginning with the characters you typed. Use the mouse or updown arrow keys to select the desired question. When you have picked the desired question, click the OK button to add this question as a coding question for this model. Ascribe™ Desktop Ascribe™ Desktop is not a "browser-based" application as Ascribe™ Web is today. Instead, it is a Windows application that executes on your local machine. This approach offers seamless updates to the program, more options for improved performance and visual experience, and better control of security options. Install Ascribe™ Client Applications: Desktop and Crossword ClickOnce enables the user to install and run a Windows application by clicking a link in a web page. The core principle of ClickOnce is to bring the ease of deployment of web applications to the Windows user. Such web applications are called Smart Client or Rich Internet applications (RIAs) because they are web applications that have some of the characteristics of desktop applications. Language Logic moved to smart client technology when developing Desktop and Crossword. This move eliminates programming dependency on the Internet Explorer which is more restrictive than desktop applications, and allows the development team more programming power to meet user requests such as the ability to highlight and move multiple codes at one time. Both Ascribe™ Desktop and Ascribe™ Crossword are considered Language Logic client applications. System Requirements Language Logic client applications run on Microsoft Windows operating systems only. They are not available for 64 bit platforms. They require Microsoft Windows 2000, XP, Vista, or Windows 7. 426 Ascribe™ Desktop A connection to the Internet is required both for installation of client applications, and when client applications are running. You can not run Crossword or Desktop without an active Internet connection. Language Logic client applications require either Microsoft .Net Framework Client Profile or Microsoft .Net CLR version 3.5 SP1. The Client Profile is a subset of the full .Net Framework 3.5 SP1. The Client Profile allows users to install just those components needed to allow the “click-once” install technology. Note: Installation of the Client Profile or Microsoft .Net CLR version 3.5 SP1 requires local administrator privilege on the user’s computer. Crossword and Desktop are Web Services applications using port 80. Firewalls and proxy servers will need to allow access for these applications. Installing Crossword/Desktop You can install the Crossword or Desktop application on your computer from the Ascribe™ software. 1. From the Home page on Ascribe™ Web, click Downloads on the left-hand menu. 2. From the Downloads page, click the link for Ascribe™ Desktop or Ascribe™ Crossword. 3. Results of clicking the link: 427 PrintDoc01 020813 a. If the required Client Profile is already installed, the links will connect to the Desktop and Crossword applications themselves. b. If the required Client Profile is not already installed, the links will connect to the setup programs for the Desktop and Crossword applications. The setup program will prompt the user to install the Client Profile (see next step.) 4. Once .NET CLR 3.5 or the Client Profile is installed, the “Support Information” area on the “Downloads” page displays the version installed. If the information under “Installed frameworks” states “Required framework not installed,” the user will need to attempt to re-install the software by clicking the link. Remember: users need local administrator privilege on their computer to install the .Net Framework. If you do not have local administrator privilege, you should contact your IT department to install the required framework. Common Key Usage Ascribe™ Desktop uses several certain keyboard keys to invoke common global actions. The keys and their resulting action are listed below. Key Action F1 Displays the help page for the current context F2 Displays the Home View tab. Equivalent to clicking the 'Home' button F3 Moves the mouse cursor and keyboard focus to the 'Find a Study' box F5 Refreshes the current context (i.e., window/tab in the Data Pane) F7 Toggles grid editing on or off F8 Same behavior as clicking the 'Back' button Terminology The following terms apply to Ascribe™ Desktop. Term Definition Main Menu The topmost menu of the application located across the top of the application. 428 Ascribe™ Desktop Main Toolbar The topmost toolbar of the application located immediately below the Main Menu. Navigation Pane The left side of the primary window area immediately below the Main Toolbar. Contains two tabs: the Navigator Tree and the Workspace. Data Pane The right side of the primary window area immediately below the Main Toolbar. Contains one permanent tabbed window (the Home View) and one or more tabbed windows based upon user activities. Filter Bar The area of the application immediately below the Navigation Pane. Contains two buttons: one to filter studies and one for questions. Navigator Tree One of the tabs of the Navigation Pane. Studies and Codebooks are displayed in a tree-like list. Workspace One of the tabs of the Navigation Pane. It is an area, initially empty, that can contain user provided items such as studies, questions, codebooks and files. Home View The first and only permanent tab of the Data Pane. It contains either the Welcome screen or a display based upon the user selection in the Navigation Pane. Status Bar The strip across the bottom of the application. It contains three distinct panels. The left-most panel typically displays connection status. The middle and right panels provide transient information related to current activities. Context Menu The menu of options that appears when you right-click an item. The options available in these menus will vary based upon what item is being selected and, possibly, if more than one item is selected. Tabbed Window One of the windows available in the Data Pane. The Home View tabbed window is an example. All functions and views selected by the user are presented in a tabbed window contained inside the Data Pane. Control Cell The cell in the upper left corner of a table of items. Log On Upon startup of Ascribe™ Desktop, a "Logon screen" displays that requests the same information as the current Ascribe™ Web logon. Type the Ascribe™ site, your user name, and password, just as you do in Ascribe™ Web. Upon successful connection to the Ascribe™ servers, Ascribe™ Desktop displays the main Navigation Window. Tip: You can determine connection status by the message located in the lower left corner of the main display window. 429 PrintDoc01 020813 A new Ascribe™ "session" is created just as it is when a user logs into Ascribe™ Web. Session hours, time tracking, and other statistics are maintained in exactly the same fashion as the Web. Note: Your password can be changed at the time of logon from the Logon screen. Click the link at the lower left of the Logon screen after you have entered your current password. Type the new password in each of the boxes that display and then click "Change" password. Log Off You can close Ascribe™ Desktop by any of these options: Select the Exit option from the Main Menu/File option Select the Log Off option from the Main Menu/File option Click the red "X" in the upper right corner of the Application Main window. Refresh The process of refreshing involves updating the current information about an item or list of items from the database and then presenting the screen with the revised information. You can refresh Ascribe™ Desktop by any of these options: Click the Refresh button on the Main Toolbar Press F5 Select the Refresh option from various context menu options. What is updated from the database when the Refresh process starts depends on what currently is selected and how the Refresh process was invoked. This table summarizes the various refresh options. If Refresh is started by... And Pressing F5 or clicking the 'Refresh' button The Home View Tab is visible Whatever is displayed in the in the Data Pane Home View will be refreshed. This typically will be a list of studies, questions, or codebooks. Pressing F5 or clicking the 'Refresh' button A tabbed window other than the Home View is visible in the Data Pane Right-clicking a tabbed Window's context menu 'Refresh' option 430 Then Whatever item or related items are being displayed in that window will be refreshed. For example, in the codebook editor, the codebook and all the codes will be updated from values in the database. Same as above Ascribe™ Desktop Right-clicking an item's context menu Refresh option in the Navigator Tree, the Workspace, or a Home View list The information for the particular item will be updated from the database. If the item is a list collection of items, such as a Questions node in the Navigator Tree, then each item in the list is updated. Grid Editing Grid editing refers to the ability of the user to edit individual cells of a table view of studies, questions, or codebooks in the Home View in an Excel ®-like manner. This allows you to be able to change multiple attributes of an item quickly or the same attribute of multiple items. When Grid Editing mode is enabled, a "properties window" displays in the lower half of the Data Pane. This property window displays the properties of whatever item is selected in the upper half table. As the user changes selection in the table, the properties window changes to reflect the new item. The properties window allows for some properties (those in the section labeled " Editable") to be changed. When more than one item is selected in the table, a reduced set of properties is displayed in the properties window. These are the properties that can be changed for more than one item. For example, when one study is selected in a list of studies, the property study ID can be changed. However, when more than one study is selected, the study ID property is not available because the same study ID cannot be assigned to more than one study. Grid Editing is enabled by using the Control Cell context menu of the table. Main Display and Navigation The Ascribe™ Desktop main window has five main visual components. These include: Main Menu, Toolbar, Navigation Pane (left side), Data Pane (right side), and Status Bar (at bottom of screen.) Main Menu The Main Menu has five top-level options, each with several sub-options as described in the table below. Option Functionality File Log On You can use this option to log on to Desktop if you have logged off and still have the Desktop screen open. Log Off Allows you to log off Desktop, but the application screen remains visible. Restore Study Restores a saved study to Ascribe™. Restore an ACM Restores a saved automated coding model to Ascribe™. Exit Ends the Ascribe™ session and closes the application. This is the same process as closing the application by clicking the close button 431 PrintDoc01 020813 ("x") in the upper right corner of the window's title bar. Exiting the application will end the current "session" in the same fashion as logging off the Ascribe™ Web version. View Refresh This option is functionally equivalent to pressing the F5 key or clicking the main toolbar Refresh button. (For more information, see Refresh.) Show Navigator Toggles the display of the left side navigation pane. Show Toolbar Toggles the display of the main toolbar at the top of the Ascribe™ window. Show FilterBar Toggles the display of the studies and questions Filter buttons at the bottom of the navigation pane. Welcome View Navigates to the Welcome page of the Desktop. Studies Status Navigates to the Studies Status page. Jobs Navigates to the Jobs page. Loads Navigates to the Loads page. Samples Navigates to the Samples page, which lists all samples on this site. Tools Codebook Merge/Map Tool Opens the Codebook Merge/Map tool. Study Merge Tool Opens the Study Merge tool. Study Import Tool Opens the Study Import tool. Options Displays the Options Dialog Form. Window Close All Tabs Closes all tabbed windows other than the Home View Tab. Help Help Topics Displays the help system. About Provides information about the current version of Ascribe™ Desktop. Toolbar The Toolbar provides quick "one-click" access to frequently used functions. The options are described in the table below. Tool Home 432 Function Restores the display to the "Home" view. The Home View sets the Navigator pane to its top levels and the Data pane to either the "Welcome" screen or the "Studies List" based upon the users settings in the Options dialog (see Options). Clicking this button is functionally equivalent to pressing the F2 key. Ascribe™ Desktop Refresh Same as the Refresh menu option above. (For more information, see Refresh.) Back Returns the Home View to the previous screen. If a tab other than the Home View is active, (e.g. currently visible), the current Home View tab becomes active displaying the last active Home View screen. This is functionally equivalent to pressing the F8 key. Find a Study Executes a search based upon the values entered into the "Find a Study" text box. Click this button or press the Enter key after entering text into the "find" box. The results of the "Find" are displayed in the Home View tab as a list of studies. The search will check for and display any studies whose study id or study name contain the text. Note that the text is treated as a regular expression and as such, can provide the ability for powerful search expressions. Find a Question Executes a search based upon the values entered into the "Find a Question" text box. Click this button or press the Enter key after entering text into the "find" box. The results of the "Find" are displayed in the Home View tab as a list of questions. The search will check for and display any question whose question id or question label contain the text. Note that the text is treated as a regular expression and as such, can provide the ability for powerful search expressions. The Find a Question option is only available when the display context provides a list of questions to be searched. This basically means a questions list must be displayed in the Home View tab or a study must be selected in the Home View studies list, the Navigator Tree, or the Workspace. Help Displays the Ascribe™ Desktop Help topics Navigation Pane The left half of the main window contains the Navigation pane. It displays three components: Navigator Tree, Workspace, and Filter Bar. Navigator Tree The Navigator Tree displays a Tree structure with four top-level items or nodes: Studies Account Codebooks Automated Coding Models "My Folders" Expand a node by either clicking the "+" on its left or double-clicking the node itself. Collapse a node by clicking the "-" on its left or double-clicking on it. Nodes expand to display the following items: 433 PrintDoc01 020813 Node Expands to Studies A list of study nodes A study A questions node and a codebooks node Questions A list of question nodes Codebooks A list of codebook nodes A question Nothing Account Codebooks A list of codebook nodes A codebook Nothing Automated Coding Models A list of models A model A training examples node and a coding questions node Training Examples A list of Languages; there must be at least one Language to hold the training examples Coding Questions A list of coding questions "My Folders" A list of "My Folders" study folders or "My Folders" codebooks folders A "My Folders" folder A list of selected study nodes or codebook nodes Workspace and Filter Bar The Workspace tab of the Navigation Pane provides a place to save frequently used items or items currently being worked on. See Using the Workspace for more information. The Filter bar is located at the bottom of the Navigation Pane. The buttons on this bar allow users to filter which studies and questions will be displayed in the navigator Tree and the Home View tables. The two available filters are described in the table below. Filter Functionality Study Filter Toggles the Study Filter selection options. Click initially to display a list of study status options available. Check the box beside each status to be displayed. Click the button again to close the displayed list. The selections are saved and will remain in effect until changed again. The Navigator pane and Data pane are immediately updated to reflect the new filtering. Question Filter Toggles the Question Filter selection options. Click initially to display a list of question types available. Check the box beside each type to be displayed. Click the button again to close the displayed list. The selections are saved and will remain in effect 434 Ascribe™ Desktop until changed again. The Navigator pane and Data pane are immediately updated to reflect the new filtering. Data Pane The right half of the main window contains the Data pane. It will display different information based upon which item (node) is selected in the left side Navigation pane. The Data pane will format the information in one of two ways: a folder view or a table view. The selected left side Navigation pane item and its corresponding Data pane display are described in the table below. Click this node… See this in the Data Pane… Studies A table of study properties listing each study meeting the filter criteria. A study A folder view displaying a count of questions in the study and a count of responses. Questions A table of question properties listing each question meeting the filter criteria. A question A folder view displaying a count of responses and codes in the codebook. Codebooks A table of codebook properties for each codebook in the study. A codebook A folder view displaying a count of responses and a count of codes in the codebook. Account Codebooks A table of codebook properties for each account codebook. An account codebook A folder view displaying a count of questions that share the codebook and a count of codes in the codebook. Automated Coding Models A table of model properties for each model on this site. A model A folder view displaying the number of Languages and coding questions for the model and the training date and time Training Examples A table of properties for the Languages in the model. A language A folder view of the number of responses/training examples in the language and the number of codes in the codebook. Coding Questions A table of properties for the coding questions in the model. A coding question A folder view of the number of responses in the coding question and the number of codes in the codebook. "My Folders" Nothing. A "My Folders" folder A folder view for a study or codebook as described above. General Navigation Concepts In the main display of Ascribe™ Desktop, here are the items that can be navigated to: a study, a question, a codebook, an automated coding model, training examples, and coding questions. All other items to be located will derive from one of these. Other items such as 435 PrintDoc01 020813 codes, responses, respondents, loads, etc., can be navigated to by clicking on one of these items or through menu options. The primary items can be located and selected either in the Navigator Tree, the Workspace, or a Home View tab list. Questions and studies also can be quickly located using the "Find" options of the Main toolbar discussed above. Generally, to perform some task for an item, the user should right-click that item. The resulting menus that display after the right-click are called context menus because the content of the menu and the availability of options will vary based upon where and what is being clicked. To speed up navigation to common tasks, use the double-click options in the Options Dialog Form to set the double-click action to go to the activity you normally would perform for the item. Study Context Menu Options Here are the options when you right-click a study: Option Description Remove from Folder This menu option is available only when the study is in a "My Folders" folder. It simply removes the study from its containing folder. It does not delete the study from the Ascribe™ account. Refresh Causes the study data to be reread from the database. Create Shortcut in Workspace Create a shortcut to the question in the Workspace. Questions Displays the study's question list in the Home View tab. Respondents Displays the study's respondents list in the Home View tab. Codebooks Displays the codebooks used by questions in the study in the Home View tab. Only codebooks that contain codes are displayed. Loads Displays the Loads page in the Home View tab. Edit... Opens the Study Edit screen. Delete Selected Studies Removes the selected studies from the Ascribe™ account after acceptance in the confirmation dialog that is displayed after selecting the option. This is a permanent deletion of all questions, responses, and coding related to the study. Load Responses / Process File Opens a dialog box where you can load data. See Load Data to a Study for more information; this section explains how to load data on Ascribe™ Web, but it also applies to Desktop. Save Study Opens a dialog box where you can save a study. See Save a Study for more information; this section explains how to save a study on Ascribe™ Web, but it also applies to Desktop. Merge Study Opens a dialog box where you can merge a study. See Merge Study for more information; this section explains how to merge a study on Ascribe™ Web, but it also applies to Desktop. 436 Ascribe™ Desktop Question Context Menu Options Here are the options when you right-click a question: Option Description Refresh Causes the question data to be reread from the database. Creates Shortcut in Workspace Create a shortcut to the question in the Workspace. Edit... Opens the Question Properties Editor dialog. This option provides four selections of items to be deleted: Delete Selected Questions... Codes Only - removes all codes from the responses of the question Quality Codes Only - remove all the quality codes from the responses of a question All Responses - removes all responses from the question. Question - removes the question from the study. Load Details Displays the Load Details for a question. Codebook Editor Opens the Codebook Editor for the question's codebook. Create Codebook Copy as Account Codebook Copies the question's codebook including all codes to a new codebook. The newly created codebook is designated as am account codebook with the text provided by the user. Presents the coding options available for the question. Currently the following coding options are available in Ascribe Desktop: Code Text AutoCode - Opens the Text AutoCoder for the question's responses. Multiple Responses – Opens the Multiple Responses Coding window. This option presents the various ACM options available for this question: ACM Extract Response Sample... Create an ACM Find ACM Start ACM coding job Review ACM coding This option displays the Create Quick Sample dialog. You can create a sample from a question. Codebook Context Menu Options Here are the options when you right-click a codebook: Option Description Remove from Folder This menu option is available only when the codebook is in a "My Folders" folder. It simply removes the codebook from its containing folder. It does not delete the codebook from the Ascribe™ account. 437 PrintDoc01 020813 Refresh Causes the codebook data to be reread from the database. Create Shortcut in Workspace Create a shortcut to the question in the Workspace. Codebook Editor Opens the Codebook Editor for the codebook. Label... Prompts the user to enter a codebook ID and designates the codebook as an account codebook. Unlabel… Removes the account codebook designation from the codebook. If the account codebook is not presently in use by a question or a model, the codebook will be deleted when you remove the account codebook designation. Create Codebook Copy as Account Codebook... Copies the codebook including all codes to a new codebook. The newly-created codebook is designated as an account codebook with the text provided by the user. Delete Selected Codebook Removes the codebook from the account, if possible. The removal is permitted only if the codebook is unused. This option presents these choices: ACM Extract Response Sample... Create an ACM Find ACM This option displays the Create Quick Sample dialog. You can create a sample from a question. Automated Coding Model Context Menu Options Here are the options when you right-click a model: Option Description Refresh Causes the model data to be reread from the database. Create Shortcut in Workspace Create a shortcut to the model in the Workspace. Coding Questions Displays the list of coding questions in the Home View tab. Training Example Editor Navigates to the Training Example Editor for this model. Codebook Editor Opens the Codebook Editor. Edit Automated Coding Model... Opens the Edit ACM dialog where you can change the properties of the model. Delete Selected Automated Coding Models Removes the selected models from the Ascribe™ account after acceptance in the confirmation dialog that is displayed after selecting the option Validate... Checks the setup of the model for errors. Train... Starts the training job for the model. Approve... Approves the model so you can use it to code questions. Unapprove... 438 When you unapprove a model, it becomes unavailable to use Ascribe™ Desktop for coding. Unapprove a model if you no longer want it to be used for coding. You must also unapprove a model if you want to modify it, or if you want to retrain it by running a training job again. When you unapprove a model, you also discard any training results. To use the model for coding again, you must run the training job, approve the model, and then run any desired coding jobs. Save Automated Coding Model Extract Response Sample... Saves the model as a file on your computer. Creates a sample from the training examples in the model. See How to Create a Sample for more information. Options and Customization Navigate: Tools/Options Ascribe™ Desktop allows for considerable customization of display and behavior. The starting point for such customization is the Options Form. You can open the Options Dialog Form by navigating to the Tools Menu option on the main Toolbar then selecting "Options". You also can customize the Table View in the Data Pane and customize Code Usage Views. Options Dialog Form The options form consists of three tabs: Environment - options related to navigation, mouse behavior, and file locations Display Options - options related to font and color display User - options related to the behavior of Ascribe™ Desktop Account - account options related to study and question setup, loading data, etc. The following table describes the options for Environment: Environment Options Value General options Open New Window If Needed If this option is checked, each time a new function such as the Codebook Editor or a view is requested, Ascribe™ will open it in a new tabbed window. Otherwise, only one tabbed window other than the Home View will ever be opened, and it will be re-used as functions and view change. When switching to any tabbed window other than the Home View, the Toolbar will be Hide Main Toolbar When Not In Home View hidden if this option is checked. Returning to the Home View tab will display it again. Hide Navigation Pane When Not In Home When switching to any tabbed window other 439 PrintDoc01 020813 View than the Home View, the Navigation pane will be hidden if this option is checked. Returning to the Home View tab will display it again. Synchronize Navigator Pane with Home View Lists When checked, the Navigator pane will change to match whatever is clicked in the Home View List. When the box is not checked, the Navigator is static and does not change when you click an item in the Home View List and navigate to another area. At Startup Show... This option determines where the Navigator Tree or the Workspace is visible in the Navigation Pane upon start up of Ascribe™ Desktop. When I Click The Home Button Show This option determines whether the Home View studies list or the Welcome page is displayed when the "Home" or the F2 key is pressed. Mouse Click Options When A Question Is Double-Clicked This setting determines the action taken when a question is double-clicked either in the Workspace, the Navigator Tree, or a Home View questions list. When A Study Is Double-Clicked This setting determines the action taken when a study is double-clicked either in the Workspace, the Navigator Tree, or a Home View studies list. When A Codebook Is Double-Clicked This setting determines the action taken when a codebook is double-clicked either in the Workspace, the Navigator Tree, or a Home View codebooks list. When An Automated Coding Model Is Double-Clicked This setting determines the action taken when a model is double-clicked either in the Workspace, the Navigator Tree, or a Home View models list. Default Directories Save Workspace Files Here This option specifies the location where files dragged to or copied to the Workspace should be saved. If blank (the default), the user is prompted each time such action occurs. If an entry is made here, it must be a valid path to a location accessible by the user. Save Job Files Here This option specifies the location where files downloaded from previously run jobs should be saved. If blank (the default), the user is prompted each time such action occurs. If an entry is made here, it must be a valid path to 440 Ascribe™ Desktop a location accessible by the user. The following table describes the Display Options: Display Options Value Fonts (You cannot change the fonts of menus or toolbar buttons.) Main Display Font The Main Display Font defaults to Verdana 9.75pt Regular. You can change this font. Codebook 'Net' Font The Codebook 'Net' Font defaults to Verdana 9.75pt Bold. You can change this font. Codebook 'Code' Font The Codebook 'Code' Font defaults to Verdana 9.75pt Regular. You can change this font Restore Defaults If you click this button, the fonts will return to the defaults listed above. Colors Table 'Stripe' Color Main Background Color Restore Defaults You can change the Table 'Stripe' Color. It defaults to a light peach color You can change the Main Background Color. It defaults to white. If you click this button, the colors will return to the defaults listed above. The following table describes the options for User: User Options Value General options Disable Spellcheck for Codebook Editing If checked, this option will disable the use of spell checking in the Codebook Editor. Normally this will be selected if 1) MS Word does not exist on the user machine or 2) spell check is not desired in order to improve performance while in the codebook creation stage. Keep Filters Between Sessions This option, if checked, will keep response filters between sessions of Ascribe™. Workspace options Remove Items Older Than This option allows the user to specify how long items will remain in the Workspace. If set to "Do not Remove," items will remain until the user manually removes them. 441 PrintDoc01 020813 Otherwise, Ascribe™ will remove items as their age reaches the specified time based upon the date the item was added to the Workspace. At Start Up, Order By This option allows the user to specify where items will appear in the Workspace. If set to "Do not change order," items will remain in the locations specified by the user by manually moving them. Otherwise, Ascribe™ will order the items according to the criteria specified and position them accordingly. At Startup, Preload Data For Recent Items If checked, Ascribe™ will attempt to download the data related to each Workspace item from the database in the background. This can improve performance and avoid delays the first time such data is needed. For a study, questions are downloaded. For a question, responses are downloaded, For codebooks, the codes are downloaded. Note: To avoid potentially long processes, only data for the five most recently added Workspace items are downloaded. Save Active Window To Workspace At Close If checked, Ascribe™ will save the active window (other than the Home View), to the Workspace when the application closes. Set Background Color To The Workspace background color will be set to the same as the computer's background color or to a color specified. Code Usage Display Options... This button opens a dialog form that allows the user to establish how code usage information is displayed. (For more information, see Customize Code Usage Views.) The following table describes the options for Account: Account Options Value Setup Options File Format See Data Tab (Output Format) Layout See Data Tab (Output Format) Card Number Format See Data Tab (Output Format) Respondent ID Column See Data Tab (Output Format) Respondent ID Columns See Data Tab (Output Format) Card Number Column See Data Tab (Output Format) Card Number Columns See Data Tab (Output Format) 442 Ascribe™ Desktop Question/Codebook Options Allow Duplicates In Codebook If checked, duplicate code output values will be allowed in a codebook. Use Codebook Colors In Reports If checked, codes will be displayed in reports with their assigned color. Otherwise, codes will be displayed with black text. Question Task – Translate See Tasks Tab Question Task - Transcribe Include In Crosstab Question Task - Code Code Numbering Spread See Tasks Tab See Tasks Tab See Tasks Tab Establishes the increment between code values when renumbering a codebook. Load Options Study ID Delimiter (in filename) The options in this section are used when Ascribe™ processes one of the default load file types received via our FTP site. See the Load File Types page for a complete description of each option. Auto-Code Closed And Other Specify Questions See Load File Types Auto-Code Text Responses See Load File Types Add Missing Questions See Load File Types Merge Codebooks See Load File Types Miscellaneous Options Track 'Non-Study' Time If checked, Ascribe™ Desktop will add entries to track activity time for those activities not specifically related to a particular study. Typically, these activities would involve viewing information such as jobs lists, loads views, etc. Customize the Table View in the Data Pane Ascribe™ Desktop will display information about studies, questions, responses, and codebooks in a table view with columns representing the properties of a particular item (e.g., Question ID, Question Label, etc.). In any of these table views, you can change the order of the columns by selecting the column header and dragging (move the mouse while holding the left mouse button down) to the desired position. You can also select which columns should be displayed by right-clicking the upper left most cell of the table. This action will display a menu of choices to allow for choosing columns, setting column widths, etc. Any customization you make will be saved and restored for future sessions. 443 PrintDoc01 020813 Note: When Ascribe™ Desktop is opened for the first time, a default set of columns is displayed for each type of item (i.e., study, question, etc.) These will be the same columns displayed in Ascribe™ Web for the similar page (studies or questions). Customize Code Usage Views At several places throughout Ascribe™ Desktop, the user has the option of displaying counts of code usage. That is, Ascribe™ will show how many responses a code has been applied to. There are several options available that will control how such counts are presented. In the User tab of the Options Dialog form available from the Tools menu, you can click the "Code Usage Display Options" button to display these options: Options Value This selection determines how the code count values will be presented. The choices and their appearance are listed below. Display Mode Counts Only -- Code Description [10] Percentages Only -- Code Description [8.0%] Counts and Percentages -- Code Description [10/8.0%] Display Zero Values As If checked, zero count values will be display the text provided in the adjacent box. If not checked, count values of zero will display 0 (zero). Display Excluded Codes As If checked, codes marked as excluded from counts in the Codebook Editor will display the text provided in the adjacent box. If not checked, blanks will be displayed. Basis For Percentage This selection will determine the value to be used as the denominator when calculating code usage percentages. Using the Workspace The Workspace is one of the two tabs in the Navigation pane, along with the Navigator Tree. It is intended to serve as a convenient place to hold Ascribe™ items temporarily between sessions so they can be quickly retrieved in subsequent sessions. It operates much like the desktop space on the computer. The Workspace stores a shortcut to the actual item, not the actual item. When a shortcut is added and removed from the Workspace, the underlying item is not affected. The Workspace can hold 7 different types of shortcuts: Shortcut Study Description A shortcut to an Ascribe™ study. Question A shortcut to an Ascribe™ question. Codebook A shortcut to an Ascribe™ codebook. 444 Ascribe™ Desktop View Link A shortcut to an Ascribe™ function or view. For example, the Codebook Editor function or the Study Status View can be saved here. A shortcut to a Web URL. File A shortcut to a file located on the local computer or a network location. Container A 'folder' that can hold one or more of the other shortcuts. Workspace Context Menu Options Right-click an open space in the Workspace to access its context menu. The available options in the context menu are shown below. Options Description New Container... Adds a new container to the Workspace. The dialog form that appears provides an opportunity to enter a container label and text that is displayed when the mouse hovers over the shortcut. Redraw Workspace Redraws the items in the Workspace and places the items in evenly spaced columns and rows based upon the size of the Workspace at the time of the redraw. Organizes the Workspace items based upon the selected criteria and then redraws per above. The available criteria are: Arrange Workspace By Date Added - arranges the items based upon the date they were added to the Workspace with the most recent item displayed first. By Type - arranges based upon type (i.e., all studies listed together.) Remove All Shortcuts From Workspace... Removes all items currently in the Workspace. The user will be prompted to complete this operation. Paste Pastes either links or files into the Workspace which have been copied elsewhere. Add, Remove, or Change Workspace Items To add a study, question, or codebook to the Workspace, choose the "Create shortcut in Workspace" option from its context menu. Optionally, drag selected items from a table displayed in the Home View tab of the Data Pane. To add a view or function window to the Workspace, choose the "Create shortcut in Workspace" option from the tabbed window's context menu. (This is the menu that appears when the tab itself is right-clicked). 445 PrintDoc01 020813 To add a new link to the Workspace, copy the link text and then paste into the Workspace using the "Paste" option of the Workspace context menu. To add a new file to the Workspace, copy the file from Windows Explorer display and then use the "Paste" option of the Workspace context menu. Or, drag the file from Widows Explorer or an email attachment and drop onto the Workspace. To add a new container to the Workspace, right-click an open area of the Workspace and then select "New Container" from the available options. To remove an individual item from the Workspace, select the "Remove Workspace shortcut" option from its context menu. Workspace items can be added to a container by dragging them onto the container. To remove individual Workspace items, first select the container. Then select the item in the table of Workspace container items in the Data Pane and choose the "Remove from Container" option in the item's context menu. All items in a container can be quickly removed by selecting the "Empty to Workspace" menu option of the container's context menu. This process will add each item in the container to the Workspace. Then, items can be removed individually as described above. Understanding Workspace Context Question, codebook and view shortcuts in the Workspace can have a context associated with them. A context means that Ascribe™ knows what should be done when a shortcut is double-clicked. For example, a question might have a Codebook Editor context, an autocoding context, a multiple response coding context, etc. Similarly, a codebook might have a Codebook Editor Context, an ACM trainings context, a Codebook Merge context, etc. View shortcuts will always have a context associated with them. For views, a context represents what the view is displaying. It might be an ACM trainings table, a Study Status table, a jobs table, etc. When a Workspace item has a context, double-clicking will return the user to that context. When an item does not have a context, the double-click result respects the setting in the Options dialog for that item type. When items are added to the Workspace by dragging or by right-clicking the item in the Navigator Tree or a Home View table, there is no context associated with the item. To add a Workspace item with a context, or to add a context to an existing Workspace item, the user must select "Create shortcut in Workspace" from the context menu of the tabbed window in the Data Pane. This menu is available by right-clicking the tab itself. When an item has a context associated with it, this is indicated by the symbol in the lower left corner of the shortcut image. Hovering over this symbol will display the context associated with the item. Using 'My Folders' Ascribe™ Desktop introduces the concept of "My Folders". These are folders or nodes in the Navigator Tree that can contain studies or codebooks. The Navigator Tree includes a top level folder named "My Folders" which is a permanent folder in the Tree. All other "My Folders" folders are children of this folder. The intent of the "My Folders" group is to allow 446 Ascribe™ Desktop easy navigation to frequently used items or as a way to organize studies and codebooks in a different manner than the primary navigation tree. A study or codebook can appear in multiple folders. Codebooks Folders To create a new codebook folder, right-click the "My Folders" node and select "New Folder...." Next select "Codebooks Folder...." Finally, enter a descriptive label for this folder in the text box that appears. To add codebooks to the folder, simply drag a codebook onto the folder from a table view of codebooks in the Home View tab or a question from a table view of questions in the Home View tab. To remove a codebook from the folder, right-click the item and then select the "Remove from folder" menu option. Note: This process simply removes the codebook from the containing folder. It does not delete the codebook from Ascribe™. The codebook can be added back again in the manner described above. Studies Folders To create a new studies folder, right-click the "My Folders" node and select "New Folder...." Next select "Studies Folder...." The studies folders provide a bit more flexibility in maintaining the folders contents than the codebooks folder does. As a result, a dialog box will appear with several options as described below: Folder Type Description Custom This option creates a folder that works exactly like a codebooks folder. It is a folder whose contents are static and maintained by the user by dragging studies into it and removing items from the context menu option. Studies Assigned To This option creates a folder that will contain all the studies assigned to the supervisor selected in the adjacent drop-down list of supervisors. As studies are added, removed, or changed to reflect a different supervisor, the contents of this folder will be modified. As a result, studies cannot be "manually" added or removed from this folder. Studies For Client This option creates a folder that will contain all the studies assigned to a particular client. As studies are added, removed, or possibly changed to reflect a different client, the contents of this folder will be modified. As a result, studies cannot be "manually" added or removed from this folder. Studies Active In Last This option creates a folder that will contain all the studies that have been "active" within the designated number of days. "Active" is defined as any activity for which time accounting information is tracked. This typically includes coding, loading data, and editing the codebook. The contents of this folder are automatically maintained. As a result, studies cannot be "manually" added or removed from this folder. 447 PrintDoc01 020813 After you select a folder type, enter a name in the folder label field. Note The “My Folders” settings and content are saved by user. Thus, the folders will appear with the same content for a given user regardless of the computer on which Ascribe™ Desktop is accessed. Account Codebooks in Desktop Account codebooks are intended to allow you to find codebooks, especially those to be reused, by codebook ID rather than via a study or question. A special folder is included in the Navigator Tree to display a list of these codebooks. Only codebooks designated as "account codebooks" will be included in the account codebooks folder. Creating Account Codebooks Creating account codebooks is very simple. Right-click a codebook in either the Navigator Tree, the Workspace or a Home View table of codebooks. If the codebook is currently not an account codebook, the menu option "Label..." will be available. Choose that option and then enter the desired label in the text box that appears. Clicking OK will designate the codebook as an account codebook. Clicking cancel will not. Codebooks can also be designated as an account codebook by checking the "Account Codebook box in the Codebook Editor. In addition, any codebook can be copied as an account codebook by choosing the "Create copy of codebook as account codebook" option from a question or codebook context menu. Marking Codebooks as Unlabeled It is very easy to "unlabel" a codebook. Right-click a codebook in either the Navigator Tree, the Workspace, or a Home View table of codebooks. If the codebook is currently marked as labeled, the menu option "Unlabel..." will be available. Choose that option and then confirm your choice in the dialog that displays. Clicking OK will mark the codebook as unlabeled. Clicking cancel will not. Codebooks can also be marked as unlabeled by un-checking the "Account Codebook" box in the Codebook Editor. Note Removing the account codebook designation affects its inclusion in the list of account codebooks in the Navigator Tree. It can also affect its existence on the account. If the account codebook is not used by a question or model on the account and you remove the account codebook designation, the codebook will no longer be accessible. If an account codebook is not assigned to a question or a model, removing the designation is equivalent to deleting it. Desktop Codebook Editor 448 Ascribe™ Desktop The Codebook Editor provides functions to add, edit, remove or rearrange codes in a codebook. In addition, response data and word lists can be accessed to facilitate building codebooks. The Codebook Editor screen has two display areas: the Codebook Properties Editor and the Codes Editor. The Codebook Properties Editor is displayed in the top portion of the window and allows editing of attributes associated with the codebook itself. Below the Codebook Properties Editor is the Codes Editor. The Codes Editor allows for adding, editing and removing codes from the codebook as well as re-ordering codes. The Codes Editor also hosts the toolbar for quick access to all the functionality of the Codebook Editor. Codebook Properties Editor You can change the attributes of a codebook with the Codebook Properties Editor. Changes to these properties are updated immediately to the database as soon as the cursor moves off the field. Property Description Codebook ID A unique identifier for the codebook. Use this identifier to provide a means to access the codebook from external sources. It should be unique within all codebooks. An error will occur if you try to set this ID to a value that already exists for another codebook. When this box is checked, the codebook will be designated as an account codebook. This allows it to be displayed in the Navigator Tree. As such, it can be located without being associated with a question. Account Codebook NOTE: Removing the account codebook designation affects its inclusion in the list of account codebooks in the Navigator Tree. It can also affect its existence on the account. If the account codebook is not used by a question or model on the account and you remove the account codebook designation, the codebook will no longer be accessible. If an account codebook is not assigned to a question or a model, unchecking this box is equivalent to deleting it. (See Account Codebooks in Desktop.) Spread Between Codes This setting determines how the default value for inputID and outputID are calculated. You may specify any number greater than zero. When the No Duplicate OutputID's check box is cleared, the default value for a new code is determined by inspection of all of the siblings of the code. The default value will be the first whole multiple of Spread Between Codes greater than the outputID of all siblings. Only siblings with numeric outputID's are considered. Description A description of the codebook. You can use this description to provide information about the codebook, which is particularly useful for shared codebooks. No Duplicate OutputIDs If this check box is selected, Ascribe™ will guarantee that the outputID of each code in the codebook is unique. When you 449 PrintDoc01 020813 assign a new value to the outputID of a code, Ascribe™ will change the outputID if necessary to ensure that it is unique within the codebook. If Ascribe™ changes the outputID, it will assign a large number. This is a visual indication that you need to correct the outputID. This check box is disabled if the codebook contains duplicate codes. To enable the check box, you must first ensure that all codes have unique outputID's. Note that uniqueness applies to all codes in the codebook, including nets. The Codebook Properties Editor can be displayed or hidden by using the applicable toolbar button as described in the Editor Toolbar section. Hiding it can be useful if more screen area is needed for the Codes Editor section. Codes Editor The Codes Editor is in the lower portion of the Codebook Editor window. It contains the majority of the functionality and has most of the screen area. The Codes Editor has four functional tabs located at the bottom of the window. Editor Tab - This is the main tab for working with codes. Usage Tab- This tab provides a display of all the questions that use the codebook being edited. This is informational only. Questions cannot be added or removed here. Duplicates List Tab - This tab provides a display of all the codes that have the same output ID. You can edit the code properties from this screen. Code Builder Tab - This tab provides a display of responses and is useful when building a codebook based upon review of the response data. Editor Tab The Editor tab is where most of the codebook editing activity occurs. Codes can be manipulated in one of four ways: the main toolbar, the context menu options, dragging and dropping, and the Code Properties Window. Editor Toolbar The main toolbar has a number of useful functions as well as options to change the views and screen layout. Function Description View Changes the view between the Edit view and the Counts view. The Counts view displays the count of code usage for each question that shares the codebook. The Edit view displays the codebook and allows for the various manipulations described here. There are three choices available with this toolbar option. Remove Items Remove all Items - Choosing this option will attempt to remove all items from the codebook. If any codes in the codebook have been applied to responses, the process will fail with no codes being deleted. Code usage can be determined by switching to the Count view as described above. Remove all Nets - Choosing this option will remove all nets 450 Ascribe™ Desktop from the codebook. The remaining codes will all be top level codes. Remove all Codes - Choosing this option will attempt to remove all codes from the codebook leaving the net structure in place. If any codes in the codebook have been applied to responses, the process will fail with no codes being deleted. Code usage can be determined by switching to the Count view as described above. Copy This toolbar option will copy the existing codebook to a new labeled codebook. After choosing this option, the user is presented with a dialog box requesting the new label. Upon completion of the copy, the newly copied codebook will appear in the Labeled Codebooks section of the Navigator Tree. It is not assigned to any question at the time it is copied. Renumber This option will renumber the entire codebook based upon the selected renumber options. For more information, see Codebook Renumber Options. Sort This option will sort the entire codebook based upon the selected sort options. For more information, see Codebook Sort Options. Spell Check This option begins a spell check of the entire codebook. In order to use spell checking, the computer must have at least Word 2003 installed. The spell check process makes use of the Word spell check dictionaries. As a result, custom dictionaries can be added for use by the spell checker. For more information, see the Microsoft Word documentation and Spell Check Options. Choosing the toolbar option will toggle the expanded state of the Codes Editor. In other words, if expanded, the codebook Expand/Collapse will be collapsed to just nets. If currently collapsed, each net and sub-net will be expanded. This button enables or ends the use of Grid Edit Mode. Grid editing allows for Excel®-like editing of individual cells in the Codes Editor table. Grid Edit NOTE: With Grid Edit mode off, a cell can be edited by double-clicking in the cell. Edit mode ends when moving to another cell. With Grid Edit Mode On, cells automatically are editable as you move into the cell. Clicking this toolbar button will offer four options for changing the appearance of the Codebook Editor. Display Options Codebook Properties - If checked, the Codebook Properties Editor described above will be displayed. Otherwise, it will be hidden. Code Properties - If checked, the Code Properties area of the Codes Editor will be displayed. This is an area of the screen below the table of codes that provides for additional information about selected codes. For more information, see Code Properties Display. 451 PrintDoc01 020813 Code Usage - If checked, the Codes Usage area of the Codes Editor will be displayed. This display occupies the same screen area as the Code Properties area. After you select a code, all of the responses that have this code applied will display. Codebook Notes - If this option is checked, a "floating" window displays, and you can add notes to the codebook. Existing notes will be displayed. Notes are saved in the database as part of the permanent codebook information. Newly added notes are saved each time the enter key is pressed and when the window is closed. The floating window can be set to always stay on top by selecting that option from the windows context menu (right-click the top title bar of the notes window to display the context menu.) Editor Context Menu Options The Editor context menu provides options for inserting and deleting codes and nets as well as providing the ability to use the sorting, renumbering, and spell checking options on only the portions of the codebook selected. To access the menu, right-click a code or select multiple codes and then right-click them. Option Description Refresh Reloads data from the database and updates the screen for the entire codebook. This option is here as a convenience and provides the same functionality as the Refresh button on the Ascribe™ Desktop's main toolbar. Expand All Expands each net, which means to display the codes for each net. Collapse All Collapses all nets, which means to hide the codes belonging to a net. Only the top level nets display. New Net Inserts a new net as a sibling (at the same level) as the selected code. The Editor’s behavior with this function differs depending on how many codes are selected at the time the option is executed. If one code is selected, an empty net is created as a sibling of the selected code. If more than one code is selected, the new net is inserted, and all the selected codes are made children of the newly inserted net. Insert Codes Inserts a new code as a sibling of the selected code. To change the description of the newly created code, double-click the description cell. NOTE: The Editor will insert as many new codes as there are selected codes in the Editor when the option is executed. Paste codes... This option allows the user to paste codes and nets into the codebook from another source. The codes will be inserted as a sibling of the selected code (or the first selected code if more than one is selected at the time the option is executed). Prior to the actual update of the codebook for the new items, a view of the new items will be presented. The paste operation can then be either accepted or cancelled. (See Paste codes 452 Ascribe™ Desktop into a Codebook for more information.) Edit... Displays the Code Properties Edit screen. Delete Selected Codes Removes the selected codes or nets from the codebook. If a net selected for removal has child codes, an error will be indicated and the process terminated with no deletions occurring. Similarly, if a code selected for removal has been applied (either to a response or as a quality code), an error will be indicated and the process terminated with no deletions occurring. Change Selected Nets To Codes If a net should be changed to a code, use this option. Change Selected Codes To Nets If a code should be changed to a net, use this option. Decrease Indent The decrease indent option will shift all selected codebook items up one level. This means that the selected codebook items will no longer be children of a net; rather, they become siblings of the net (i.e., they will be at the same level as the net.) Increase Indent The option will make all selected codebook items children of the first net above the selected items in the codebook. Renumber Codebook... This option allows for applying the renumbering options to selected codebook items rather than the entire codebook. For more information, see Codebook Renumber Options. Sort Codebook... This option will apply the sort options to the selected codebook items. For more information, see Codebook Sort Options. Spell Check Codebook... This option begins a spell check of the selected codebook items rather than the entire codebook. In order to use spell checking, the computer must have at least Microsoft Word 2003 installed. The spell checking process makes use of the Word spell check dictionaries. As a result, custom dictionaries can be added for use by the spell checker. See the Microsoft Word documentation for more information. For more information, see Working with Spell Check Results. Editor Drag/Drop Options The drag/drop operations provide the ability to move codebook items within the current codebook or between two different codebooks. If a group of selected codebook items are "dropped" onto a target code, the items will be moved to a position immediately below the target code as siblings of that code. If a group of selected codebook items are dropped onto a target net, the dropped items will become children of that net. 453 PrintDoc01 020813 To move selected codebook items to a position immediately below a net as siblings rather than children, position the mouse in the area between the net and the item below it while continuing to hold the mouse button down for dragging. Code Properties Window The Code Properties Window displays the editable properties of a code and offers the ability to change them just as the Code Properties Edit screen does. Unlike that screen however, the Editor Code Properties Display remains on the screen until the option is selected to hide it. This makes it convenient to use when the properties of many codebook items are being changed. The display of code properties will change as different codes are selected in the codes table. When more than one codebook item is selected in the codebook items table, the Code Properties Display changes to show only those properties that can be altered for more than one code at a time. Specifically, code attributes, input ID, and color can be changed. To show or hide the Code Properties Display, click the Display options button in the Editor Tab toolbar and then check or uncheck the "Code Properties Display" option. Usage Tab The Usage tab displays the questions that use the codebook which is being edited. Questions can not be added or removed here. The screen displays the study ID, study name, question ID, and question label. Duplicate List Tab The Duplicate List tab displays all codes that have the same output ID. If you right-click a code, you can edit the code properties and change the output ID. Or you can go to the Editor tab and change the output ID there. Code Builder Tab The Code Builder tab allows you to view the response data associated with questions that share this codebook in several ways to help you create new codes. As with the Code Editor Tab, the functionality and display options are executed by either toolbar selections or context menu options. Code Builder Toolbar The toolbar of the Code Builder Tab has a number of useful functions as well as options to change the views and screen layout. Option Description Click this button to change which responses are displayed. Response View - Choose this view to display all responses from the selected response sources. View Selected Sample - Choose this option to display a random sampling of responses from the selected response sources. Sample options included either a percentage of responses or a specific number of responses. Responses Source Click this button to display a list of questions that share the 454 Ascribe™ Desktop codebook being edited. Responses displayed based upon the view option above will come from the selected questions. Search Expression Enter a valid regular expression in this box and then press the Enter key or click the search icon to display a list of responses that match the search expression. Word List Click this button to toggle (display or hide) the Word List. When the Word List is displayed, the tab is split in to two panes. The left side displays a count of each word in the selected response set. The right side displays responses that contain the selected word on in the left pane. Work Pad This button toggles (displays or hides) the Work Pad. See Using the Code Builder Work Pad for more information. Code Builder Context Menu Options The Code Builder context menu provides different options based upon whether the Word List is displayed or not. Option Description Word List Options Add Selection To Ignored Item List This option will add the selected words to the ignored word list. All words in the ignored word list will be excluded from display in the Word List. This ignored word list will remain set between sessions until changed by the user. Make Codes... Each of the selected words will be added as codes to the current codebook if this option is selected. Provides a dialog box to edit the Word List options: Delimiters - A list of individual characters that will be used to delimit (separate word boundaries). Enter these characters as a single string such as .!?" Characters to Ignore - A list of characters that will be stripped from the display in the Word List display. Enter these characters as a single string such as .!?#$%&*" Edit List Options... Words to Ignore - A list of words which will be stripped from the display in the Word List display. Enter the words as a list of words in the box provided. Minimum List Count - determines the number of duplicates necessary for all words or a part of speech to be displayed. Sentences ignore the minimum list count. The minimum list count defaults to 1. If you change it, the change sticks from session to session. Responses List Menu Options Reset List Resets the response list to display all responses for the selected view (not just responses containing the selected word). 455 PrintDoc01 020813 Hide/Show Word List Toggles (shows or hides) the Word List. Add Selection To Search Places the highlighted text of the current selected response into the Search Expression box. Highlight Selection Highlights the current selection wherever it is found in the response list. Clear Word Selections Removes any highlighted selections from the response list. Copy Copies the selection in the current row to the Windows Clipboard. Copy to Work Pad Copies the selection in the current row to the Work Pad. Make Code Adds a new code to the codebook with the text of the current selection. Using the Code Builder Work Pad The Work Pad is a pop-up window containing a list of text items that can be used to hold ideas or text found while reviewing the response data of a question. To add text segments to the Work Pad, select text and then right-click it to get the context menu. Select Copy to Work Pad. Items can also be added if you double-click below the last item in the Work Pad. A blinking cursor will appear indicating the list will accept entry of text. The Work Pad can be forced to remain on top of all other windows by choosing the "Always on Top" option from the Work Pad context menu. The Work Pad context menu can be displayed by right-clicking in the title bar of the Work Pad window. Items can be dragged from the Work Pad to the codebook to add them as codes. Items in the list can also added as codes by using the list items context menu "Make Codes" option. Remove items from the list by using the context menu of the Work Pad items. Codebook Renumber Options When the Renumber option is selected either from the Code Editor context menu or the toolbar button, a dialog with the following options is presented. Option Description Items To Be Renumbered If the Selected Items option is chosen, only those items highlighted in the Code Editor will be renumbered. Otherwise, all codes in the codebook will be renumbered. Value This setting determines which of the code values will be renumbered. Items To Renumber Choose Codes to renumber just the codes in a codebook. Choose Nets to renumber just the nets, including all subnets. Choose Both to renumber all items in the codebook. Optionally, checking "Renumber blank values only" will not renumber any items that already have a value. 456 Ascribe™ Desktop Start Code Value Specifies the value to assign to the first code or net that is renumbered. Spread Between Codes The amount by which each code value will be incremented. Specifies the increment applied to each value as codes or nets are renumbered. For example, if the Start code value is 100, and the Spread between codes is 10, the values assigned will be 100, 110, 120, etc. Digits Specifies the minimum number of digits to appear in the number assigned. A blank value or zero means that the value should not be left padded with zeros. A number greater than zero means the value should be padded to the left with zero digits to this number of digits. The maximum allowed value for Digits is 20. Codebook Sort Options When you sort the codebook, you have these options: Option Description Items To Be Sorted If the Selected Items option is chosen, only those items highlighted in the Code Editor will be sorted. Otherwise, all items in the codebook will be sorted. Sort Method This setting determines which of the code values will be renumbered: input code value, output code value, or code description. Working with Spell Check Results When a spell check process is completed without finding errors, a message will be displayed to that effect. If errors are found, a dialog box will be displayed. Each error instance will be displayed beginning with the first item in the codebook that contains a spelling error. As each error instance is disposed of by one of the button options, the next error instance will be presented. Option Description Item Represents the current spelling error found as a result of the most recent spell check process. In the case of the codebook spell checking, this box displays the entire code description with the error noted by a red text color. Change To This is the text that will replace the current spelling error text when one of the 'Change' buttons is clicked. Suggestions The suggested changes for the current spelling error. When a choice is selected here, it will replace existing values in the "Change To" box. Button Options Ignore Click to ignore this instance of the current spelling error. Other instances of this error will be presented as they are found. 457 PrintDoc01 020813 Ignore All Click to ignore all instances of the current spelling error. If this same spelling error is found in other codebook items, it will not be presented to be corrected. Change Click to replace this instance of the current spelling error with the new text. Change All Click to replace all instances of the current spelling error with the new text. Add Click to add the currently identified spelling error to the Microsoft Word dictionary. This process will remove the item as an error and prevent its detection as an error in future spell check processes. Options This button is currently not available. Close Closes this instance of the spell check results dialog. Paste Codes into a Codebook Codes can be pasted into the codebook from a text document or an Excel ® spreadsheet. Before you paste codes, you should prepare the source document so that the outputID of the code appears to the left of the description of the code, separated by a tab (or in a different cell in Excel®). You can also indent the codes, using tabs or empty cells to indicate the nesting level. Here is an example of a text document that you could paste into a codebook: 100 200 300 Code Code 210 220 Code 1 2 Code 2 Sub 1 Code 2 Sub 2 3 Text AutoCoder The Text AutoCoder provides the same functionality as the "auto-code at load." That is, uncoded responses are compared to previously coded responses from questions sharing the same codebook with the target question. Exact verbatim matches are coded with the same codes and in the same order as the previously coded responses. Start the Text AutoCoder To start the Text AutoCoder process, right-click a question either in the Navigator pane, a Home view question table, or a Workspace question icon. Select the "Code" menu option and then the "Text AutoCoder" option. Process Overview The Text AutoCoder starts with a list of uncoded text responses for the selected question. The actual coding is a two-step process. First, the uncoded responses are matched to previously coded responses. The source of the matched responses and matching criteria can be controlled with the Text AutoCoder toolbar options described below. If matches are found, they are displayed in the column labeled "Codes to Apply." Codes are then applied or added to the permanent coding for the question. The entire process is controlled by buttons on the Text AutoCoder toolbar. 458 Ascribe™ Desktop Toolbar Options Option Description Question To Code This option provides a drop-down list of other questions in the current study. It provides a convenient way to change which question is to be auto-coded. It is initially set to the question whose menu option was selected. Changing to a different question will reset all in process activities and display the uncoded responses of the newly chosen question. Studies in Scope This button provides a list of all studies that contain at least one question that shares a codebook with the question selected to be auto-coded. Checking a study will include all previously coded responses from questions with the shared codebook in the list of potential responses to be matched. The "Options" button provides for control over the matching process. The individual matching options are described below. Each can be selected/ unselected by checking or unchecking the box next to the option. Case Sensitive Comparisons - If selected, responses will only be considered matched if the case of each verbatim is the same. Options Keep Order of Mention - This option determines which previously coded responses will be considered for matching. It applies when there is more than one previously coded response with the same set of applied codes. If this option is selected, then each of those previously coded responses must have the set of codes applied in the same order. Otherwise, that verbatim will not be considered valid for matching to uncoded ones. If the option is not selected, then the group of previously coded responses with the same set of codes will be considered as valid for matching. Reset This button removes all matched coding. Apply Codes Click this button (available when there is at least one match response) to apply the codes to the responses permanently. After the process is complete, the newly coded responses are removed from the list. Match Click this button to begin the matching process. Remove Matches After responses have been matched and the matched codes are displayed, the codes can be removed in one of two ways. First, all matches can be removed by clicking the "Reset" button in the toolbar. Second, matched codes can be removed selectively by choosing one or more responses from the list, right-clicking to display the context menu, and then choosing the "Remove codes to be applied" option. Codes from the selected response will be cleared from the list and not applied when the "Apply" button is clicked. Multiple Responses Coding Tool 459 PrintDoc01 020813 The Multiple Response Coding Tool allows users to apply codes to responses and modify existing coding in similar way as "Review Responses" in Ascribe™ Web. The Multiple Response Coding Tool screen is divided into three sections: a top toolbar, the Response Pane, and the Tabbed Pane. The Response Pane has two columns: the coding source and the codes applied. The Tabbed Pane displays a tabbed view of the Codebook Window, the Expression Tester Window, the Word List Window, the Edit Response Window, and the Other Responses Window. Top Toolbar The main toolbar provides the following options. When you display responses, the Responses Source displays at the bottom of the Codes Applied column so you can tell the source of the responses. Option Codebook Description Click this button to open the Codebook Editor in the Navigation Pane. Once the Codebook Editor is open in the Navigation pane, all editing functionality is available there. Codes can be added, moved, or deleted (if unused). Code properties can also be changed. Word List Click this button to show or hide the Word List depending upon its current state. View Options Click this button to open the View Options window. This toolbar option provides a set of options for selecting or removing examples for display. The options are: Show Examples Display Responses All - Displays all examples for the selected Language Coded Only - Displays all coded examples Uncoded Only - Displays all uncoded examples Clear Display - Clears the Examples Pane display Search... - Opens the Expression Tester Window to enable searches for examples containing specific text. Respondent Filter Responses Source This toolbar option displays the Respondent Filter window to allow filtering of responses displayed based upon the respondents’ answer to another question. It is the same filter used in Ascribe™ Web. (See Filter - Select a Response Filter for more information.) Click this button to select the source of the responses for display, search, filtering, etc. When you display responses, the Responses Source displays at the bottom of the Codes Applied column so you can tell the source of the responses. Options are: All Shared Questions - Only questions that share the codebook with the current question are listed. If no questions share the codebook with the current question, this option will be disabled (grayed out). Individual Questions - Individual questions that share the 460 Ascribe™ Desktop codebook with the current are listed. You can select one or more questions. Available Samples - Samples that share the codebook with the current question and are in the current study are listed. If no samples are available, the word 'none' will display. For Samples for more information. Create Sample - Opens the Create Quick Sample dialog; you can create a sample from the questions that share the codebook of the current question or just from this question. See the Multiple Responses Coding Tool section of Samples for more information. Response Pane The Response Pane is the large area on the left side of the window below the toolbar. It displays responses and codes applied based upon the selection criteria. Upon initial entry to the window, the area is blank. The response display consists of two columns: the left column displays the Coding Source Text. If the responses are from shared questions, the question ID displays when you hover over the response. The right column displays the list of codes, if any, that have been applied to the response. Hover over a code to see who applied the code and when the code was applied; this information displays at the bottom of the left column. Context menu options for each column are available when you right-click a response or a code. The menu options available in the Coding Source Text column are: Option Description Find Code After you highlight some text and right-click it, you can search the codebook for that text. The first instance of that text in the codebook is highlighted in the Codebook Window of the Tabbed Pane. The Find box in the Codebook Window will also contain the search text so you can continue searching the codebook by using the Find box. Add to Notes After you highlight some text and right-click it, you can add the selected response text to the notes for the response. To see the added note, click anywhere else on the screen, and the note will display below the response. Edit Response Opens the Edit Response Window for the current response. Mark as Quality Reviewed Marks selected responses as quality reviewed. Clear Quality Reviewed Clears the quality reviewed status for selected responses. Clear All Responses Un-selects all currently selected responses. Select All Responses Selects all responses in the list. 461 PrintDoc01 020813 Save Selected Responses as New Sample You can create a sample from selected responses. See the Multiple Response Coding Tool section of Samples for more information. Hide Selected Responses Hides selected responses. Display Hidden Responses Displays all previously hidden responses. Other Responses Opens the Other Responses Window for the current respondent. Deletes all codes applied to selected responses. First, highlight the response or responses, and then control-click a response. Remove All Codes from Select this option from the menu. A dialog box displays and Selected Responses asks you to confirm the deletion. Type OK in the text box and click the OK button. The menu options available in the Codes Applied column are: Option Description Clear All Selections Unselects all selected codes. Select All Same Code Selects all codes that are the same as the current code. Remove Selected Codes Deletes the code from the list of codes applied to each of the selected responses. Tabbed Pane The Tabbed Pane on the right side of the Coding Tool Window displays various functional windows based upon the user's navigation choices. The Tabbed Pane can optionally be hidden. To hide, right click in an empty area of the top toolbar. Uncheck the Display Tab Pane option. When the Tabbed Pane is hidden, the windows described below will appear as "floating" when you use the toolbar options to display them. You can also "float" the tabbed windows if you click one of the tabs and select Float. A "floating" window can be closed if you right-click its header and select Close. To have it display in the Tabbed Pane again, use the toolbar options to display it. Question Information The upper portion of the Tab Pane provides an area to display information about current question. Information displayed includes the ID, the label, and the text of the question. If shared questions are involved, the question information for the currently selected response will display. If more than one response is selected, the Question Information window displays "???". To hide the Question Information window, right-click in an empty area of the top toolbar. Uncheck the Display Question Information option. Codebook Window 462 Ascribe™ Desktop The Codebook Window displays the codebook for the current question. Although similar in appearance to the Codebook Editor, this display of the codebook does not allow editing of the codebook itself. To edit the codebook while in this coding tool, open the Codebook Editor in the Navigation Pane. Several menu options are available for using the codebook and controlling the display. Option Description Expand All Expands all nets in the codebook to display each net's children. Collapse All Collapses (hides) all but the top level codes and nets. Provides a list of options that control display of the codebook. View Options... Display help available symbol - If selected, a symbol appears in front of the code description when code help is available. Codes only - If selected, only codes are displayed in alphabetical order. Displays options for selecting responses to display in the Response Pane. Select Responses... Match any selected code - If selected, responses with any of the currently selected codes applied to them are displayed. Match all selected codes - If selected, only responses with all the currently selected codes applied to them are displayed. Expression Test... Opens the Expression Tester Window with the selected code's regular expression text. Expression Tester Window The Expression Tester window allows users to search responses for text that matches the search criteria. This is a powerful search tool using regular expressions. The window offers a text box for entering the search text as well as a filter to determine which responses to search. The filter options available are: Show Uncoded Responses Only - Searches only responses that have not had at least one code applied. Show Coded Responses Only - Searches only responses that have had at least one code applied. Show All Responses - Searches all responses for the selected questions. From Current Result Set - Searches only responses from the previous search results. Responses that match the search criteria will be displayed with matches highlighted. Whatever filter option you select stays selected until the next time you use the Expression Tester. For example, you select Show Coded Responses Only. The next time you open the Expression Tester, that option is still selected. You can select a different filter every time you use the Expression Tester. 463 PrintDoc01 020813 Word List Window The Word List window displays a list of words contained in the response set. If the Tabbed Pane is visible, the Word List is displayed in a tab in that panel. If the Tabbed Pane is not displayed, the Word List window is displayed as a "floating" window. The Word List displays two columns. The left column lists the words. The right column displays the count of occurrences of the word in the response set. Click on either column heading to sort the list, either in ascending order or descending order depending on its prior sort order. View Options You can view all words or determine words to display based on their part of speech (nouns, verbs, adjectives, adverbs, noun phrases, or sentences.) Display Responses with the Word List Double-click a word to display all responses that contain that word. Display Words that are Near Each Other You can also see two words that are "near" one another. "Near" is defined as within 20 characters, which is about 3-5 words. To use this function, drag a word from the Word List to another word in the list. When you drag, you will see faint text that says '(Word1) NEAR (Word2)'. Word1 is the word you are dragging, and you drop Word1 on Word2. You will see responses where Word2 is within 20 characters after Word1. The word you drag will come before the word on which you drop it. For example, the Word List displays these words: juice, kids, and tangy. To see responses with a phrase like 'tangy juice,' you would drag tangy to juice. To see responses with a phrase like 'juice for kids,' you would drag juice to kids. Add Words to the Ignored Words List Right-click a word to add it to the list of ignored words. Word List Options The Word List also has a toolbar at the top with a button to edit options related to the Word List. See Edit List Options for more information. Edit Response Window The Edit Response Window displays the verbatim, translation, transcription, and notes fields for a response. You can edit all fields except the verbatim. Changes are saved when moving from field to another. When this window is displayed, selecting a response from the list in the Response Display Panel will change the response information in the Edit Response Window to the newly selected response. 464 Ascribe™ Desktop The toolbar at the top of this Edit Response Window has one option: "Advance after Enter". If this option is left unchecked, pressing the enter key when in any of the edit boxes will move the cursor to the next available box and save any changes made in the current box. If this option is checked, pressing the enter key saves the current changes, remains in the current box and sets the response to the next one shown in the Response Display Panel. Other Responses Window The Other Responses Window displays the responses (Coding Source Text) and codes applied to all other responses for the selected respondent. The toolbar at the top of this Other Responses Window has one option: "Question Types". This button provides a choice of which Ascribe™ question types should be displayed in the list. When this window is displayed, selecting a response from the list in the Response Display Panel will change the response information in the Other Responses Window to the newly selected respondent. Set View Options This View Options Window allows the user to set various properties that control the view and display of items in the Response Display panel. The settings established here are set for the user and will remain in place from session to session. In addition, many of the options are shared with the same option in the Ascribe™ Web Review Responses window. Option Description Coding Source Check this option if you wish for the content of the coding source for the question to be displayed. Typically, this is the response verbatim and would generally be checked. Additional Display Use this drop-down list to select an additional aspect of the response to be displayed. Choices will always include verbatim and notes. Additionally, if translations and transcriptions are available to the account, those will be listed as well. This section of the View Options allows filtering in addition to respondent filters. These filters are unique to this coding window and together with respondent filters determine what responses are displayed. The filters are: Filters Additional Display: Responses can be filtered based upon whether a value for the Additional Display field exists for the response. For example, if "Transcriptions" are selected as the additional display field, then the selection options would allow viewing only responses with transcription, without transcriptions, or both. Review Status: The second option allows filters to be set based upon the quality reviewed status of a response. Coder: The third option can be used to filter by the user who applied the code. To set that filter, select the user from the user drop-down list. Only users who have applied codes will be displayed in the list. Coding Method: The fourth filter option is based upon how 465 PrintDoc01 020813 the code was applied: manual, autocoded, or ACM. This setting determines how the responses listed in the left pane will be sorted. The options are: By Respondent ID By Question ID, Respondent ID (useful with shared codebook questions) By Response (the verbatim text ) By Transcription text By Translation text Sort Order By Notes text By Text Length - sorts based upon the length of the coding source text. By Codes Applied - sorts based upon the number of codes applied. The sort for all but text length and codes applied are done alphabetically. Since many of the transcription or translation values can be empty, the sorting could appear to produce unusual results when those options are selected. The sort direction (either Ascending or Descending) can be specified by choosing either the ' Asc' or ' Desc' options. These options determine what information in addition to code description will be displayed in the codebook window: Codebook Display Show Details - If checked, counts of the codes applied will be displayed after the code description Show Code Values - If checked, the code output ID will be displayed before the code description. Show Net Values - If checked, the net output ID will be displayed before the code description. Respondent ID Color There are two color choices here. One establishes the color used to render the respondent ID when the response has not been marked as quality reviewed. The second color selection determines the respondent ID color when the response has been marked as quality reviewed. The colors can be changed by clicking on the appropriate color box and choosing a new color from the Color window. Show Respondent ID If checked, the respondent ID is displayed with the coding source text. Highlight Matched Text If checked, text that matches the Expression Test search will be highlighted using the selected color. The highlight color can be changed by clicking on the color box and choosing a new color from the Color window. Code with Multiple Responses Coding Tool This section explains how to code using the Multiple Responses Coding Tool. Select Responses to Display 466 Ascribe™ Desktop Responses can be selected for display in several ways: Use the top toolbar's Display Responses option to quickly display all responses for the selected questions. Use the Codebook Window's Select Responses option to display responses for selected codes. Use the Expression Tester Window to search for and display responses matching the text criteria. Once you have displayed responses in the response pane, you can use menu commands and drag and drop operations to alter coding of the selected responses. Code a Single Response To code a single response, first ensure that no responses are highlighted. Choose Clear all Responses from the Response Display Panel's menu to clear the highlight (unselect) from all responses. To code a response, click on the desired code in the Codebook Window, and drag it to the response you want to code. The response turns blue. Drop the code on the response. You can also drag a code from the Codes Applied column and drop it on another response. You can apply more than one code at a time if you control-click multiple codes in the Codebook Window and then drag them to a response. Code Multiple Responses To code multiple responses, first ensure that the responses you want to code are selected. Double-click a response to select it. To select all responses, choose Select all Responses from the Response Display Panel's menu. Click on a code in the Codebook Window, and drag it to any response. The response turns blue. Drop the code on the response. Note When multiple responses are selected, dropping a code on any of them applies the code to all selected responses. You can apply more than one code at a time if you control-click multiple codes in the Codebook Window and then drag them to selected responses. Swap Codes Swapping codes replaces the code applied with a different code from the codebook. To swap codes, first select the responses for which you want to swap codes. In any of the selected responses, click on the code you want to swap. Drag the code to the Codebook Window, and drop it on the code you want to replace it with. The code will be swapped in all of the selected responses. Remove Codes There are several ways to remove codes, depending on what codes should be removed. 467 PrintDoc01 020813 To remove a single code from a single response, right-click the code and select Remove Selected Codes. To remove the same code from all responses, right-click the code and choose Select All Same Code. Right-click the code again and select Remove Selected Codes. To remove multiple codes from a single response, control-click each code, and then right-click one of the codes. Choose Remove Selected Codes. To remove multiple codes from multiple responses, control-click each code, then right-click one of the codes, and choose Remove Selected Codes. Another way is to right-click a code and choose Select All Same Code. Right-click the code again and select Remove Selected Codes. Do the same action for every code you want to remove. To remove all codes from all responses, right-click a response and choose Select All Responses. Control-click a response and choose Remove All Codes from Selected Responses. A dialog box displays. Enter OK in the text box, and press the OK button to continue. If you want to unselect any codes, right-click the code and select Clear All Selections. Hide a Response You can remove a response from the list of responses displayed (so that it will not be affected by coding operations). To hide a single response, right-click the response and select Hide Selected Responses. To hide multiple responses, control-click the responses, then right-click one of the responses, and select Hide Selected Responses. To see the hidden responses, right-click any response and select Unhide Selected Responses. Display Notes, Transcriptions, and Translations To display notes, transcriptions, or translations in the Response Display Panel, click View Options in the top toolbar. Select the desired option in the Additional Display list. Edit Notes, Transcriptions, and Translations To change the text of a note, transcription, or translation, right-click on the response and select Edit Response. Add Selected Text to a Note To add text from a response to the notes for that response, first click on a response. Next drag the mouse over the desired text to select it. Right-click on the response and select Add to notes. The selected text is appended to the notes. You must select Notes in the Additional Display list of the View Options for this feature to be available. Change Review Status To change the review status of a response, right-click and select either Mark as Quality Reviewed or Clear Quality Reviewed. This will change the review status of all highlighted responses. If no response is highlighted, the status of the single response you clicked on is changed. 468 Ascribe™ Desktop The review status of a response is indicated by the color of the respondent ID. To select color preferences, click View Options in the buttons pane. Also note that responses can be marked or cleared of the Quality Reviewed status regardless of whether codes have been applied to the selected responses. Codebook Merge/Map Tool The Codebook Merge/Map Tool has two modes: Merge - This mode allows users to merge the coding of two codebooks into one. In Ascribe™ terminology, this means users can share a codebook between two questions. In this process, one codebook is considered the target and one is considered the source codebook. The target codebook is the one to survive the merge process and will be the one shared among the questions selected for the merge operation. Because codes from one codebook (the source) will be replaced by codes from another (the target), the Code Merging process cannot be completed unless all codes from the source are matched to a code in the target. If, during the merging process, one code fails to be successfully replaced, the entire process will be terminated without any change. Map - This mode allows users to map the codes of one codebook to another. In this process, one codebook is considered the target and one is considered the source codebook. The source codebook will have the designated code ID (Input, Output, or Dictionary) of each code matched to a target codebook code set to the target codes value. Unlike Code Merging, the mapping of codes can be accomplished with less than all source codebook codes being mapped to target codebook codes. The Codebook Merge/Map Tool has three primary viewing panes. The top left portion of the screen displays the target codebook. The top right portion displays the source codebook either as a traditional view of a codebook or a matched view. See Working with the Match View for more information. Finally, the lower portion of the screen displays the list of target/source matches. Merge Codebooks To begin, select the Coding Merge mode by clicking the Mode button on the toolbar then selecting the "Coding Merge" option. Next, drag a question or codebook item onto the area marked "Drop target codebook here". This codebook will serve as the target. Next, drag the question or codebook that will serve as the source codebook. Once each codebook has been identified, match source codes to target codes by dragging them onto a target code. Matched codes will be displayed in the bottom portion of the screen. Optionally, choose the match option from the Merge Mode toolbar to attempt to "auto match". Merge Mode Toolbar Options Option Description Mode The options are either Coding Merge or Code Mapping. Save Click this button to save the current state of the tool (i.e., target codebook, source codebook and matched codes) to a 469 PrintDoc01 020813 disk file. This option is available when at least one code has been matched. Restore Restores a previously saved Codebook Merge/Map Tool state file. Unmatched Codes Remaining This portion of the toolbar is informational only. It provides a count of the source codes not yet matched to a target code. All codes must be matched before the Merge Code function can actually be invoked. Four reset options are available when this button is clicked: Right Source (right side) - Removes the source codebook from the tool and resets the matched list. Resets Reset Target (left side) - Removes the target codebook from the tool and resets the matched list. Reset Matches - Clears all matches. Reset All - Removes both the target and the source codebooks from the tool and removes all target codes from the match list. Click this button to automatically match codes from the target to the source codebooks. There are four options available to use for the match criteria: Code Description - The source and target code descriptions must match. Input ID - The source and target code input IDs must match. Output ID - The source and target code output IDs must match. Match Codebook Net Position - The source and target positions in the codebook must match. This option is quite restrictive and means the codebooks must have essentially identical structures for codes to match. These options can be used together or individually. If used together, each of the criteria items must be equal between the source and target codes for the items to be considered matched. Equality is determined by a case insensitive comparison of the items. Once the match process is complete, the top right side display switches to the Match View. See Working with the Match View for more information. Merge Coding When all source codes have been matched to a target code, this button will available. Upon clicking the button, the user is presented with a list of the questions currently using the source codebook. One or more questions should be chosen to have their coding merged by checking the box beside each question. Map Codebook Codes To begin, select the Code Mapping mode by clicking the Mode button on the toolbar and then select the "Code Mapping" option. 470 Ascribe™ Desktop Next, drag a question or codebook item onto the area marked "Drop target codebook here". This codebook will serve as the target. Then, drag the question or codebook that will serve as the source codebook. Once each codebook has been identified, match source codes to target codes by dragging them onto a target code. Matched codes will be displayed in the bottom portion of the screen. Optionally, choose the match option from the Mapping Mode toolbar described below to attempt to "auto match". Mapping Mode Toolbar Options Option Description Mode The options are either Coding Merge or Code Mapping. Save Click this button to save the current state of the tool (i.e., target codebook, source codebook and matched codes) to a disk file. This option is available when at least one code has been matched. Restore Restores a previously saved Codebook Merge/Map Tool state file. Unmatched Codes Remaining This portion of the toolbar is informational only. It provides a count of the source codes not yet matched to a target code. All codes must be matched before the Merge Code function can actually be invoked. Four reset options are available when this button is clicked: Right Source (right side) - Removes the source codebook from the tool and resets the matched list. Resets Reset Target (left side) - Removes the target codebook from the tool and resets the matched list. Reset Matches - Clears all matches. Reset All - Removes both the target and the source codebooks from the tool and removes all target codes from the match list. Click this button to automatically match codes from the target to the source codebooks. There are four options available to use for the match criteria: Code Description - The source and target code descriptions must match. Input ID - The source and target code input IDs must match. Match Output ID - The source and target code output IDs must match. Codebook Net Position - The source and target positions in the codebook must match. This option is quite restrictive and means the codebooks must have essentially identical structures for codes to match. These options can be used together or individually. If used together, each of the criteria items must be equal between the source and target codes for the items to be considered matched. Equality is determined by a case insensitive 471 PrintDoc01 020813 comparison of the items. Once the match process is complete, the top right side display switches to the Match View. See Working with the Match View for more information. Map Codes When at least one source code has been matched to a target code, this button will available. Upon clicking the button, the user is presented with a selection of code properties that will be used for the mapping. Only one of these optional properties can be selected. Once selected, each code in the source codebook that has been mapped to a target code will have the selected property set to the target code's property value. Working with the Match View The Match View is displayed in the top right panel in lieu of the source codebook display. There are two ways to display the Match View. It is automatically displayed after the Match option has been executed from the toolbar. And, it can be displayed on demand by choosing the "Switched to Match View" from the source codebook display context menu. The match view aligns matched source codes to their target code position. Matched codes text is displayed in green. Unmatched codes are displayed at the bottom of the list in red. If more that one source code is matched to a target code, the text in the match position of the target is displayed as "***". To display the target and source lists aligned to their matched positions, hold the CTRL key down while scrolling the target codebook or selecting a target codebook item. The Match view also has its own context menu of options. Those options are describe in the following table. Option Description Switch to Codebook View Changes the display of the top right panel to the Source Codebook View. Create Selected Codes in Target Codebook Selected items in the Match View list will be added to the target codebook and matched to that newly added target code. This option is available only for source codes that have not yet been matched (i.e. those codes displayed in the color red.) Drag and Drop Options The Codebook Merge/Map Tool offers several options using drag/drop functionality. Codes in the Source Codebook View or Matched View can be dragged on to a target code to match them. When dragging from the codebook view, nets can be dragged onto a target source code to match all codes that are children of the net. Codes selected in the Matched View can also be dragged on to a floating Codebook Editor of the target codebook to add those codes to the target codebook in the desired position. Unmatch Codes 472 Ascribe™ Desktop Codes can be unmatched by selecting " Unmatch the Codes" from the Mapped Code List in the bottom portion of the screen. Study Import Tool The Study Import Tool allows users to extract questions and codebooks from one or more source files into a selected target study. Optionally, a study can be created with selected questions. Currently, information can be imported from three types of files: Ascribe™ XML, Dimensions MDD, and Confirmit CXML. The Study Import Tool screen is divided into three sections: the standard top toolbar containing functional controls, the main data display panel, and a left-side vertical toolbar providing navigation controls. The navigation toolbar controls allow the user to choose the data displayed in the main display panel. Study information, the file's question list, or the file's codebook list can be displayed. Top Toolbar The main toolbar has the following options. Option Description Setup File Enter the name of the file to be imported. Use the ... button to browse for the file. The file type drop-down box will be set automatically when a file is selected based upon the extension of the file. The default extensions are mapped as follows: xml - Ascribe™ File Type mdd - Dimensions cxml - Confirmit If the file extension is one of the known extensions described above, but does contain valid content, use this drop down to select the appropriate type. Refresh This toolbar option will reload the current file and refresh each of the screens. Add New Click this button to add a new study with the current information and selected questions. Left-Side Toolbar The left-side toolbar has three selection options. They are used to change the information displayed in the main data display panel. Option Description Changes the display to study setup information. Changes the display to list the questions contained in the selected file. The number in parenthesis below the image indicates the count of questions in the selected file. 473 PrintDoc01 020813 Changes the display to list the codebooks contained in the selected file. The number in parenthesis below the image indicates the count of codebooks in the selected file. Add a New Study To begin, enter the file to be imported in the Setup File box in the top toolbar. This file must be of a known format. The current supported formats are Ascribe™'s saved study format (xml), Dimensions's meta data file (mdd), and Confirmit's exported Ascribe™ setup file (cxml). If the file selected has an extension that matches one of the known types (i.e. xml, mdd, cxml), the file type will be set automatically. The Importer tool will process the file and then display the information. If the extension does not match a known type, the file type will be set to "Unspecified". If the file actually is a known format, simply change the file type to the correct format. The file will then be read and displayed. Attempts to process a file of unknown format will result in an error message. If you wish to create a new study with the existing setup information, simply click the Add New button in the top toolbar. If you wish to create the study and add selected questions, switch to the Questions list by selecting the Question Image (the middle image) in the left toolbar. Select one or more questions. Then click the Add New button. For convenience, the Questions list has menu options available to select, hide, and show all or selected question types. Add Questions to an Existing Study To begin, enter the file to be imported in the Setup File box in the top toolbar. This file must be a known format. The current supported formats are Ascribe™'s saved study format ( xml), Dimensions's meta data file (mdd), and Confirmit's exported Ascribe™ setup file (cxml). If the file selected has an extension that matches one of the known types (i.e. xml, mdd, cxml,) the file type will be set automatically. The Importer tool will process the file and then display the information. If the extension does not match a known type, the file type will be set to "Unspecified". If the file actually is a known format, simply change the file type to the correct format. The file will then be read and displayed. Attempts to process a file of unknown format will result in an error message. Switch to the Questions list by selecting the Question Image (the middle image) in the left toolbar. For convenience, the Questions list has menu options available to select, hide, and show all or selected question types. In the navigator pane, open the study node of the desired study. Drag selected questions from the Study Importer questions list to the Questions collection node of the selected study. Study Merge Tool 474 Ascribe™ Desktop The Study Merge Tool allows users to copy responses from one or more source questions into a selected target question. Responses will be assigned new respondent IDs and will have any coding applied copied as well. This functionality is useful when it is desirable to build a repository of coded responses for a particular codebook. The Study Merge Tool screen is divided into three sections. The top left panel displays a list of target questions. The top right portion displays a list of questions that share a codebook with a selected target question. Finally, the bottom portion of the screen displays a list of the mapped target/source question combinations. In order for responses and coding to be copied from one set of questions to a target question, the target question and source questions must share the same codebook. Before the copy process can begin, at least one source question must be “mapped" to a target question as described below. To begin, drag a study item onto the area marked "Drop the target study here". This study will serve as the source of the target questions. A list of the questions appears in the top left panel. As a question is selected from the Target question list, a list of questions, if any, that share the codebook with the selected target question will be displayed in the right side panel. If no questions share a codebook, a message to that affect is displayed in the top right panel. If all shared questions have been mapped, an empty list appears. Study Merge Tool Functions To map source questions to a target question, select the desired source questions and then choose the "Map selected questions" option from the source question list context menu. To unmap source questions from a target question, select the desired mapped list item from the bottom portion Mapped Question list and then choose the " Unmap the questions" option from the source question list context menu. To automatically map questions based upon question ID, click the "Match Question IDs" button from the Study Merge Tool toolbar. To execute the merge process once questions are mapped, click the "Copy Mapped Questions" button from the Study Merge Tool toolbar. To clear the current target study, click the "Reset" button from the Study Merge Tool toolbar and then choose the "Reset target (left side)" option. To clear the all current question mappings, click the "Reset" button from the Study Merge Tool toolbar and then choose the "Reset Matches" option. Study Merge Tool Copy Process Description When the Merge process is executed, the Study Merge Tool will copy each response from the source question to the target question. When doing this, the copied response has its respondent ID changed to a new unique ID. This ID is based upon the internal unique database identifier of the source response. If, while attempting to copy, the source response 475 PrintDoc01 020813 is found to already exist, it will be ignored. This means that if a Merge is attempted a second time on the same pair of target/source questions, no responses will be copied unless new responses have been added to the source question since the last Merge was executed. Note If a study's questions are merged into a target questions using the Study Merge Tool and then saved and restored, the responses are considered new responses. If the newly restored study's questions are again merged, the responses will be added again to the target questions responses. During the copy process described above, the original respondent ID is saved to the notes field of the copied response. Samples Samples are a way to look at a portion of the data; they are simply a named sub-set of the responses. The sample responses are still part of the source responses and do not exist separately. If you delete the source of the sample, the sample is also deleted. A sample cannot be edited and remain valid. Samples can be helpful if you load a large number of responses and only want to code part of them. Or you can just use it as a way to filter the responses by a particular word or concept and be able to return to that sub-set. The sample could be used for quality review of coding or to create training examples for ACM. How to Create a Sample Desktop provides several ways to create a sample. From the Navigator Pane, right-click one of these items and select Extract Response Sample to create a sample: Source Sample Where Used A question whose codebook is not shared with other questions or a model The sample will be taken only from that question. Multiple Response Coding - The sample is available for the source question only. A question whose codebook is shared with other questions in the same study The sample will be taken only from that question; you do not have access to the shared questions when creating a sample. Multiple Response Coding - The sample is available to any question in the same study that shares the source question's codebook. A question whose codebook is shared with questions across studies The sample will be taken only from that question; you do not have access to the shared questions when creating a sample. Multiple Response Coding - The sample is available to any question in the same study that shares the source question's codebook. The sample is not available 476 Ascribe™ Desktop across studies. A question whose codebook is shared with a model The sample will be taken only from that question; you do not have access to the model when creating a sample. Multiple Response Coding - The sample is available to any question in the same study that shares the source question's codebook. Training Example Editor The sample is available when the model shares the source question's codebook. A codebook that is not shared with other questions or a model The sample will be taken only from the question with this codebook. Multiple Response Coding - The sample is available only for the question with this codebook only. A codebook that is shared with questions in the same study The sample can be taken from all questions in the study that share this codebook. Multiple Response Coding - The sample is available to any question in the same study that shares this codebook. A codebook that is shared with questions across studies The sample can be taken from all questions that share this codebook, including questions in other studies. If you take a sample from across studies, you cannot use this sample, unless the codebook is also shared with a model. A codebook that is shared only with a model and no other questions The sample will be taken only from the question with this codebook. Multiple Response Coding - The sample is available only for the question with this codebook only. Training Example Editor The sample is available when the model shares this codebook. A codebook that is shared with a model and other questions (in the same study or across studies) The sample can be taken from all questions that share this codebook, including questions in other studies and the model. Training Example Editor If you take the sample from across studies and/or the model, the sample is available only in the Training Example Editor. 477 PrintDoc01 020813 A model *If you create a model from a question, you can take a sample from only that question. *If you create a model from a codebook, you can take a sample from all questions that share the codebook. Training Example Editor If you take the sample from across studies and/or the model, the sample is available only in the Training Example Editor. Note You can create samples from closed-ended questions, but currently, there is no way to use them. From the Questions page, right-click a question and select Extract Response Sample. The same rules in the Navigator table apply. From the Multiple Response Coding tool: Source Responses Source dialog - select one or more questions (when questions are shared in the study) Response Pane 478 Sample The sample can be taken from the question or questions you select in the Responses Source dialog. (A sample can be taken from any questions in the study that share this codebook; however, first, you must select the shared questions in the Responses Source dialog, which causes those responses to display in the Response Pane. Then you click Responses Source again and select Create Sample.) Click the check box next to a question or questions and click Create Sample at the bottom of the dialog. Display some responses in the Response Pane in whatever way you would like. Select (highlight) the responses from which you want to create a response, right-click, and choose Save Where Used Multiple Response Coding - The sample is available to any question in the same study that shares the source question's codebook. Training Example Editor The sample is available if the model shares the source question's codebook. Multiple Response Coding - The sample is available to any question in the same study that shares the source question's codebook. Training Example Editor The sample is available if Ascribe™ Desktop Selected Responses as New Sample. (Note: Since you determine which responses are selected, this sample is not random.) the model shares the source question's codebook. Note You cannot create a sample from a sample. The Response Source must be a question. Create Quick Sample Dialog Box The Create Quick Sample dialog displays after you select an item and choose to create a sample. The dialog has these fields: Field Description Sample ID Give the sample a unique ID; should be 30 characters or less. Description Optional description of sample Randomly Select This Count of Responses Enter a number in this field if you want to randomly select a specific number of responses. (This field does not display when you use the Save Selected Responses as New Sample option in the Multiple Responses Coding tool.) Randomly Select This Percentage of Responses Enter a percentage (whole number only) in this field if you want to randomly select a certain percentage of responses. (This field does not display when you use the Save Selected Responses as New Sample option in the Multiple Responses Coding tool.) Sample Sources This section displays this information for available questions for the sample: study ID, study name, question ID, question label, and number of responses. How to Use a Sample You can use a sample in the Multiple Response Coding tool or the Training Example Editor. In the Multiple Responses Coding tool, click the Responses Source button. The Responses Source dialog displays. Available samples are listed in the dialog. Click the check box next to the sample, and the responses display in the Response Pane. You can use a sample as you would other responses (coding, quality review, etc.) The source of responses displays at the bottom of the Response Pane to let you know whether responses are from a question or a sample. You can only view one sample at a time in the Multiple Responses Coding tool. The source of responses displays at the bottom of Response Pane. In the Training Example Editor, click the Import Examples button. The Import Examples dialog displays. Available samples are listed in the dialog. Click the check box next to the sample, and the sample will be imported as training examples for the model. 479 PrintDoc01 020813 View a List of Samples To see a list of samples for your account, choose View from the Main Menu and select Samples. The Samples page has these columns: Column Description Sample ID The sample ID Target Sample Responses The Target Sample Responses are a count of the original sample when the system randomly generates the sample. Target Sample Responses and Current Sample Responses start out with the same number. If the sample is changed, the two fields will not match. This difference lets you know that the sample is no longer valid. (For example, one of the questions used to create the sample is deleted. Those responses are no longer available in the sample, which makes the sample no longer valid.) If a sample is created by user selection with the Multiple Responses Coding tool, Target Sample Responses will be zero since the sample is not randomly generated by the system. Sample Description Optional description of sample Current Sample Responses Number of responses actually in the sample. Target Sample Responses and Current Sample Responses start out with the same number. If the sample is changed, the two fields will not match. This difference lets you know that the sample is no longer valid. Studies in Sample Number of studies used to create the sample Codebooks in Sample Number of codebooks used to create the sample Sample Owner The person who created the sample is the owner. Owners can delete their own samples; administrators can delete any sample. Sample Mode Description of how the sampled was created (random selection of count or percentage or user selection.) Delete a Sample You can delete samples from the Samples page, if you own the sample or if you are an administrator. Deleting the sample does not delete responses. To delete a sample, rightclick and select Delete. 480 Automated Coding Model (Desktop) This section gives an overview of automated coding models. For instructions on how to create and use models with Desktop, see ACM Process on Desktop. An automated coding model (ACM) learns to code automatically based on examples that you provide. See ACM Theory of Operation for more information. Ascribe™ Desktop adds several powerful tools that can be used in conjunction with Ascribe™ Web to manage the ACM process. You train a model on a specific codebook. A given codebook may have only a single model associated with it. When you first create the model, you will need to specify the codebook to be trained. We recommend that you give the codebook a descriptive ID before you create the model. You may want to make the codebook an account codebook to make the codebook easier to identify. To train a model for a codebook, you must first create a model for that codebook. The model appears in the Automated Coding Models folder in the Navigator. When you create the model, you specify certain options for the model and the training process, such as whether you want the ACM to compute estimated accuracy for each code. Next, you must provide a set of training examples. The model learns how to suggest codes by examining these examples. A model is able to code in multiple languages. If your codebook will be used for multiple languages, you must also list the languages you want to train. These appear in the Languages table for the model. When you train a model for multiple languages, you must provide a question ID for the question in the study that contains the language response. This information is used to separate responses into their proper language at both training and coding times. Once you set up a model, you start the training job. When the job completes, you should review the results to determine whether the model may be used to code questions. If you are satisfied with the results, you must approve the model to make it available for coders to use. After you approve a model, you add coding questions to the model. A question may be associated with only one model for coding purposes. If a given question is associated with an approved model, a coder may use the model to code the question. A coder may use a model to code a question by starting an ACM coding job. When the job completes, the coder may review and optionally accept the coding suggested by the model in the Review ACM Coding window. ACM Process on Desktop The ACM process on Ascribe™ Desktop has the following steps: Create the model - There are several ways to create a model with Desktop. A recommended way is to right-click the question that has the codebook you want for the model and select ACM/Create an ACM. 481 PrintDoc01 020813 Enter information in the Edit ACM dialog box - After you enter the information, click the OK button. A Training Example folder and a Coding Questions folder are added for the model, and you can view them in the Navigator. The Language node is automatically added to the Training Example folder; the Language default is en (for English.) (You can add Language nodes for other languages if it is a multi-lingual model. See Multi-Lingual Model for more information.) If you create a model from a question, the question is automatically added as a coding question. If you create a model from a codebook, no coding question is automatically added. Provide training examples - There are several ways to provide training examples with Desktop. You can load a file to the Language node in the Navigator or import examples with the Desktop Training Example Editor. You can copy examples from a study question if you display the question in Home View and drag it to the model's Language node in the Navigator. For more information, see Provide Training Examples. Train the model - After you've given each code enough good examples, you want to train the model. During the training process, the training examples for each code are analyzed to create patterns, rules, and classifiers for that code. From the Navigator, right-click the model and select Train. A training job is submitted, and you can view it on the Jobs page. View training results with the Desktop Training Example Editor - Use the Training Results Tab from the Desktop Training Example Editor. The Training Results page displays the trend and accuracy of the model. Examine this report to see which codes may need more or different training examples to receive better accuracy or trend scores. Use the Desktop Training Example Editor to improve training examples - Import or enter more examples on this page and code them. After you add more training examples, you need to train the model again. Approve the model - Once you are satisfied with the training results, you need to approve the model. After a model is approved, it can be used to code a question. From the Automated Coding Models page or the Navigator, right-click the model and select Approve. Share the codebook of the coding question with codebook used by the model - The codebook of the coding question must be shared with the codebook used by the model. To share a codebook, the coding question should be in the Data Pane/Home View. Select the question so it is highlighted. Place your cursor on the model in the Navigator. Drag from the model to the question in the Data Pane. The Copy/Share dialog displays; select Share and click OK. Add a coding question or questions to the model - Drag and drop from the Question List in the Data Pane to the Coding Questions for the model in Navigator. Start the ACM coding job - Right-click a coding question and select Start Coding Job. A dialog box displays with some options. See Coding Time Options for more information. You can view the coding job on the Jobs page. Review the ACM coding - Right-click a coding question which has been coded with ACM and select Review ACM coding. The Review ACM Coding tool opens. You also can accept the suggested codes with this tool. 482 Ascribe™ Crossword Improve the model - If you want to change or add training examples, you first must unapprove the model. When you change the status to unapproved, the training results are discarded. It does not delete any codes that have been applied to responses. To unapprove a model, right-click it and select Unapprove. Navigate to the Desktop Training Example Editor and continue the process add/change examples, train the model, review training results, approve the model, start the ACM coding job, and review the ACM coding until you are satisfied with the coding. List of Automated Coding Models You can view a list of the automated coding models in two places: in the Navigator or in the Home view. Automated Coding Models Folder in the Navigator The Navigator has an Automated Coding Models folder. If you click the plus sign next to the folder, a list of the models displays under the folder. If you click the plus sign next to a model, the Training Examples folder and Coding Questions folder display. Click the plus sign next to Training Examples folder to see the Languages for the model. Click the plus sign next to the Coding Questions folder to see a list of coding questions for the model. To see a list of the Languages in the Data Pane, click the Training Examples folder. To see a list of the Coding Questions in the Data Pane, click the Coding Questions folder. Automated Coding Models Table in Home View To see a table of models in the Home view, left-click the Automated Coding Models folder in the Navigator. The table contains these columns: Column ID Description An identifier for the model. The ID must be 30 characters or less. You can change this identifier if the model is not approved. The ID must be unique in this list of models. If the model is not approved, you can change this value either in the Edit ACM dialog. (You can open the Edit ACM dialog if you double-click anywhere in this table.) Description An optional description of the model. Codebook The ID of the codebook associated with this model. You cannot modify the ID of the codebook in this table, but you can change it in the Codebook Editor. Code Map Mode Unless all of the questions you will use at ACM coding time share the same codebook, you will need to specify a way to match codes in the codebooks. See the Match Codes By field in the Edit ACM dialog. If the model is not approved, you can change this value in the Edit ACM dialog. (You can open the Edit ACM dialog if you doubleclick anywhere in this table.) 483 PrintDoc01 020813 Training Examples The number of training examples in the model and includes examples from all Languages. Coding Questions The number of coding questions for the model. Training Languages The number of Languages in the model; there must be at least one Language in the model because the Language holds the training examples. Date Trained If a training job has been run, this column shows when the job was run. Training Job Status This field displays a status of Completed if the training job has run successfully, and a status of Failed if it did not run successfully. Click Completed or Failed to see the job details. Approved By If the model is approved, this column shows the user name of the person who approved it. Minimum Confidence When a coding job is run, and the user elects to automatically apply codes, those codes with a confidence value not less than this number will be applied to the response. If the model is not approved, you can change this value in the Edit ACM dialog. (You can open the Edit ACM dialog if you double-click anywhere in this table.) Compute Accuracy If this box is checked, the VCS™ will calculate the expected accuracy of each code in the codebook. This computation produces interesting information, but can be time consuming. Training takes about eight times longer when VCS™ computes accuracy. If the model is not approved, you can change this value in the Edit ACM dialog. (You can open the Edit ACM dialog if you doubleclick anywhere in this table.) Compute Trend When this box is checked, the VCS™ will calculate the trend in expected accuracy for each code. If the model is not approved, you can change this value in the Edit ACM dialog. (You can open the Edit ACM dialog if you doubleclick anywhere in this table.) Allow Auto Code If checked, codes can be applied automatically when a coding job is run using this model. If not checked, codes cannot be applied automatically at coding time, and the user must use the Desktop - Review ACM Coding Window to apply suggested codes. Use Lemmatisation You can look up lemmatisation in Wikipedia for a more detailed description, but basically lemmatisation involves grouping words with related meanings. Hence lemmatisation might group "soft" and "softness" as closely related. The ACM normally uses lemmatisation to allow it to generalize a given example as being applicable to closely related responses. If you need to consider such related words as distinct concepts, you can turn off lemmatisation for an ACM. This will generally cause the 484 Ascribe™ Crossword ACM to code fewer responses than it would if lemmatisation were enabled, but allows you to tune the ACM by providing more examples so that it will distinguish closely related words from one another. Hierarchical Codebook You may specify that the codebook for an ACM is hierarchical. This means that the codes below each net are related conceptually; in other words, the codes below a net are refinements of the concept represented by the net. Similarly, sub-nets are refinements of the concept represented by their parent net. If the entire codebook is constructed in this fashion, it is hierarchical. Specifying that a codebook is hierarchical will result in faster coding, and may also provide higher accuracy. Cleaning Regular Expression If a noise filter is used, it is listed here. Language Question ID The ID of the question that determines the language of a response. If blank, the model may be used for only one language. If not blank, each study used at training or coding time must contain a question with this ID. If the Language Question ID is not blank, more than one language may be specified in the Training Languages table. See Training for Multiple Languages. If the ACM training is not approved, you can change this value either in the Edit dialog, or by clicking twice in the ACM Trainings table. Examples Last Trained The number of examples that were trained during the last training job. Codes Last Trained The number of codes that were trained during the last training job. Training Result If a training job ran successfully, this field will have a check mark. Edit ACM Dialog The properties dialog is displayed when you create a new model or modify an existing model. This dialog is not available if the model is approved. If you want to modify the model properties, the model must not be approved. You can display the Edit ACM dialog if you right-click a model in the Navigator or double-click the model on the Automated Coding Models list when it is displayed in the Home View. The dialog box contains these entries: Field Description ID The unique identifier for the model. No two models can have the same identifier. The ID must be 30 characters or less. Description An optional description of the training. Noise Filter The regular expression used to remove noise from the response at coding time. See Noise Filter. 485 PrintDoc01 020813 Match Codes By Specifies how codes are matched between the model's codebook and the coding questions' codebook, if the codebooks are not shared. The options are: None - No method is used to match codes in different codebooks. In this case, all of the questions used for coding must share the trained codebook. Description - Codes match if the description of the two codes is the same, ignoring differences in case alone (e.g. capital 'A' is considered the same as lower case 'a'). Input ID - Codes match if the input IDs of the two codes are the same, ignoring differences in case alone. '12a' matches '12A', but '003' does not match '03'. Output ID - Codes match if the output IDs of the two codes are the same, ignoring differences in case alone. '12a' matches '12A', but '003' does not match '03'. Numeric Input ID - The match is performed by treating the input ID as a number, if possible, and comparing the numbers rather than the text. If the input ID cannot be treated as a number, then a text comparison is used. In this case, '12a' matches '12A', and '003' matches '03'. Numeric Output ID - The match is performed by treating the output ID as a number, if possible, and comparing the numbers rather than the text. If the output ID cannot be treated as a number, then a text comparison is used. In this case, '12a' matches '12A', and '003' matches '03'. Language Question ID This entry should be used only if the model has more than one language, and in that case, it is required. This entry specifies the ID of the question (usually a closed-ended question) that distinguishes the language of the response. See Multi-Lingual Models for more information. Use Lemmatisation You can look up lemmatisation in Wikipedia for a more detailed description, but basically lemmatisation involves grouping words with related meanings. Hence lemmatisation might group "soft" and "softness" as closely related. The ACM normally uses lemmatisation to allow it to generalize a given example as being applicable to closely related responses. If you need to consider such related words as distinct concepts, you can turn off lemmatisation for an ACM. This will generally cause the ACM to code fewer responses than it would if lemmatisation were enabled, but allows you to tune the ACM by providing more examples so that it will distinguish closely related words from one another. Codebook is Hierarchical You may specify that the codebook for an ACM is hierarchical. This means that the codes below each net are related conceptually; in other words, the codes below a net are refinements of the concept represented by the net. Similarly, sub-nets are refinements of the concept represented by their parent net. If the entire codebook is constructed in this fashion, it is hierarchical. Specifying that a codebook is hierarchical will result in faster coding, and may also provide higher accuracy. 486 Ascribe™ Crossword Approved If checked, this model is approved and can be used to run ACM coding jobs. Approved By If the model is approved, this field shows the user name of the person who approved it. Compute Accuracy If checked, accuracy of the model will be computed when the training job is run. This increases the time required to run the training job. Compute Trend If checked, the trend of the accuracy will be computed when the training job is run. This increases the time required to run the training job. Allow Auto Code If checked, codes can be applied automatically when a coding job is run using this model. If not checked, codes cannot be applied automatically at coding time, and the user must use the Desktop Review ACM Coding Window to apply suggested codes. To apply the codes at coding time, you must also select Apply codes to uncoded responses when you start the coding job. Minimum Confidence This option is only available if Allow auto coding is checked. When a coding job is run, and the user elects to automatically apply codes, those codes with a confidence value not less than this number will be applied to the response. See Coding Using Binary Classifiers for more information about confidence scores. You can view the confidence score of a suggested or applied code in the Review ACM Coding window. Training Example Languages To see a list of the Training Example Languages in the Data Pane, click the Training Examples folder under the model in the Navigator. This list has the following columns: Column ID Description The Language ID; it defaults to EN, for English. You can add Languages if you right-click the Training Examples folder in the Navigator and choosing Add New Training Language. Question Text The question text from the Questions Property dialog. Question Label The question label from the Questions Property dialog. Language Response This field will be blank if only the model only contains one Language. If the model has more than one Language, then the Language Response field will have the value that identifies the language in the closedended language question. The Language Response value is entered when you add an additional Training Language to the model. You can add Languages if you right-click the Training Examples folder in the Navigator and choosing Add New Training Language. See MultiLingual Model for more information. 487 PrintDoc01 020813 Examples The number of training examples in the model for this Language. Examples Coded The number of training examples that have codes applied. You can choose what columns display if you right-click the control cell (the first cell in the left corner of the table) and select Show columns. A dialog opens where you can click the columns you want to see. Right-click menu options for the table are: Option Description Refresh Causes the model data to be reread from the database. Remove Training Language... Load Examples... Training Example Editor Deletes the training language from the model. Opens the Load Responses dialog. Click the search button (the button with ...), which allows you to locate file on your computer and select it. After you find the file and click OK, a job is submitted which will load the data. You can view the job on the Jobs page. Opens the Training Example Editor for the model. Multi-Lingual Model You can set up a model to train and code in multiple languages. When working in multiple languages, the language of a particular response must be determined by a language question in the study. The response to the language question determines the language of the response. You enable multiple languages in a model by specifying a Language Question ID. A question with this ID must be present in each study used for coding with the model. If the model specifies a Language Question ID, then you are able to add multiple languages to the Languages table. Languages Table The Languages table lists the languages you will use for both training and coding with a model. It is very important that the languages specified in this list match the languages found in the responses used at training and coding time. Training or coding a German response using the English language setting will not yield good results! At least one Language must be specified for each model. If the Language Question ID property of the model, then you may specify more than one Language. 488 Ascribe™ Crossword When the Language Question ID property of the ACM model is not blank, certain rules apply both at coding time: A question with the specified ID must be present in the study for each coding question. The Language response column in the Languages table may not be blank. For each response considered at training and coding time, the respondent must have also given a response to the question specified by the Language Question ID. In addition, this response must match one of the specified Language response values in the Languages table. Otherwise, the response is ignored. Add a Language To add a Language, right-click the model's Training Examples folder in the Navigator. Choose Add New Training Language. A dialog box displays. Select the Language ID, enter the Language Response value, and enter an optional description. Provide Training Examples Desktop allows you to provide training examples for a model in a variety of ways. Copy Responses From a Study Question to the Model In this method, you display the question in the Home view and the model's Language node in the Navigator and drag responses from the question to the Language node. (It will be easier to accomplish if you turn off the Synchronize Navigator Pane with Home View Lists option. From the main toolbar, click Tools/Options. Uncheck the Synchronize option on the Environment tab.) Here are the steps: Left-click the Questions folder of the study in the Navigator. The questions in the study display in Home View. Left-click the plus sign next to the model. Left-click the plus sign next to the Training Examples folder. The Language node displays. Left-click the question in the Home View to highlight it and drag it to the Language node in the Navigator. When the plus sign appears, drop the question onto the Language node. The Copy Responses Dialog displays. Choose the appropriate options in the dialog and click OK. Here are the options in the Copy Responses Dialog: Option Description Copy All Responses As Examples This section allows you to copy all responses from a question as examples. You can also choose to copy responses that are Quality Reviewed and/or Coded Only. Select A Sample Of Responses To This section allows you to copy a temporary sample of responses from a question. (This sample will not be saved, and you cannot access it again.) You can choose how the sample is created through one of these options: 489 PrintDoc01 020813 Copy Randomly Select This Count of Responses or Randomly Select This Percentage of Responses. Copy Codes Applied If Possible If the question's codebook is the same as the model's codebook, you have the option to copy the applied codes. If the codebooks are not the same, this option is unavailable. When you copy responses from a question, a given response will be copied only if the coding source is not blank and the coding source is not an image or sound file. The criteria selected in the Copy Responses dialog also determines if a response is copied. Load a File of Responses to a Model Using the Right-Click Menu in the Navigator You can load the usual file types to a model using the right-click menu. To load a file, rightclick the Language node (under the Training Examples folder) and select Load Examples. The Load Responses dialog displays. Click the search button (the button with ...), which allows you to locate file on your computer and select it. After you find the file and click OK, a job is submitted which will load the data. You can view the job on the Jobs page. Import Examples Using the Training Example Editor The Editor Toolbar in the Training Example Editor has an option for importing examples. You can import examples from a file or from available samples. For more information, see Import Examples. Listen for Copy Using the Training Example Editor The Editor Toolbar in the Training Example Editor has an option called Listen for Copy. When you select this option, copied text will be imported into the New Examples tab, where you can edit it, code it, and add it as new examples. For more information, see Listen for Copy and New Examples Pane. Enter New Examples Using the Training Example Editor For this information, see New Examples Pane. Desktop Training Example Editor The Desktop Training Example Editor allows you to add examples to a model in a variety of ways. You can also code examples, modify coding, and view training results. The Desktop Training Example Editor looks similar to the Multiple Responses Coding tool and has similar features. Like the Multiple Responses Coding tool, the screen is divided into three sections: a top toolbar, the Examples Pane, and the Tabbed Pane. The Example Pane has two columns: Example Text and Codes Applied. The Tabbed Pane displays a tabbed view of these windows: Codebook, Training Results, Search, Word List, and New Examples. Quick Start The Desktop Training Editor works in a similar way as the Multiple Responses Coding tool. While these sections of the help system refer to "responses," the same actions apply to 490 Ascribe™ Crossword training examples. You can use these sections to see how to code using the Training Example Editor: Select Responses to Display Code a Single Response (There is a difference for the Desktop Training Example Editor. You cannot drag a code from the Codes Applied column and drop it on another example.) Code Multiple Responses Swap Codes Remove Codes For information on how to import or create training examples, see these sections: Provide Training Examples Import Examples New Examples Pane For information on how to remove examples from the model, see Delete Examples. After you run a training job, you can view the training results with the Training Example Editor. See Training Results Window for more information. Editor Toolbar The main toolbar has a number of useful functions as well as options to change the views and screen layout. Function Description Language This drop-down box lists the Languages used in the model. If only one Language is used, it will be displayed and use of the drop-down box will be disabled. Codebook This button opens the Codebook Editor and displays the entire codebook associated with the model in the Workspace portion of the application. When the codebook is displayed, any code can be dragged onto an example in the Example Text column or in the New Examples Pane to apply that code. This button also serves another function. If you float the Codebook Window and then close it, it is removed from the Tabbed Pane. To restore it to the Tabbed Pane, click this button. Word List Click this button to show or hide the Word List depending upon its current state. Training Results Click this button to show or hide the Training Results pane depending upon its current state. View Options Click this button to open the Training Example Editor View Options window. Display Examples This toolbar option provides a set of options for selecting or removing examples for display. The options are: 491 PrintDoc01 020813 Show Examples All - Displays all examples for the selected Language Coded Only - Displays all coded examples Uncoded Only - Displays all uncoded examples Clear Display - Clears the Examples Pane display Expression Tester... - Opens the Expression Tester Window to enable searches for examples containing specific text. This option allows you to import examples: From a File - Opens a dialog box where you can enter a file name or search for a file on your computer From a Sample - If there are available samples, they will be listed here, and you can copy them into the model. Import Examples Create Sample - You can also create a temporary sample from questions that share the model's codebook and import it. A dialog box opens and you can select the desired options. Note that you cannot name or save this sample. After you create the sample, it is automatically added to the examples in the model. From a Study Question - This option is not available in the Training Example Editor at this time. You can copy examples from a study question if you display the question in Home View and drag it to the model's Language node in the Navigator. See Copy Responses from a Study Question for more information. Listen for Copy The Editor will monitor the computer's clipboard for copy operations. Text copied into the clipboard from any application that enables copying selected text will be placed into the New Examples Pane. If you already have examples in the Coding Source Text column, the copied examples are added there. Otherwise, they are added in the Text window, where the text can be edited and then added to the Coding Source Text. Training Example Editor View Options This View Options Window allows the user to set various properties that control the view and display of items in the Training Example Editor. The settings established here are set for the user and will remain in place from session to session. Option Description Filters This section of View Options allows filtering by review status: any review status, not reviewed, and reviewed only. This setting determines how the responses listed in the left pane will be sorted. The options are: By Response (the verbatim text ) - sorts alphabetically Sort Order By Text Length - sorts based upon the length of the coding source text. By Codes Applied - sorts based upon the number of codes applied. 492 Ascribe™ Crossword The sort direction (either Ascending or Descending) can be specified by choosing either the ' Asc' or ' Desc' options. These options determine what information in addition to code description will be displayed in the Codebook window: Codebook Display Show Details - If checked, counts of the codes applied will display after the code description Show Code Values - If checked, the code output ID will display before the code description. Show Net Values - If checked, the net output ID will be display before the code description. Respondent ID Color There are two color choices here. One establishes the color used to render example text when the example has not been marked as quality reviewed. The second color selection determines the example text color when the example has been marked as quality reviewed. The colors can be changed by clicking on the appropriate color box and choosing a new color from the Color window. Highlight Matched Text If checked, text that matches the Expression Test search will be highlighted using the selected color. The highlight color can be changed by clicking on the color box and choosing a new color from the Color window. Examples Pane The Examples Pane List, located on the left side of the Training Example Editor, consists of two columns. The left column displays the example text. The right column displays codes applied to the example. Context menu options (right-click menus) provide access to the functionality available to manipulate this list. You can code or change the coding of the training examples in this pane. Drag codes from the codebook to the example, just like you do in the Multiple Responses Coding tool. The Examples Pane List context menus provide options for selecting, editing, and moving examples as well as clearing selections and managing the codes assigned to the examples. Right-Click Menu for Example Text Column Here are the menu options: Option Description Find Code After you highlight some text and right-click it, you can search the codebook for that text. The first instance of that text in the codebook is highlighted in the Codebook Window of the Tabbed Pane. The Find box in the Codebook Window will also contain the search text so you can continue searching the codebook by using the Find box. Split Text as New Examples This option, when chosen, will split the existing example into a series of new examples. The split process is based upon normal sentence punctuation. When you select an 493 PrintDoc01 020813 example and choose this option, the split example displays in the New Examples window of the Tabbed Pane. From that pane, you have options to edit the example or add it as a new example. Add Selection as New Example After you highlight some text in an example and choose this option, the text displays in the New Examples window of the Tabbed Pane. From that pane, you have options to edit the example or add it as a new example. Remove Selected Examples This option deletes selected examples from the model. Move to Different This option moves the selected response to another Training Language Language (only available if the model has more than one Language.) Edit Example This option causes the example to display in the New Examples window where you can edit it. You can also edit an example by double-clicking it. Clear All Examples Clears (removes the highlight from) all selections in the Example Text column. Select All Examples Selects all examples in the Example Text column. Hide Selected Examples Hides selected examples. Display Hidden Examples Displays all previously hidden examples. Remove All Codes from Selected Examples Removes all the codes from the selected examples in the Example Text column. Right-Click Menu for the Codes Applied Column The context menu options for managing the codes assigned (right clicking the right-side column) are described below. Option Description Clear All Selections Clears (removes the highlight from) all selections in the Codes Applied list. Select All Same Code Selects all codes in the Codes Applied list that are the same as the current selected code. Remove Selected Codes Removes the selected codes from the list of codes assigned to the examples in the Example Text column. Delete Examples With the Desktop Training Example Editor, you can delete examples from the model. You can remove them one at a time or remove multiple examples at one time. 494 Ascribe™ Crossword First, you must have the example or examples displayed in the Examples Pane. (You have multiple ways to display examples. You can use the Display Examples option on the toolbar to display all coded, all uncoded, or all examples. You can use the Search option to display only examples that contain the search criteria. Or you can right-click a code in the Codebook Window, choose Select Responses/Match any selected code to display only examples that have been coded with this code.) To remove one example, right-click it and select Remove selected examples. To remove multiple examples at one time, select all of the examples. (You have multiple ways to select responses. You can use the Ctrl-Click or Shift-Click methods. If you want to remove all of the examples in the Examples Pane, right-click one of the examples and choose Select all examples.) Once the codes are highlighted, right-click one of them and select Remove selected examples. Training Example Editor Tabbed Pane The Tabbed Pane on the right side of the Training Example Editor displays various functional windows based upon the user's navigation choices. The Tabbed Pane can optionally be hidden. To hide, right click in an empty area of the top toolbar. Uncheck the Display Tab Pane option. When the Tabbed Pane is hidden, the windows described below will appear as "floating." Codebook Window (Training Example Editor) The Codebook Window displays the codebook for the model. Although similar in appearance to the Codebook Editor, this display of the codebook does not allow editing of the codebook itself. You can drag codes from this window to the examples to code them. For more information, see Quick Start. Several menu options are available for using the codebook and controlling the display. Option Description Expand All Expands all nets in the codebook to display each net's children. Collapse All Collapses (hides) all but the top level codes and nets. Provides a list of options that control display of the codebook. View Options... Display help available symbol - If selected, a symbol appears in front of the code description when code help is available. Codes only - If selected, only codes are displayed in alphabetical order. Displays options for selecting examples to display in the Examples Pane. Select Responses... Match any selected code - If selected, examples with any of the currently selected codes applied to them are displayed. Match all selected codes - If selected, only examples with all the currently selected codes applied to them are displayed. 495 PrintDoc01 020813 Expression Test... Opens the Expression Tester Window with the selected code's regular expression text. Add Training Examples This option allows you to add training examples for a specific code. When you select this option, the New Examples window displays. Click the plus sign, which opens the Text window. Enter a new example, and press Enter. The selected code is automatically applied. Click Accept to add the example to the model. Add Code Text as Example This option adds the text of the selected code as a new example. When you select this option, the New Examples window displays. The example displays, and the code is automatically applied. Click Accept to add the example to the model. Training Results Window To review the results of an ACM training job, right-click the model and select Training Example Editor. From the Editor, select Training Results from the Tabbed Pane. This window displays the training results. It is important to review these results before marking a model as approved. Here are the columns on this page: Column Description Code The code that was trained. Double-click a code to see the examples that were used to train the code. The examples display in the Example Text column along with the codes that were applied. Responses The number of examples used to train the code in this language. Accuracy The estimated accuracy for this code, using the F1 metric. Trend The trend in the estimated F1 accuracy. New Examples Pane The New Examples Pane allows you to enter and manipulate new examples. It has two sections: the Text window in the upper section, and the Codes window in the lower section. The initial view may display the Codes window with the Text window closed. To open it, click the plus sign on the toolbar. The Text window displays either the current example selected in the Examples Pane or is blank. The Codes window allows you to code the new examples. Enter a New Example in the Text Window To enter a new example, click the plus sign ( ) on the toolbar. The Text window opens, and you can type in a new example. When you press Enter, the example is moved from the Text window to the Codes window. When you are finished typing new examples, click the plus sign again to close the Text window. 496 Ascribe™ Crossword The examples are uncoded at this point, but you can code them here or in the Examples pane. If you do not want to code them here, click the Accept button to add the examples to the model or use the right-click menu options to manipulate the examples. Right-Click Menu Options for the Text Window All or portions of the text in the Text window can be selected. The selected text can be added as a new response, either by dragging it to the Coding Source Text column or by using a menu option available from the window's context menu described below. Option Description Edit Selection As New Example Allows you to edit all or part of the text in the Text window. When editing is complete, press Enter to move the edited text to the Coding Source Text column. Add Selection As New Example Split Text As New Example Allows you to add the selected text as a new example. To use this option, select part of the text that displays in the Text window. Right-click the selected text and choose this option. The selected text is copied to the Coding Source Text Column, where you can code it or accept it as a permanent addition to the training examples. This option will split the existing example into a series of new examples. The split process is based upon normal sentence punctuation. Code New Examples in the Codes Window To code a new example, open the Codebook Editor in the workspace (click the Codebook icon on the toolbar) or float the Codebook Window from the Tabbed Pane (right-click the Codebook tab and select Float.) Select an example in the Coding Source Text column and highlight it. Drag a code from the codebook to the Codes Applied column and drop it. When you finish coding, click the Accept button to add the examples to the model or use the right-click menu options to change the codes or examples. The examples do not have to be coded when you add them to the model. Right-Click Menu Options for the Codes Window The context menus for the Codes window give you more options for working with new examples. Both the Coding Source Text column and the Codes Applied column have rightclick menu options. Here are the options for the Coding Source Text column: Option Description Select All Examples Selects and highlights all of the examples in the Coding Source Text column. Clears the highlight from all examples in the Coding Source Clear All Examples Text column. Remove Codes From Selected Examples First, you should select the example or examples from which you want to remove codes. Then right-click one of the selected examples and choose this option. The codes are 497 PrintDoc01 020813 removed from the Codes Applied column. Remove Selected Examples From List First, you should select the example or examples that you want to remove. Then right-click one of the selected examples and choose this option. The selected examples are removed from the Coding Source Text column. Remove All Items From List Removes all examples from the Coding Source Text column, whether the examples are selected or not. Add Selected Items To Training Examples First, you should select the example or examples you want to add to the model's training examples. Then right-click one of the selected examples and choose this option. The selected examples are added to the training examples, making it a permanent addition. If any codes have been assigned, the codes will be applied to the examples. Add All Items As Training Examples Adds all items in the Coding Source Text column to the model's training examples, whether the examples are selected or not. This process makes the examples permanent additions to the training examples. If any codes have been assigned, the codes will be applied to the examples. Here are the right-click menu options for the Codes Applied Column: Option Description Clear All Selections Clears the highlight from all selections in the Codes Applied column. Select All Same Code Remove Selected Codes Selects all codes in the Codes Applied column that are the same as the currently selected code. Removes the selected codes from the Codes Applied column. Use the Accept Button to Add Examples The Accept button on the toolbar adds all examples in the Coding Source Column as a permanent addition to the model's training examples. If any codes have been assigned, the codes will be applied to the examples. If you do not want to accept all examples in the Coding Source Column, highlight the examples you want to add and use the right-click menu to only add those examples. Use Listen for Copy to Create New Examples The Listen for Copy option can be used to create new examples in the New Examples Pane. Text copied into the clipboard from any application that enables copying selected text will be placed into New Examples Pane. If you already have examples in the Coding Source Text column, the copied examples are added there. Otherwise, they are added in the Text window, where the text can be edited and then added to the Coding Source Text. Create New Examples from Current Examples 498 Ascribe™ Crossword Click the New Examples tab to open the New Examples Pane. Click the plus sign to allow entry of new examples. Display some examples in the Examples Pane, using whatever criteria you would like. Click an example, and it will display in the Text window. You can edit the text here and press Enter to move it to the Coding Source Text column. Note that you are not changing the original example; you are copying it and creating a new example. Once the new example is in the Coding Source Text column, you can use the right-click options or click Accept to add it to the training examples. See Right-Click Menu Options for the Text Window for more information. Split Text as New Examples This option is available as a right-click menu option in both the Example Text column in the Examples Pane and the Text Window in the New Examples Pane. When chosen, this option will split the existing example into a series of new examples. The split process is based upon normal sentence punctuation. When you select an example in the Examples Pane and choose this option, the split example displays in the New Examples Pane. From here, you can code it and add it as a new training example. ACM Coding The automated coding model is ready to code questions after it has been trained and approved. You also need to associate the question with the model before you start the coding job. After you submit the coding job, you can view it on the Jobs page. When the job finishes, you can look at the coding with the Review ACM Coding window. Associate the Coding Question with the Model You also have to associate the question with the model by doing two things: Share codebooks - The codebook of the coding question must be shared with the codebook used by the model. To share a codebook, the coding question should be in the Data Pane/Home View. Select the question so it is highlighted. Place your cursor on the model in the Navigator. Drag from the model to the question in the Data Pane. The Copy/Share dialog displays; select Share and click OK. Add the coding question or questions to the model - Drag and drop from the Question List in the Data Pane to the Coding Questions for the model in Navigator. Start the Coding Job You can start the coding job in two ways. In the Navigator, you can find the coding question under the Coding Questions folder for the model. Right-click the question and select Start coding job. Or, you can display the Coding Questions table in the Data Pane by double-clicking the Coding Question folder under the model in the Navigator. The Coding Questions table lists all of the coding questions for the model. Right-click a question and select Start coding job. Coding Questions Table 499 PrintDoc01 020813 This table lists all questions that may be coded using this model. (You see this page in the Data Pane when you click a Coding Questions folder for a model.) A given question may appear in the Coding Questions table of no more than one model. You can remove questions from a model at any time, whether or not the model is approved. To remove a question, right-click on the question and select Remove selected questions for ACM coding. The columns in this table are: Column Description Study ID The name of the study where the question is found. Question ID The question ID Question Label The question label Codebook Mapping Unless all of the questions you will use at ACM coding time share the same codebook, you will need to specify a way to match codes in the codebooks. See the Match Codes By field in the Edit ACM dialog. If the model is not approved, you can change this value in the Edit ACM dialog. Responses The number of responses in the question. Responses Coded The number of responses that have codes applied. ACM Coding Date If an ACM coding job has been run, this column shows when the job was run. Coding Job Status This field displays a status of Completed if the coding job has run successfully, and a status of Failed if it did not run successfully. Click Completed or Failed to see the job details. You can choose what columns display if you right-click the control cell (the first cell in the left corner of the table) and select Show columns. A dialog opens where you can click the columns you want to see. Right-click menu options are: Option Description Refresh Causes the model data to be reread from the database. Edit Opens the Questions Properties dialog for editing. Remove Selected Questions for ACM Coding Deletes a question from the list of coding questions for a model. Start Coding Job Starts the ACM coding job for this question. The model must be approved; otherwise, the option is disabled. Review ACM Coding Opens the Review ACM Coding window; the option is disabled if the question has not been coded with the ACM coding process. 500 Ascribe™ Crossword Codebook Editor Opens the Codebook Editor. Codebook Merge/Map Tool Opens the Codebook Merge/Map tool. Coding Time Options After you select to start the coding job, a dialog box displays. You can choose from these options: Option Description All Responses With this option, the coding process will suggest codes for all responses, even if the responses have codes applied. It does not uncode and recode responses, but will suggest new codes if applicable. Uncoded Only With the option, the coding process will suggest codes only for uncoded responses. Apply Codes to Uncoded Responses This option is only available if you selected Allow Auto Code in the Edit ACM dialog and set a minimum confidence. If you allow auto coding and select to apply codes to uncoded responses, suggested codes that meet the minimum confidence are applied. If the suggested codes do not meet the minimum confidence, they are not applied. You can apply suggested codes in the Review ACM Coding window. For more information about confidence scores, see Coding Using Binary Classifiers. Review ACM Coding When the coding job completes, use the Review ACM Coding window to review and accept the coding suggestions of the model. Review ACM Coding Window In this window, you review the coding suggested by the model. You also have the option to accept the coding suggestions. You cannot change the suggested codes or the applied codes in this window. To change the coding, use the Multiple Responses Coding tool. The Review ACM Coding window is similar in format to the Multiple Responses Coding tool and the Training Example Editor. It has three sections: a toolbar, the Response Pane, and the Tabbed Pane. The window also has a View Options dialog. If you use the Paging Options, page numbers display under the toolbar. The numbers change color as you view each page. Dark blue indicates the page you are viewing. Green indicates pages you have viewed, and aqua indicates you have not view that page. If you right-click a page number, you have the option to clear visited pages and return the page number colors to aqua. Accept Suggested Codes Use the Display Responses option in the toolbar to view the responses and the suggested codes. If the codes appear in gray, they have already been applied. (They were applied 501 PrintDoc01 020813 because Apply Codes to Uncoded Responses was selected at coding time or because a previous coding job may have been run.) Also note that the display of the responses may depend on your choice of the ACM Suggestion Confidence Threshold. As you review the coding, you can accept the suggested codes in two ways: Use the button on the toolbar to apply all displayed codes. Select one or more responses in the Response Pane. To select multiple responses, hold the Shift or Ctrl key as you click the responses. When you have selected the desired responses, right-click one of the responses and select Apply codes for selected responses. This action applies all of the codes suggested for each selected response. old You do not have accept any or all of the suggested codes. If you do not accept them, they are not applied. To change the coding of a response, use the Multiple Response Coding tool. ACM Suggestion Confidence Threshold The confidence setting appears both at the top of the Codes column and in the View Options dialog. They control whether a given code appears in the Codes column. Recall that during the ACM coding job, a confidence number is assigned to each code in the codeframe for each response. See How the Verbatim Coding System Works for more information about the confidence number. These confidence numbers range from -100 to 100. A score of -100 means the ACM is very confident that the code should not be applied to the response, and 100 means that the ACM is very sure that the code should be applied to the response. Numbers closer to zero mean the ACM is less confident. The value of zero is special. It is never assigned as a confidence number (you will not find any codes with a confidence number of zero). The ACM has been carefully tuned so that if all codes with a confidence number above zero are applied, you will likely obtain the accuracy of coding predicted at training time. The confidence number displays in parentheses after each suggested code in the Codes column. You can change the threshold in the View Options dialog. If a confidence score is equal to or above the Suggestion threshold, it appears in the Codes column. Toolbar The toolbar has these options: Option Description Apply All Displayed Codes Click this button to apply suggested codes to the responses that are currently displayed. Word List View Options 502 Click this button to show or hide the Word List depending upon its current state. Click this button to open the View Options for ACM Review Ascribe™ Crossword Coding window. This toolbar option provides a set of options for selecting or removing examples for display. The options are: Show All Responses All - Displays all responses Display Responses Coded Only - Displays all coded responses With Unapplied Codes Only - Displays responses that have suggested codes which have not been applied Clear Display - Clears the Response Pane display Expression Tester... - Opens the Expression Tester Window to enable searches for examples containing specific text. Response Pane The Response Pane has two columns: the Coding Source Text column and the Codes column. The Coding Source Text contains the responses that were coded. The Codes column contains the applied and suggested codes for the response. The Response Pane also has a control cell. The control cell is the top left-most cell next to the first column. The right-click menu options for the control cell are: Option Description Clear All Selections Clears any selected fields in the table. Clear All Filters Clears any filters. Show Columns Displays a list of available columns; click the box next to each column to display or hide a column. Show All Columns If some columns are not displayed, click this option to show all columns. (This option is grayed out if all columns are currently displayed.) Resize You can manually resize columns by dragging the column lines. To restore the columns to their default size, select Resize and either Column Headers, Rows, or Entire Table. Restore Default Column Order You can manually change the order of the columns by dragging the column headings. To restore the columns to their original order, select Restore Default Column Order. Copy Text This option selects and copies the responses on the current page. You need to have Word or Excel ® open first and then select Copy Text. When you switch to Word or Excel ®, rightclick and select Paste. Paging Options As a default, paging is enabled, and the responses will page in increments of 100. The paging options stick from session to session. The paging options were included to help with data sets that are fairly large and take time to render. The options are: Enable Paging - Click this box to turn the paging off and on. If you turn paging off, all of the responses display on one 503 PrintDoc01 020813 page. Page Size Options - Items per Page defaults to 100. You can enter a different number. Select the Percentage per Page option to display a percentage of the responses on each page. When you use the Percentage per Page option, you can enter numbers in the Minimum and Maximum fields to control how many responses display on a page. If you have paging enabled, you can see the total number of responses when you hover over one of the page numbers. The Coding Source Text column has the following right-click menu options: Option Description After you highlight some text and right-click it, you can search the codebook for that text. The first instance of that text in the codebook is highlighted in the Codebook Window of the Tabbed Pane. The Find box in the Codebook Window will also contain the search text so you can continue searching the codebook by using the Find box. Apply Codes For Selected Responses Clear All Responses Un-selects all currently selected responses. Select All Responses Selects all responses in the list. Hide Selected Responses Hides selected responses. Display Hidden Responses Displays all previously hidden responses. The Codes column has the following right-click menu options: Option Description Clear All Selections Unselects all selected codes. Select All Same Code Selects all codes that are the same as the current code. Apply Selected Codes Applies the currently selected codes to the responses. View Options for ACM Review Coding Window This View Options Window allows the user to set various properties that control the view and display of items in the Training Example Editor. The settings established here are set for the user and will remain in place from session to session. Option Description Confidence This section allows you to set the suggestion confidence threshold for suggested codes. Move the sliding buttons to set an upper and lower range. You may want to set the lower range to correspond to the Minimum Confidence number in the Edit ACM dialog. Sort Order This setting determines how the responses listed in the left pane will be sorted. The options are: 504 Ascribe™ Crossword By Respondent ID By Question ID, Respondent ID (useful with shared codebook questions) By Response (the verbatim text ) By Transcription text By Translation text By Notes text By Text Length - sorts based upon the length of the coding source text. By Codes Applied - sorts based upon the number of codes applied. The sort for all but text length and codes applied are done alphabetically. Since many of the transcription or translation values can be empty, the sorting could appear to produce unusual results when those options are selected. The sort direction (either Ascending or Descending) can be specified by choosing either the ' Asc' or ' Desc' options. These options determine what information in addition to code description will be displayed in the Codebook window: Codebook Display Show Details - If checked, counts of the codes applied will display after the code description Show Code Values - If checked, the code output ID will display before the code description. Show Net Values - If checked, the net output ID will be display before the code description. Respondent ID Color There are two color choices here. One establishes the color used to render example text when the example has not been marked as quality reviewed. The second color selection determines the example text color when the example has been marked as quality reviewed. The colors can be changed by clicking on the appropriate color box and choosing a new color from the Color window. Show Respondent ID If checked, the respondent ID is displayed with the coding source text. Highlight Matched Text If checked, text that matches the Expression Test search will be highlighted using the selected color. The highlight color can be changed by clicking on the color box and choosing a new color from the Color window. Tabbed Pane for Review ACM Coding Window The Tabbed Pane has these windows which are similar to the ones in the Multiple Responses Coding tool and Training Example Editor: Search, and Word List. The Tabbed Pane also has a Codebook Window which is similar to the one in the Multiple Responses Coding tool. However, you cannot apply new codes from the Codebook in the Review ACM Coding window; you can only accept the suggested codes. Also, you cannot open the Codebook Editor from the Review ACM Coding Window. 505 PrintDoc01 020813 The Tabbed Pane has one unique window: the Accuracy List. This window displays each code, the number of distinct training examples, and the accuracy and the trend from the training process. Ascribe™ Crossword Navigate: Client/Crosswords Ascribe™ Crossword provides tabular reports of coded studies, and allows you to drill down to examine the verbatim comments. It is intended to help you explore the qualitative content behind the coded results of your studies. With Ascribe™ Crossword, you can create and publish tables. You select row questions and column questions (usually socio-demographic questions) and look at the frequencies. Each cell in the table represents the number of respondents that satisfy the two criteria of the row code and the column code. You can save your tables for later use, and can even send a colleague a link to a table, so he or she can view it with a single click. For more information, see: Installing Crossword You can install the Crossword application on your computer from Ascribe™. In Ascribe™, navigate to the Home page, and select Downloads. In the Downloads page, click the Ascribe™ Crossword link. Once Crossword is installed, you can start it from the Windows Start menu. You do not need to be logged on to Ascribe™ to run Crossword. Or you can access Crossword from Client/Crosswords or Supervisor/Studies/Reports/New Crossword. There are some system requirements for using Crossword. Ascribe™ Crossword runs on Microsoft Windows operating systems only. It is not available for 64 bit platforms. It requires Microsoft Windows 2000, XP, Vista, or Windows 7. Ascribe™ Crossword requires the Microsoft .Net Framework version 2.0. If this is not installed on your computer, it will be installed automatically when you install Crossword. This installation can be time consuming. If you do not yet have the Microsoft .Net Framework version 2.0 installed, you may want to perform the installation of Crossword when your computer is idle. A connection to the Internet is required both installing Crossword and running Crossword. You cannot run Crossword without an active Internet connection. Crossword Logon When you start the Crossword application, you will be presented with a logon dialog. Enter your logon credentials for Crossword just as you would for Ascribe™. To logon to Crossword, you must have a valid Ascribe™ account. Such accounts must be created and managed in Ascribe™. You cannot create Ascribe™ accounts in the Crossword application. 506 Ascribe™ Crossword If your password has expired, you must renew your password in Ascribe™. You cannot renew your password in the Crossword application. After you logon to Crossword, you will be presented with the Open Crossword dialog, if you have any saved Crossword tables. Otherwise, you will be presented with the New Crossword Wizard. Understanding a Crossword Report A Crossword works with respondent sets. The starting point is the set of respondents who gave a given answer to a question. This is, of course, the same as saying the set of respondents who were coded with a given code to that question. A Crossword is constructed from dimensions, with each dimension containing a list of codes applied to a question. Dimensions A Dimension is a list of codes and nets. As you might expect, you will normally create a dimension from the codebook for a question. When you do this, the dimension initially looks just like the Ascribe™ codebook. You can change dimensions, rearranging the codes and nets, and changing the text and colors. When you do this, you never modify the Ascribe™ codebook. This is a fundamental rule of the Crossword. You cannot change anything in the Ascribe™ study using the Crossword, but you can change the way the data displays. These changes are confined to the Crossword, and do not affect the underlying data in the study. In a dimension, we call each net or code generically a node. If the node is a code, it is associated with a question and a code. Its respondent set contains those respondents who replied to the question and were coded with that code. If the node is a net, its respondent set is the union of the respondent sets for each of its child nodes. Hence, a net represents the set of unduplicated respondents to all subordinate codes. The Crossword Table The table displays some or all of the cells in the dimensions. The value displayed in a cell of the table is the number of respondents in the respondent set for that cell. This can be displayed as a simple count, or as a percentage. The table is two dimensional. You can control the way the cells in the cube are arranged in the two dimensional table by working with the dimensions and the properties of the table. Crosswords Page Navigate: Client/Crosswords This page lists the Crossword reports available to you. In the filter bar, you can enter a filter to search for Crosswords by the name of the Crossword or the study ID. The filter will select items that start with the characters you type. Type * as the first character to find items that contain the characters following the *. The items visible to you depend on the type of your user account. If you have administrator privilege, you can see all of the Crossword reports in the list. You can delete any Crossword. You can open your Crossword reports, and those that are marked Public. 507 PrintDoc01 020813 If you are an associate, but do not have administrator privilege, you can see your Crosswords and those that are marked Public, and can open any of these reports. You can delete only your Crosswords. Contacts do not have access to this page, but can view Crosswords created by others for studies assigned to their Company. To do so, they should start the Crossword application from the Windows Start menu. To delete a Crossword, right-click and select Delete.... You can delete only your own Crosswords unless you have administrator privilege. To add a new Crossword, right-click and select New... Enter the ID of the study for which you want to make a Crossword report. Click OK, and you will be redirected to the New Crossword page. To view a Crossword, right-click and select Crossword. If you are an administrator and view a Crossword that is not Public and is not your own Crossword, the Crossword program will not display the report. Open Crossword To open a saved Crossword, click the Open button in the toolbar. The list displays all of the Crosswords you have saved, and also all Crosswords that have been saved by others with the Public option selected. Select the Crossword you want to open, then click the OK button. New Crossword This page allows you to construct a new Crossword quickly. The Crossword program allows you to create more complex reports than can be created in this page, but this is a good starting point for learning how to use the Crossword program. The list at the left displays the questions in the study. You can add questions to the side of the report, or along the top of the report (the "banner"). To add questions, select one or more in the list on the left and click the Side or Banner button. To remove questions from the side or banner, select them and click the Remove button. When you have selected your questions, click the OK button to create the Crossword. If you wish to view the Crossword immediately, check the Start Crossword box before you click the OK button. You can give your new Crossword a name by entering the name in the Name box before you click the OK button. If you do not enter a name, a default name containing the study ID will be used. New Crossword Wizard To create a new Crossword, click the New Crossword Wizard button in the toolbar. The wizard guides you through creation of a simple Crossword table. Note that when you create a new a new Crossword when an existing one is opened, the properties of the existing ones will apply to the new one. 508 Ascribe™ Crossword Select Study for Crossword Select the study whose data you want to view in the Crossword table. Click the Next button. Alternatively, you may click the Finish button after selecting your study. In this case, a Crossword will be created with an empty table and no Dimensions. You can then construct Dimensions using the New Dimension Wizard. Select Row Question(s) Select the question(s) you want to appear in the rows of the Crossword. The codebook for these questions will be used to construct dimensions, and these dimensions will appear in the rows on the left side of the table. You can view the codebook of a question by doubleclicking it or by right-clicking and selecting "View Codebook" from the popup menu. Click the Next button to select one or more questions for the columns of the Crossword, or click Finish if you do not want questions to appear in the columns. To select more than one question, press Control while you click the questions. Select Column Question(s) Select the question(s) you want to appear in the columns of the Crossword. You can view the codebook of a question by double-clicking it or by right-clicking and selecting "View Codebook" from the popup menu. Click Next or Finish. Select Hidden Question(s) Select the hidden question(s) you want to use for filtering the questions displayed in the crossword or that you may want to display at a later stage. (The use of filters is optional.) In order to select all the remaining questions that were not selected as row questions or column questions, you can right-click and click "Select all remaining Dimensions". You can view the codebook of a question by double-clicking it or by right-clicking and selecting "View Codebook" from the popup menu. Click Finish. Troubleshooting Your table will be empty under either of these conditions: One or more of the questions you selected has no coded responses. If any dimension in the Crossword has no respondents, the table will always be empty. There are no respondents with coded responses to either of the questions you have selected. A respondent will appear in the table only when it has a coded response in each of the displayed dimensions. Save Crossword To save the current Crossword, click the Save button on the toolbar. You can save a Crossword only if you are the user who created the Crossword. If you are viewing a public Crossword created by another user, you cannot save it. You can, however, use the Save As command to save it as a new Crossword. If you do this, you are the owner of the newly saved Crossword. 509 PrintDoc01 020813 If the Crossword does not have a name, the Properties dialog is displayed. Enter a name in the dialog and click OK to save the Crossword. Here are the options to save a Crossword: Save As - To save the open Crossword as a new Crossword, click the drop-down next to the Save button on the toolbar. Assign a new name to the Crossword in the properties dialog, and click OK. Save As HTML... - You can save the current crossword table in a simple html format by right-clicking it and selecting "Save as HTML..." from the popup menu. Export to Microsoft Excel – You can export the current crossword to Excel by rightclicking it and selecting “Save as Excel…” from the popup menu. Crossword Table Right-Click Menu The row and column headings of the table display dimensions. You can control the structure of these dimensions using the Show Dimensions toolbar item. You can also control display options for the table and using the Properties dialog. When you left-click in the table, the selection highlight shows the cells spanned by the respondent set for the cell. If you click a cell that is at the intersection of two codes, only the single cell is highlighted. If you click a cell that is at the intersection of one or more nets, the cells that comprise those nets are highlighted. Right-click anywhere in the table to display a menu of commands for the table. These commands are described below. Option Description Show Responses for Selected Cells… To view the responses for the cells selected in the table, right-click and select Show responses for selected cells... The Responses Dialog box displays, unless there are no respondents selected. You also can view the responses for a single cell by double-clicking the cell. You can hold the Ctrl key and left-click with the mouse to customize your selection. When you Ctrl-click a cell that is already selected, the selection is removed for that cell only. Properties… Select Properties... from the menu to display the Properties dialog. If you are viewing a published Crossword, this option is available only if the publisher has enabled it. (The Properties dialog also displays when you attempt to save a Crossword, if you have not assigned a name to the Crossword.) Transpose Click the Transpose option in the menu to move swap the Dimensions in the rows and columns of the table. Calculate Click the Calculate button to refresh the data in the table. This action is generally required when you change a dimension. This menu option is identical to clicking the Calculate button on the toolbar. 510 Ascribe™ Crossword Copy Table Click the Copy table option to copy the table to the clipboard. The table is copied in HTML and text formats. This allows you to paste it to applications that accept HTML (such as Excel), or to applications that accept text (such as Notepad). Resize Row Heights This option resizes the height of the rows in the table, based on the contents of the cells. To adjust the width of the row header, drag the border of the header. Resize Column Widths This option resizes the width of the columns in the table, based on the contents of the column headers and the cells. To adjust the width of an individual column, drag the border of the column header. Graph This lets you show the frequency of the first (total sample) column in a bar chart style graph. Save as HTML This saves the current table in HTML format to a file. Save as Excel This saves the current table in an Excel file. t-Tests Between Proportions See the section t-Tests Between Proportions for more information. The status bar displays the name of the Crossword (which can be set in the Properties dialog), and the study ID and study name. It also displays the number of rows and columns in the Crossword. Responses Dialog This dialog is displayed when you select Show responses for selected cells... in the Crossword table right-click menu. The number of responses is displayed on the top left side of the window. Also, when resizing a row manually all the other rows are resized to the same size. The toolbar commands are: Close - Closes the dialog. Show questions - Displays a list of the Questions in the Study. The list selection tracks the question for the current response in the responses list. Show details - Displays the verbatim and codes applied to the selected response. If any response has a translation, transcription, or notes, a box is also displayed for the applicable items. Filtering You can filter the table by right clicking and select "Filter..." from the popup menu. A dialog box will appear and let you filter the data in any way you want. The drop-down lists on the left let you apply regular expression patterns to the different text fields. To apply a regular expression, type on top of one of the textboxes and press Enter. If you do the same operation again, the two regular expressions will be "added", meaning that they will be " ANDed" together. Whenever a filter is applied, the column headers will turn yellow. To remove the filter(s), right-click and select "Remove Filter(s)". Saving 511 PrintDoc01 020813 If you click the "Save" button on the top left side, you will be able to save the list of responses in either a tab delimited text file, XML file or HTML file. Crossword Properties Dialog Here are the fields in the Crossword Properties dialog: Field Description Name The name of the Crossword. A name is required to save a Crossword. Crossword names need not be unique. The name appears in the Open Crossword dialog when you open a saved Crossword. Description A description of the Crossword. A description is not required. The description appears in the Open Crossword dialog when you open a saved Crossword. Access These options determine access for others to this Crossword. Changes you make to the access of a Crossword have affect on other users only after you save the Crossword. The options are: Public - If a Crossword is public, it can be opened by other users in your Ascribe™ account. Although other users can open the Crossword, and can modify their local copy of the Crossword, only the user who created the Crossword can save it. Published until - Check this box to publish your Crossword. The publication will not be effective until you save the Crossword. When you clear this box in a published Crossword and save it, the publication is revoked. Once you revoke a publication any links generated for that publication become invalid. Should you again publish the same Crossword, new links can be created for the Crossword; however, links created for the previous publication remain invalid. The publication will expire after the date you set in the date control. Permissions - Specify the access rights within a published Crossword using these check boxes: Dimensions - The user can create new dimensions and change existing dimensions. Properties - The user can access this Properties dialog. Cell Content Display -If "Distinct Responses Count" is selected, the count of respondents appears in each cell. If the count is zero, the cell is blank. If "Responses Count" is selected, the count of respondents appears in each cell as well but the difference is for nets. Net cells will represent the sum of the cells that they contain rather than unduplicated counts. In other words, if the same respondent gave two codes that belong to the same net, he or she will be counted twice in the net count. Percentages - Specify whether to display a percentage in the cell as well as what cells the percentages should be based on. It is also possible to specify the number of decimals used for percentages (from 0 to 2). Output Code Value - If checked, the output code values will be 512 Ascribe™ Crossword displayed in front of row codes. Display Codes and/or Nets - You can ask Crossword to display only codes, only net or both (default) in the rows. Coded and/or Uncoded Respondents - You can ask Crossword to display two special rows. "Total Coded Respondents" represents the number of respondents who gave at least one answer that was coded in the current question. "Total Uncoded Respondents" (which appears at the bottom of each question) represents the number of respondents who did not give at least one answer that was coded. Minimum Percentage Difference to Highlight Cell - If you are displaying any kind of column percentages, you can ask Crossword to highlight cells whose percentage differs by a minimum amount from the total sample. Cells whose percentages are higher will be displayed in light blue, those whose values are lower than the average sample will be displayed in green. Leave the value to 0 if you do not want to use this feature. Hide These check boxes control whether rows and columns that contain no entries in any of the cells are displayed. Color You can change the colors of the titles, the nets and/or the codes of all the dimensions of your Crossword at once by using any of the 3 buttons available to you. Nets You can ask Crossword to display nets in bold and/or to display net rows in gray for greater visibility. The deeper the net level (sub nets and sub-sub nets...), the lighter the gray. Sort You can sort the Crossword table by frequency. Nets will be sorted first and then codes within nets as well. Show Dimensions You can manipulate the dimensions of the Crossword by clicking the Show Dimensions toolbar button. When the button is clicked, the Dimension Editor appears to the left of the table. Note that you can copy a code node and insert it multiple times in a dimension. This is in contrast to an Ascribe™ codebook, in which a given code appears in one location only. Dimensions List The Dimension list appears at the top of the Dimension Editor. The list shows each of the dimensions. The check box on a dimension controls whether the dimension is displayed in the dimension view below the list. Right-click a dimension to display a menu of commands for the dimension. These are: Field Description Add to Rows This option appears only if the dimension is not currently displayed in the table. Click this option to display the dimension in the rows of the table. Add to This option appears only if the dimension is not currently displayed 513 PrintDoc01 020813 Columns in the table. Click this option to display the dimension in the columns of the table. Add as Grid This option appears only if the dimension is a "Codebook Dimension". This means that you added it by selecting "Add Dimension from a Codebook" in the New Dimension Wizard. Click this option if you want to see all the questions that share that codebook appear next to each other. Move to Rows This option appears only if the dimension is currently displayed in the columns of the table. Click this option to move the dimension to the rows of the table. Move to Columns This option appears only if the dimension is currently displayed in the rows of the table. Click this option to move the dimension to the columns of the table. Remove from Table This option appears only if the dimension is currently displayed in the table. Click this option to remove the dimension from the table. The dimension is not deleted, but becomes a hidden dimension. Hidden dimensions act as filters