Lecture №18 Nursing Research PART IІ Types of Non-probability Sampling Convenience (Accidental) Sampling Quota Sampling Purposive Sampling Network Sampling Theoretical Sampling Non-Probability Sampling Theoretical Sampling Quota Purposive Sampling (Non-Randomized) Convenience Sampling Network Caution Areas on Data You see what you look for You look for what you know Appropriate statistical strategies for certain types of numbers If you are a hammer, the world looks like a nail Dealing With Data (ch. 11) Developing Data Collection Forms Planning Data Collection Process Planning he Organization of Data Planning Data Analysis Planning Interpretation & Communication of Findings Evaluation of the Plan Data Collection Tasks Recruiting Subjects Maintaining Consistency Maintaining Controls Protecting Study Integrity Problem-Solving Physiological Measures: Reliability and Validity Accuracy Selectivity the amount of reproducibility in measurement Sensitivity the ability to identify that which is really want to sometimes called specificity Precision measurement that has the most precise identifiers for the level of measurement sought The amount of a changed parameter that can be detected Sources of Error Data Collection Problems People Problems Researcher Problems Institutional Problems Event Problems Measurement Validity Measurement Reliability Computer Support for Data Data Input Data Storage Data Retrieval Statistical Analysis Numbers and Use of Numbers Nominal (subjective) Ordinal (subjective) A scale that is subjective but shows a direction, e.g. pain scale, cancer staging, all Likert scales Interval (objective) A Named category given a number for convenience, e.g. males are 1 and females are 2 Numbers where the interval between them is meaningful, and there is no absolute zero but an arbitrary zero, e. g. a temperature. These numbers can be less than zero. Ratio (objective) Numbers where there is an absolute zero which means it is absent or there is a denominator that allows for comparison of meaning and . e. g. number of cases or infections per 100 hospital days, stage 2 skin breakdown per 100 patients. Bivariate Data Analysis Independent Groups Nominal Data Chi squared (Two or more samples) Phi (Two samples) Cramer’s V (Two samples) Contingency Coefficient (Two samples) Lambda (Two samples) Bivariate Data Analysis Independent Groups Ordinal Data Mann-Whitney U Kolmogorov-Smirnov (two-sample test) Wald-Wolfowitz Run Test Spearman Rank-Order Correlation Kendall’s Tau Kruskal-Wallis One-Way Analysis of Variance by Rank (three or > samples) Bivariate Data Analysis Independent Groups Interval or Ratio Data t Test for independent samples Pearson’s Correlation Analysis of Variance (Two or more samples) ANOVA Simple Regression Multiple Regression Analysis (two or more samples) Bivariate Data Analysis Dependent Groups Nominal Data McNemar Test Cochran Q Test (three or more samples) Ordinal Data Sign Test Wilcoxon Matched-pairs, Signed-Ranks Friedman Two-Way Analysis of Variance by Ranks (for three or more samples) Bivariate Data Analysis Dependent Groups Interval or Ratio Data t Test for Related Samples Analysis of Covariance (for three or more samples) ANCOVA Multivariate Data Analysis Interval or Ratio Data Multiple Regression Analysis Factorial Analysis of Variance Analysis of Covariance Factor Analysis Discriminate Analysis Canonical Correlation Structural Equation Modeling Time-Series Analysis Working with Descriptive Data: A Toolkit for Health Care Professionals Using Descriptive Statistics Correlational Descriptive Predictive Descriptive Model Testing Descriptive Statistics vs. Tools Inferential Statistic Analysis Statistics (regression, correlation, t-test, Ftest, Multivariate testing etc.) Descriptive Statistic Analysis Tools to display information Critical Path Process (p. 524) 1. 2. 3. 4. 5. Select the process Define the process Form a team Create the critical path Make the path a working document Critical Pathway for Complaints of Chest Pain in ED ED Patients c/o chest pain No previous symptoms Good Health Min. Risk factors Previous symptoms Has some risk factors Previous CAD many risk factors O2, IV, Bloods, EKG O2, IV, Bloods, EKG ASA, Nitroglycern O2, IV, ASA, Beta, Blocker, Morphine, Cardiac Cath Lab CCU Force Field Analysis Driving Issues for Moving Minimum Grade at DSN From 72% to 74% Driving Forces (support efforts) Restraining Forces (conflict with efforts) Comparable to Other Schools Recent drop in NCLEX rates Faculty requests Significant Change in Policy More students would fail DSN had 90-94% NCLEX rates with 72% Indicators to be Used in Hospitals Quantitative measures Related to one or more dimensions of performance Help provide data that (when analyzed) give information about quality Direct attention to potential problems Types of Indicators Sentinel-event indicators Aggregate-data indicators Rating for med errors and patient complaints Continuous-variable indicators Serious injury or death indicator Number of new bed sores per day Rate-based indicators Infections per 1000 patient days Run Charts Probably most familiar/used tool Used to identify trends/patterns in a process over time Helps track if target level has been attained/maintained Run Chart – Trend Chart Used for Self Comparison Quarterly report of new bed sores for Unit X 2008 40 20 0 1st Qtr Unit X 2nd Qtr 3rd Qtr 4th Qtr Unit X Comparison Run Charts – Trend Charts-(Dangerous because these are not ratio numbers) Quarterly report of new bed sores for Units A, B, & X for 2008 30 20 Unit B Unit A Unit X 10 0 1st Qtr 2nd Qtr 3rd Qtr 4th Qtr Unit X Unit A Unit B Histograms Bar charts that display: Patterns of variation The way measurement data are distributed Snapshot in time May be more complex to establish; consult statistics textbook if needed Comparison Run Charts – Trend Charts-(Dangerous because these are not ratio numbers) Quarterly report of new bed sores for Units A, B, & X for 2008 30 25 20 Unit X Unit A Unit B 15 10 5 0 1st Qtr 2nd Qtr 3rd Qtr 4th Qtr Comparison Run Charts – Trend Charts for Delta Hospital (can be compared equally) Quarterly report of new bed sores per 1000 patient days for Units A, B, & X for 2008. 16 14 12 10 Unit X Unit A Unit B 8 6 4 2 0 1st Qtr 2nd Qtr 3rd Qtr 4th Qtr Control Chart This is the control chart for infections from I.V.s on Unit X With 3 case per 1000 patient days as the standard (std) for 2008. 0.005 Max. Std. 0.003 0.000 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec x x x x x x x x x Min. x x x Pie Charts Descriptive data Shows a distribution by category Compared to the Whole Pie Distribution of new bed sores for hospitalized patients at Delta Hospital Total of 140 new bed sores reported in 2008 36 43 37 Unit X Unit A Unit B Scatter Diagrams Graphs that show statistical correlation between 2 variables Used when group wants to: Test a theory Analyze raw data Monitor an action taken Scatter Diagram Process Min. Program Passing rates in % 76 74 72 NCLEX Scores by % 100% Surveys Survey’s can carry a risk to them. Also know what Likert Scale you are using and why (1-4, 1-5, 1-10 most common). These are Ordinal Numbers Naturalistic Inquiry— (Ch. 3) Qualitative Research Methods Phenomenology Ethnography Auto-ethnography Grounded Theory Descriptive Qualitative Historical ? Non-Probability Sampling Theoretical Sampling Quota Purposive Sampling (Non-Randomized) Convenience Sampling Network Observational Measurement Unstructured Structured Category Systems Checklists Rating Scales Emic (from within) Etic (from external view point) Phenomenology Research: “The Lived Experience” Phenomenology is a science whose purpose is to describe the appearance of things as a lived experience. It allows nursing to interpret the nature of consciousness in the world. It can be descriptive or interpretive (hermeneutic). It is a philosophy, an method, and an inductive logic strategy Design Characteristics Purposive samples of 7-20 usually going for saturation. Instrument is the researcher Data collection is by interview of groups or individual that are verbatim, taped, and field notes. Data collection is directly tied to analysis, that eventually is coded or structured into themes. Unique Features of Phenomenology Most of the literature review is conducted at the end of the data collection. It is believed the CF biases the data collection and analysis. Like Grounded Theory but without a BSP or bias already in mind. It is conducted by gathering interview data from others. It is never quantitative, but some would prefer to try and keep it objective. Five Steps of the Method Shared Experience is presented Transform the lived experience into an experience the subject would agree with Code the data Put it into written form and create confirmation of the data texts. Create a complete integration of all of these for a research document NOTE: In come cases, researchers need to have Bracketing to control an over-riding bias or emotional response Qualitative Research Rigors The Five Standards (Ch. 13) Descriptive Vividness Methodological Congruence Theoretical Connectedness Analytical Preciseness Heuristic Relevance Defining Naturalistic Rigor Standards 1 and 2 Descriptive vividness narratives are texturized, thick, and full of details the writer shows connections and level of membership Methodological congruence details of exactly how the data is gathered with ethical rigor. Does the method match the design? Defining Naturalistic Rigor Standards 3, 4 and 5 Analytical preciseness the data is transformed across several levels of abstraction moving raw data to clusters, interpretations, or theory Theoretical connectedness ensuring the theoretical schema is clear and related to the data being collected and a lens for analysis Heuristic relevance readers must recognize the phenomenon as applicable, meaningful, & recognizable Other Types of Rigor Using Trustworthiness Trustworthy questions Trustworthy approach Trustworthy in analysis Trustworthy and authenticity of data Ethnography Research Defined as: “Learning from People” By Spradley Four Types of Ethnography Classical Systematic Defines the structure of a culture. Interpretive (hermeneutic) Years in the field, constantly observing and making sense of actions. Includes description and behavior. Attempts to describe everything bout the culture. To study the culture through inference and analysis looking for “why” behaviors exist. Critical Relies on critical theory. Power differentials, who gains and who loses, what supports the status quo. Historical Roots Early 1900s had several introductions Herodotus wrote about travel in Persia Malinowski’s Study of Trobriand Islanders Hans Stade wrote about his being in captivity by the wild tribes of Eastern Brazil The School of Sociology in Chicago, where the city was a laboratory from all the immigrants (dancers, muggers, case studies) Observation Methods Emic From within the research itself as a member or participant of some type. Etic From the outside looking in like a camera. It can be a peripheral issue or external observer member. Fundamental Constructs Is usually “etic” on the outside like a camera Sometimes they are “emic”, on the inside as one of the actors (more in sociology) Researcher is the instrument Fieldwork is where the work occurs Focus is on culture Involves cultural immersion There is a tension and reflexivity between the researcher as a member or researcher as researcher Stages of Ethnography Participant observation (gain access, rapport, trust) Descriptive observation (9) (space, actors, activities, objects, act, event, time, goal, and feelings) Ethnographic record (field notes, verbatim, old records, amalgamate the information) Domain analysis Focused observation (what is now critical) Stages in Ethnography-2 Taxonomic analyzing (categorize) Componential analysis (components of the selected areas) Discover cultural themes Take a cultural inventory Write up the ethnography Rigors for Ethnography Plausibility Credibility Not exactly self evident, so you look at sources of evidence Thick Description It is very easy to accept as truth Writing in such detail as to know exactly what is going on. We could also use the Five Standards Sources of Errors Personal reactivity False inferences Gaps in writing, remembering, and interpreting Going Native Grounded Theory Research Started by Glaser and Strauss in 1967 Used extensively in nursing research Takes into account the concepts of George Herbert Mead (1934) regarding symbolic interaction theory- how we give meaning to situations, words, objects, symbols Is very individualistic in meaning Most often used to study areas which previous research exists Steps in Grounded Theory are conducted simultaneously Observation Collection of data Organization of data Review of additional literature Forming theory from the data Using Constant Comparative Analysis Data Collection Methods Have qualitative and quantitative properties Interviews (one on one, groups) Observation Records (retrospective analysis) Surveys (quantitative) Questionnaires (could be quantitative) Demographic data Constructs of Grounded Theory Conceptual framework comes from the data rather than the literature review There is always an over-riding social issues being addressed called the Basic Social Process (BSP) Researcher focuses on dominate processes rather than describing the setting, or unit You compare all data with all other data Constructs of Grounded Theory You may change data collection methods in mid stream to be more appropriate to what has already been discovered The researcher is to be doing most sequential tasks all at the same time Constant Comparative Analysis Get data, look at it, look at the literature, look at previous data, go get more data, look at more literature, look at all the data, etc. Revise the question, collection method, and keep collecting data, look at literature, compare to old data, etc. Sampling Methods Called Theoretical Sampling Based on the current question Add new groups to the sample based on what it is you have learned (may need more men in the sample, or more people over the age of 70, etc.) The sample being used moves as the theory develops Coding the data Look for positive AND negative cases related to your social process Step One: read, describe, and interpret Step Two: constant comparison and clustering Step Three: reduce it to a BSP Conducting Grounded Theory Be aware of the social life of the participants Make less assumptions in the beginning Sensitizing to the literature, Bracket if needed Layers of reality are explored, assess your own energy to go further Spend enough time with participants and data Be observant to how the participants are doing Learn the symbols being used to create this reality Sample across time Case Studies from Stake (2000) and Yin (1994) These are OBJECT or METHOD issues Object: Has to do with what you want to study not an approach to how to study it Method: Can be quantitative or qualitative method (analytically, vs. holistically) Questions are aimed at “How” or “Why”(rarely “What”) Single or multiple cases-usually1or 2 Purpose of Case Studies Seeks the unique features (particular) while also describing the common by describing: The nature of the case The case’s history and background The physical setting Other contexts (economics, political, legal, aesthetic issues) Other cases through which this case is recognized Through the informants by which the case is known Examine changes across time (multiple case) Same group of different group Case Study Rigor Yin (1994) treats this as a positivistic activity, therefore: Construct, Internal, and external validity Reliability This is not just a pilot study for quasi- or full experimental designs. It is different. Stake (2000) treats it more naturalistic Thick description is key Auditability (can it be followed by the reader) Observational Measurement Could Use all of These Unstructured Structured Category Systems Checklists Rating Scales Emic (from within) Etic (from external view point) Interview Data Collection Unstructured Structured Describing interview questions Pretesting the interview protocol Training interviewers Preparing for an interview Probing Recording interview data Coding methods Problem Revisions I am curious about the standardized treatment protocols for circumcision of a new borne. NEXT REVISION NEXT REVISION NEXT REVISION NEXT REVISION Problem Statements-Questions dictates the design What is experience of police officers who were wounded in the line of duty related to their ability to return to work? What are the unique features of Hospitals that have NP conducting all surgical admission assessments? There is (is no) statistically significant difference in iatrogenic diseases between nurse to patient ratios of 1:5 vs 1:8 on General Medical Units. Does the birthing center philosophy show a relationship to the type of care provided and if so, what is the relationship. How did the July 08 BSN cohort at DSN obtain a 99% NCLEX pass rate? Special Research Designs Triangulated, Mixed, Blended Historical Research Action Research Outcome Research Intervention Research Triangulation Blended Designs First used by Campbell and Fiske in 1959. Denzin in 1989 identified four different types. Data Triangulation Investigator triangulation Theoretical triangulation Methodological Triangulation Kimchi, Polivka, and Stevenson (1991) have suggested a fifth type Multiple Triangulation Data Triangulation Collection of data from multiple sources Intent is to obtain diverse views of the same phenomenon. (Longitudinal is different and is looking for change) Validate data by seeing if it occurs from different sources Investigator Triangulation Two or more investigators with different research backgrounds examining the same phenomenon Clarifies disciplinary bias Adds to validity of data Theoretical Triangulation Using all the theoretical interpretations that could conceivably be applied to a given area Each view is critically examined for utility and power Increased the confidence of the hypothesis Can lead to even greater T. F. beliefs Methodological Triangulation The use of two or more research methods in a single study Design level Data collection level Two major types Within-method (all are one philosophy) Across-method (across philosophies) Pros and Cons of Triangulation Very trendy in the 90’s Can be used with smaller N Combined methods may just be the rise of a new method There are philosophical risks Complex designs and therefore complex analysis Action Research: AKA clinical research, clinical inquiry, A systematic investigation conducted by practitioners involving the use of scientific techniques in order to improve their performance. Kurt Lewin (1946). Advantages of Action Research: The reflective practitioner Contributes to the knowledge base of teaching practice-self awareness Supports the professional development of practitioners –more competent in research issues Builds a collegial network Identifies problems and seeks solutions in a systematic fashion It can be used at all levels and in all areas of education Examples of Action Research Pick a topic Define the problem Select a design Select subjects Collect the data Analyze the data Application of results WHAT MAKES IT ACTION RESEARCH What Makes it Action Research Invested in rigorously empirical (positivistic), and reflective and interpretive (naturalistic) Engages people who have traditionally been called “subjects” who are active in the research process. Results have a practical outcome related to lives or work of participants. Outcome Research p.272-317 Came from evaluation research of the 70’s and 80’s Focuses on the end result of patient care and linked to the process that caused the outcome Momentum is from policy makers, insurers, and the public Level of concern: 1. Care by clinician, 2. Amenities, 3. Care by the patient, 4. Care received by community More complex that it may appear Evaluation of Outcome Research Process Evaluation Structure Evaluation Involves Standards of Care Involves Practice Styles Involves Cost of Care Elements of the Structure Philosophies of Management & Decision Making Process Evaluate Structure Issues and their impact on the care provided Lacks a set methodology Indicators of Outcome Research Many Descriptive Indicators for Nursing Care: NDNQI, Picker, Stage all bed sores on patients at admission vs. during stay and at discharge. There must be a clear link between outcome and process We see practice based web sites: AHRQ, APRNet, PBRN group, Sampling in Outcome Research Large heterogeneous samples, but not randomized. They want a full spectrum of the population. However, they want samples who were treated and those who were not treated to compare differences in outcomes. Risks, no random sample, small sample sizes are often used putting all their inferential statistics at risk for error. Intervention Research It is used to give “Causal Explanations” for what is being seen Uses quantitative and qualitative methods It is more than a single research event, but it deals with multiple issues over time Intervention Research Process Extensive search of what information is available Heavy emphasis on the intervention and refining its use Field tested to see if it will work It will involve a host of studies over time Has a host of informants who explain the local culture and what it will take to get data Intervention Research Methods Integrative lit. reviews Consumer publications Standards/ guidelines Meta-analysis Health policy analysis Personal exp. Reflections Consensus conferences Retrospective chart reviews Descriptive-Correlational studies Observation Case study Focus groups Qual. Studies Concept analysis New media Position Papers Delphi studies Outcome studies Risk for Use of Intervention Research Risk is asking the wrong question Inadequately trained interveners Poorly defined intervention Many confounding variables that can show up Too complex to manage and integrate Long time can change many factors: i.e. who is doing it, where can you still collect data, level of commitment by locations, etc. Criteria for Intervention Research Design: The intervention is-- Effective Replicable Simple to use Practical Generalizability Compatible with local customs and values Historical Research Thought of as qualitative because it lacks sampling, treating, and controls. Uses Quantitative language, i.e. validity and reliability of data—best primary sources of data. Looks at external criticism of data (where, when, by whom), and internal criticism of data (reliability, authentic, biased lens of writer) Process of Historical Research No Visible Rigor from Qualitative or Quantitative Research Outline Watch for cross-referencing Be prepared to spend months to years collecting the data Careful attention to note taking for all data collection A synthesis of all the data collected and may need an interpretive strategy Develop a writing outline Write your Historiography