991219 http://www.ul.ie/~infopolis/methods/index.html Ergonomics Methods and Tools This section is intended for practitioneers such as designers, engineers, project managers and ergonomics practitioneers in the public transport domain - with or without previous task analysis experience - who need to carry out a traveller task analysis. It includes and explains several methods of task analysis with an ergonomic point of view. It shows how and when various task analysis, task description and task simulation methods can be used. Task Analysis Methods Task analysis is the study of identifying how users do certain tasks or operate different products. This is usually done before a product is designed, or when a product needs to be redesigned based on usability testing results. Task analysis is an important aspect of user interface design, ensuring that the end product is usable and practical. Techniques used to collect data on actual or proposed task performance: [Critical Incident Technique] [Group Techniques] [Interview Techniques] [Observational Techniques] [Questioning Techniques] [ Verbal Protocols] Task Description Methods Task Description Methods are techniques which structure the information collected into a systematic format (e.g. the data from observation studies may be recorded by the use of flow charts, or operational sequence diagrams). Such formats may then serve either as reference material to enhance the understanding of the human-system involvement, or may be used more directly (e.g. the hierarchical task analysis technique can be used to identify user needs) Representing such data in a pre-specified format: [Charting and Network Techniques] [Hierarchical Task Analysis] [Task Decomposition Techniques] Task Simulation Methods In addition to classical data task collection methods, potential users can be placed in mockup or prototype versions of the system that simulate the proposed equipment and 'walk through' the planned task sequences. [Walkthrough and Talkthrough Methods] [Mock-ups and Simulators] Selecting a Method to Evaluate a User Interface When deciding on the method (or methods) to use it is important to consider: The exact purpose of the evaluation; The external limitation imposed on the evaluation process, such as the time constraints of the development cycle, cost and availability of equipment or expertise; Biases in the data collected: how much the evaluation affects the situation being evaluated (known as the ecological validity), the reliability of the methods and so on; Practical questions such as: Are the data easy to gather? How long will it take? How much will it cost? Are appropriate resources available? How well particular methods can be assimilated with industrial and commercial practices (for example, expert evaluation and interview methods can be integrated easily whereas observational techniques are more difficult). The features of the main evaluation methods are summarized in a table : Which method to use and when ? (adapted from Preece et al, 1994). Home Page I Menu I Site Map I Contact I About the Project Project Scope I Overview of Existing Systems I Traveller Needs I Ergonomics Methods I Mock-ups I Guidelines Search I Library I Bibliography I Links http://www.ul.ie/~infopolis/methods/index.html © Infopolis 2 Consortium User Interface Evaluation : Which Method to Use and When ? Method Observational Evaluation Experiments Interviews Contextual Inquiry Heuristic Evaluation / Cognitive Walkthroughs Purpose Understanding the real world, Comparing designs and engineering towards a target Standards conformance, comparing designs, benchmarking Users’ opinion, requirements identification Understanding natural usage, requirements identification Engineering to a target, conformance testing Interface development Any stage of development At least a working prototype Any level of development Any level of development Early at specification or mock-up stage User involvement Yes, some control of tasks by users Yes, little or no control Yes, often no control Yes, considerable control No Type of data Quantitative and qualitative Emphasis on quantitative Quantitative and qualitative, but more emphasis on quantitative Qualitative Some qualitative but emphasis on quantitative Practical considerations Special equipment useful but not essential Laboratory conditions preferred None Little or no equipment needed. Video/audio may be used No equipment needed Some of the Key Differences Between Laboratory Conditions and Normal (Work) Environments Work - For example, typical word processing benchmark tasks are six pages long, whereas, in practice, working with documents 30 to 50 pages long is not unusual; Time - Experiments generally have a prescribed time context, whereas in work environments people tend to have more influence on what they do and when they do it; Motivation - In an experimental context the experimenter controls the situation, whereas usually in the work context there is scope for some negotiation; Social factors - In the work environment there is normally a social network of support that does not exist in an experimental context. Task Analysis Methods: Critical Incident Technique Overview The Critical Incident Technique consists in collecting information on incidents occurring over a specific period of time, for instance a trip, in order to make corrections and prevent accidents. These incidents are events or features which could have an effect on system objectives. It is a way of identifying events that may have a positive or a negative influence on the system. The first principle of the technique is that critical incidents will be memorable to those using a system. Taking an inventory of incidents, these can be classified and analysed. It should then be possible to identify what changes can be made in the system to ameliorate problems, or to ensure positive components of system success. Application Generally, the critical incident technique is most useful in the early stages of a larger task or activity analysis, because its strength lies in the correctly rapid separation of problem areas within a system. It is also most interesting that the critical incident technique is applied to existing systems. The technique can be seen as a very open ended one, where operators are simply asked to respond to questions. A much more systematic approach of interviewing operators on a daily basis, or getting them to fill in a weekly or monthly report on critical incidents can also be used (see also Deliverable 3 : "Needs of travellers: an analysis based on the study of their tasks and activities" Infopolis 2 project; see page Library of this web site to download this document). Advantages The critical incident technique is inexpensive and provides rich information. This technique is helpful in emphasising the features that will make a system particularly vulnerable. The technique is especially appropriate for the identification of unusual events, which may not be picked up in other techniques that rely on users reporting their everyday activities. It may also have a particular interest to system users who are interested in reporting on near-misses, etc. Disadvantages A first problem comes from the type of the reported incidents. The critical incident technique will rely on events being remembered by users and will also require the accurate and truthful reporting of them. There is also the problem of reliability in terms of memory, because many critical incidents may be forgotten or distorted by events. The last point to highlight is that the technique will emphasise only rare events, and more commonplace events will be missed. Type of Critical Incident Technique Micro-incident analysis This method aims at analysing how complex systems can recover small errors but also how global reliability is obtained (this objective is analysed deeply in critical safety situations, in which complex systems play a major part and where it is important to ensure their reliability such as Air Traffic control or Urgency Medical Centre). Task Analysis Methods: Group Techniques Overview A potential weakness of methods such as interviews or questionnaires is that they are based on the activities and views of individuals on their own rather than as part of a wider social-technical situation. Also group techniques are useful for getting opinions efficiently from a set of people. The main principle is that each participant can act to stimulate ideas in the other people present, by a process of discussion. It is well-known that the collective view is greater than the individual contributions (see also Deliverable 3 : "Needs of travellers: an analysis based on the study of their tasks and activities" - Infopolis 2 project. Not published yet) Application There are various means of managing group sessions. The simplest is to have the group meet and discuss all aspects of the problems being addressed. Alternative approach are, for instance, Focus groups, Philip 66, Brainstorming, … In order to encourage discussions, the group members should be assured that their contributions will remain confidential. If any recording is made, members should also be assured that this will remain confidential. In the domain of information systems groups may be of use in understanding social issues affecting the system and the design and the evaluation of possible design solutions. Advantages Groups techniques allow the rapid collection of a wide variety of views from a range of people. Group discussion helps to summarise the ideas and information held by individual participants. Meetings can be recorded for future analysis. Disadvantages Some shy participants may be inhibited by other group members. One member of the group may appear highly confident and persuasive and may influence the opinions of the other group participants. Some people may also not think creatively in a group setting and prefer to be interviewed or to complete a questionnaire. Types of Group Techniques Focus group A focus group brings together stakeholders and / or experts to help refine the requirements of a system or to evaluate a particular system. Views are elicited by a facilitator on relevant topics. A focus group is useful at the early stage of the system development: it can assist identification of problems for a target users population, refinement of requirements for systems, and assessment of the early prototype. It helps to identify issues which may need to be tackled and provides a multifaced perspective on them (see also Deliverable 3 : "Needs of travellers: an analysis based on the study of their tasks and activities" - Infopolis 2 project. Not published yet). Group discussion Group discussions are based on the idea of stakeholders within the design process discussing new ideas when relevant to the design process. Group discussions help to summarise the ideas and information held by individual members. Brainstorming Brainstorming, developed by Osborn, is probably the oldest and best known group method. This technique has been designed to help creative thinking in initial system development. The principle is to bring together a range of people to inspire each other in the creative, idea-generation phase of the problem-solving process. The members of the brainstorming group should have a range of experience but should consist not only of experts. The advantage of this technique is that it is a rewarding experience for the participant, and it is a quick and easy method to apply. The disadvantage is that user's expectations of the new system's capabilities may be heightened to unrealistic levels and the users may therefore be disappointed later. Task Analysis Methods: Interview Techniques Overview The interview technique is a systematic collection of verbal information. It consists in asking about users opinions and attitudes to get basic information with prepared questions asked by the interviewer. The answers are either written or recorded. The interviews can be structured or unstructured. The terms structured interviews implies that the content of the interview, in terms of the questions and their sequence, is predefined. Because of the structuring the interview offers the opportunity for more systematic collection data. The unstructured interview is more free, and the interviewee develops the themes proposed by the interviewer Application Interviews can be used at any stage during a comprehensive task analysis activity. They can usefully be applied early on in an investigation for collecting basic information about the situation. An interview with the user will provide a deeper understanding of the requirements for the system and of the user tasks involved. It can also assist in finding out more about individual user perceptions (see also Deliverable 3 : "Needs of travellers: an analysis based on the study of their tasks and activities" - Infopolis 2 project. Not published yet). In order to encourage frank answers, the participant should be assured that his contribution will remain confidential. If any recording is made, the interviewer should also assure that this will remain confidential. Newman and Lamming (1995) describe a number of elements of a successful interview: Determine some basic domain knowledge before the interview so that time is efficiently utilized during the interview. Clearly state to the interviewee the purpose of the interview at the outset. Enumerate all user activities with general and follow-up specific questions. Find out how user activities are performed. Trace interconnections with other users. Uncover issues that determine and affect the performance of the user tasks. Follow up on exceptions, the unusual activities which are unlikely to occur during observation. Advantages Interview techniques are useful for identifying possible areas for more detailed analysis. Interviews are easy to conduct and direct, the unstructured interview can generate interesting points, statistical analysis can be run on the users answers. The data collected provides information about general rules and principles and is faster than observational techniques. Interview techniques are useful for investigating events which occur infrequently. The interviews can be recorded for a future analysis. Disadvantages Respondents are not committed to give correct answers and may often be influenced by what they believe the interviewer requires, or what they themselves wish to portray. The interviewer may need to acquire domain knowledge in order to know what questions to ask. There is a range of considerable bias due to the understanding by the users of the questions, and the subjective collected information might be misleading or inaccurate. The critical aspects are the choice of the place for the interview and how to conduct it. Types of Interview Techniques Unstructured interview or depth interview In an unstructured interview, the respondent must develop the theme proposed by the interviewer. The investigator does not impose a questionnaire and must follow the thoughts and reactions of the respondent. Unstructured interviews are good for investigating potential emotional and / or sensitive personal issues. Semi-structured interview Semi-structured interviews should only be carried out in a situation where broad issues may be understood, but the range of respondents' reactions to these issues is not known or suspected to be incomplete. This type of interview is mostly applicable in situations where both qualitative and quantitative feedback are required. Structured interview Structured interviews are useful in situations where the respondents range of replies may be estimated and there is a need to clarify details, opinions or ideas. Structured interviews work well when the assessment goals are clear. Task Analysis Methods: Observational Techniques Overview Observational techniques are a general class of techniques whose objectives are to obtain data by directly observing the activity or behaviour under study. A wide range of observational techniques are commonly used and combined according to the particular requirements of a study. Observations may be direct observations, where the investigator is actually present during the task (visual observations, remote observations via closed-circuit television), or indirect observations, where the task is viewed by some other means such as video recording or time-lapse photography. Application Perhaps the most obvious method by which the user can be analysed is to observe him or her performing the task in an operational situation. Although observational techniques can capture the details of many individual tasks, they may not elicit the work role. Thus, observational techniques are most appropriate when the information of primary interest is of visual or auditory form. Thus, observation methods can be especially useful for recording physical task sequence, or verbal interactions between several people. One method which can augment the information gathered is that of current verbal accounts. Here the user is asked to 'think aloud' and describe the activities he or she is performing as they are undertaken. Advantages Observational techniques produce information which cannot be collected in any other way. Indeed, detailed physical task performance, social interactions, surrounding influences can be recorded. Observation methods allow the observer to become more familiar with the task. So they are very appropriate for exploratory studies because they allow the investigator to decide what to look for. Observation studies can be used to identify and develop explanations of individual differences in task performance. Observational techniques provide data which can be compared and contrasted with information collected by another investigator, or by another method. Disadvantages Real time observation requires trained ergonomists. Observational data is the widest possible form of information, for this reason the effort which must be expended on classification and analysis is considerable. Direct observation can be an intrusive technique. The user behaviour might be influenced by the presence of the observer,. Therefore the investigator must estimate whether the type of observation selected will significantly alter the task under study. Observational situations which produce rich context data are rarely the ones which produce precise and controlled data. So data will be incomplete, inconsistent and sometimes unusable Observational techniques cannot provide information about underlying thought-processes, and so they will be of very little use for highly cognitive activities. The equipment needed for producing high quality observational data can be expensive and difficult to install, and video recording followed by later analysis is time-consuming. Type of Observational Technique Passive observation At the simplest level passive observation, where the person conducting the observation quietly observes and records the user activity, can be used to record some details of the processes which the user undertakes. Passive observation has the advantage that it minimises the intrusion which the user experiences. One problem with passive observation is the practical difficulty in recording sufficient detail of the activity. This can be overcome through the video recording of user activity which, after transcription into textual form, can provide a highly detailed analysis. Participant observation The investigator takes part in the tasks alongside other operators. It may be a useful data collection approach if skilled performance is such that actions are "semi-automatic" in which case users may find it difficult to verbalise how they are achieving the task goals. This type of observation can also be useful if aspects of team performance are being investigated to understand how the team members are organised and perform their tasks. Communication and conversational analysis Investigation of the patterns in the organisation of people's interaction (it would be possible to use the principle of analysing a conversation between two travellers, or one traveller and one transport enquiry office operator in real situations, in order to identify the needs for the traveller in specific contexts, what are the difficulties encountered, how a system could resolve these difficulties,... ). One of these methods is the language / action approach which considers the language as a means by which people act. Advantage of this method : it provides a complete and logical conceptual frame in order to investigate all kinds of conversation. Disadvantage of this method : many situations are characterised by subtle communication processes not taken into account by the model. There is also a difficulty in labelling an interaction and a message especially if they do not fit into the request or promise categories. Task Analysis Methods: Questioning Techniques Overview Questionnaires are sets of predefined questions arranged on a form and typically answered in a fixed sequence. The questionnaire is usually completed by the individual under study, but it is also possible for the questionnaire to be administered by a second party. There are essentially two types of questions : closed questions (respondents are asked to select an answer from a choice of alternatives) or open questions (the respondent is free to provide his own answer). Application Questionnaires can be used in a very broad series of applications. They can be used for evaluating specific features or issues of a system, and they can also be used in order to investigate different opinions, perceptions, knowledge or attitudes towards this system. However, designing a questionnaire is a much more complex task than it might appear. It is important to make the questionnaire easy for the user to complete. There are two quite obvious reasons for this. First in order for the analysis to be statistically significant it is necessary to generate a high return rate. If users find difficulty in completing the questionnaire they might not complete it, and only a relatively small proportion will be returned. The second reason is that it is important to avoid ambiguity in the questions which are posed. Ambiguous questions generate unreliable answers. For both these reason, it is usual to trial the survey instrument with a small proportion of the intended population so that any design problems can be addressed before the main survey is carried out. Advantages The main advantages are the area and the flexibility of administration. They can be filled in at any time by any number of individuals. Because it is relatively easy to administer a large number of questionnaires, they are especially useful for comparing the opinions or understanding of a large number of users. The use of closed questions can prevent digressions. The use of multiple choice items and rating scales can produce answers that can be weighted for their reliability and consistency. Careful decisions have to be made concerning the sample of user population and whether they are representative of the whole group of interest. Another advantage of this technique is to gather a large amount of information from many people, relatively rapidly and at low cost. Disadvantages Respondents do not necessarily give correct answers and may often be influenced by what they believe the analyst requires, or what they themselves wish to portray, especially for personal information. Anonymity and having the questionnaire completed remotely from the analyst can reduce distortions. Loss of information or subjective insights may be introduced by inappropriate perceptions of what the question is asking. The correct choice of target population minimizes this risk. There is room for considerable bias in what questions are asked and how the answers are interpreted, mainly if open-ended questions are incorporated. Response rates for posted questionnaires can be very low (30 % or less is considered as normal). Types of Questioning Techniques Multiple Choice Items Multiple Choice items provide two or more specific responses from which respondents must choose an item which is most representative of their opinions, perceptions, knowledge or attitudes. Rating Scales Rating scales can be used to obtain subjective information from respondents which gives an indication of both the nature and magnitude of their opinions about certain aspect of a task. Bipolar alternatives Bipolar alternatives is a type of rating scale, where respondents must make a choice between two alternatives. Ranking Ranking requires respondents to order several items according to precise criteria. Open-ended questions The respondents are free to write their own comments, answer or opinions on the questions Task Analysis Methods: Verbal Protocols Overview Verbal protocols are the recorded verbalisations of users as they carry out their tasks. This technique uses users' spoken comments, who verbalise how they use the system, explaining what they are trying to do and the type of problems they experience. Verbal protocols may be collected by video or simply audio tape recorder, direct notes can be also used. These self-commentaries are made while the task is being carried out in order to avoid the inevitable distortion or forgetting that could occur if the reporting were left until afterwards. These protocols are particularly useful for gaining information about the psychological or cognitive process for users' actions, which cannot be directly observed, such as the knowledge requirements or mental processing. However, the think aloud protocol, during the performance of the task, can impose an additional task-load on the user, significant when the task is difficult. This load can alter the way the user performs the task. The post-event protocol is based upon the user's memory after running the task. The self-confrontation is a protocol where users are asked to comment on their actions in front of the video recording after the system use. Application It is important that verbalisations should not interfere in any way with task performance and that users should freely report on what they are doing without any direction. Also, verbal protocols are limited by the ability of subjects to freely and honestly state why they are undertaking particular actions, without first making a conscious effort to explain these actions. When to use : To identify the likely errors that users make while learning how to use the system. To identify the expectations and preconceptions that users may bring to an interaction, revealing the internal model developed by the user concerning the way the system operates. This technique is useful for activity analysis of existing tasks, for system evaluation and for teasing out any misunderstanding or misconceptions a user has when operating a system for the first time. Advantages The main advantage is that it helps to find out what the user is thinking when performing a task or operating a prototype system. Verbal protocols provide a basis for investigating the underlying mental process of complex tasks which cannot be studied in other way. Data collection is rapid, because very few special arrangements need to be on-site, and data analysis will usually be conducted off-site. The results of applying the technique will be credible and comprehensible to non-specialist observers or users. Disadvantages Verbalisations may be biased because they interfere with the basic task, for instance, changing the speed or the execution method. This is mainly the case if the activity necessitates the use of short term memory. Verbal protocols may reflect only the results of cognitive process, rather than the process itself. Thus the depth of knowledge which can be obtained is often limited, and it may be necessary to do additional investigations. Language may be an inadequate descriptive technique for the process being studied. Many mental processes depend on visual, auditory, or mathematical imagery for their execution. Therefore it may be difficult to verbalise such events. Another disadvantage is that it is timeconsuming to analyse audio and video tapes afterwards. Types of verbal protocol Personalised Task Representation (PTR) This method is used to investigate user procedures that are not well documented. A standard set of questions can be used to probe activities in a flexible manner : discussions with the task expert are about the activity they perform, the reasons for each task action, and how the action is carried out. · Advantages of this method : it can be applied to a range of knowledge and decision based processes. · Disadvantages of this method : the data captured can produce complex data diagrams that will need to be structured afterwards into a logical form. Task Description Methods: Charting and Network Techniques Overview Charting and network techniques are graphic representations produced from the analysis of a system which can be used to describe tasks within a system. The principal aim of charting and network techniques is to use a formal graphical representation of the task which is more easier to understand than a textual description. Application These techniques provide a systematic approach to describe tasks and offer a graphical description, which allows an adaptable means of analysis. The techniques have a range of applications because the techniques try to represent interaction between human and machine or system. These techniques can also be useful for describing the action plans of a hierarchical task analysis. Advantages Charting and network techniques are able to offer a very clear representation of the tasks (tasks into clear units and diagrams which link units) and provide information in a format that can be directly used for making decision, producing requirements or creating operational procedure. The charts can be used to represent human activities, tasks and equipment in a common format. Flow charts and network diagrams can incorporate the representation of the flow of time, in order to describe the conditions under which a task is performed. Disadvantages Charting and network techniques become less satisfactory if cognitive content is increased. Only a limited amount of information can be represented in one diagram, too much information lead to complicated charts. Types of charting and network techniques Input-output diagrams This technique, developed by Singleton, provides a representation of a task as an "inputoutput diagram". The technique involves a first selection of tasks or steps, and then identifying all inputs and outputs which are necessary to execute these tasks or steps. Inputoutput diagrams are useful for understanding the role of the user in a system, showing what information and actions are necessary to realised the task. Process charts Process charts are a 'top-down flow line' which links a sequence of activities or events. Moreover, information about time taken or distances moved can be shown on the process chart. This technique is mainly used for individual task. Functional flow diagrams Functional flow diagrams are block diagrams which illustrate the relationships between different functions. Usually, these functions are expressed by a "noun-verb-number" combination, in order to identify each function and the sequences within the process. Decision-Action-Information diagrams The Decision-Action-Information diagrams, also called information flow charts, depict the process through a system in terms of the decisions and actions which must be performed. Decision blocks are usually represented as diamonds and the possible solutions are labelled on the exit lines and blocks. Scenario analysis Another method for representing data in a pre-specified format is the scenario analysis. This method allows the identification of classes of typical scenarios (like road accidents scenarios). The analysis can be made from archive documents and interviews. The advantages of this method are: introduction of cognitive factors in the model in order to understand incident genesis, all stakeholders participation in the accident process are represented in the scenario, strategies of risk management and environment characteristics are represented. Task Description Methods: Hierarchical Task Analysis Overview Hierarchical task analysis, also called HTA, is a broad task analysis method, which describes the task in terms of a hierarchy of operations and plans based on structure chart notation. The hierarchical task analysis prompts the analyst to establish the conditions when various sub-tasks should be carried out in order to meet a system's goals. This method produces a hierarchy of three levels of task analysis: Goals (external task): system state that the human wishes to achieve, Tasks: structured set of activities in some sequence to achieve goals, Operations or actions: different things that a person must do within a system; simple tasks having no control structure. Application Hierarchical task analysis can be used to deal with specific issues, such as interface design or work organization in order to help the designer articulate how tasks should be executed. Hierarchical task analysis is a useful framework in which other task analysis techniques can be applied effectively. The main characteristics of the HTA : The goal decomposition is based on a person's verbalisation (interview, thinking aloud methods) and thus on his mental representation of the activity. Each sub-goal and activity may be described by a verb and a name Each sub-goal or activity has a number The plan at each level of decomposition and on each branch of the tree indicates how the activities are done (sequential, parallel, simultaneous) and their characteristics (optional, conditional, loop). Each plan has a number. The decomposition stops when it is not possible to break down the cognitive activity any more by the person, or when the analyst considers that more decomposition would not be useful for the application. Advantages This method is easy to learn and to use. The hierarchical structure of this task analysis approach allows the analyst to concentrate on crucial aspects of the task within the context of the overall task. Also other specific techniques of task analysis may be applied. This method is best developed as a collaboration between the task analyst and user involved in operations. Thus the analyst should operate in accordance with the perceived needs of people who are users of the system. · This method can form the basis of many other assessments, such as the communications analysis. The hierarchical task analysis is commonly used and widely accepted in cognitive task analysis. The HTA is very powerful because it can be applied to different types of physical and mental activities and different domains of applications. Disadvantages The hierarchical task analysis requires a lot of time, skill, and effort to use. The method must be carried out with the collaboration of managers, engineers and operating staff, and this collaboration involves agreement, time and effort from a lot of people. Errors and " unforeseens ", inevitable in the performance of a task, invalidate a part of the plans. It is difficult to represent in the plan goals which apply to every activity, interrupted activities or 'ad hoc' activities The hierarchical task analysis applies only to procedural activities and not to heavily parallel activities. Task Description Methods: Task Decomposition Techniques Overview The task decomposition is a structured way of developing the information from a task description into a series of more detailed declarations about particular issues. It starts from a set of tasks descriptions which describe how each element within a particular task is carried out. Next, further information is collected about each task element, for example, displays used or time taken. Then this information is presented using a sub-heading, so that the total information for each step is boken down into a series of statements. Decomposition is aimed at determining the control and information requirements of each step so that it is possible to ensure that the necessary interfaces for a task will be provided. Application The task decomposition is an information collection technique which is used to systematically expand upon the basic description of the activities which must be performed in each task element. The task decomposition includes three steps: obtain a task description, choice of decomposition categories, collection and presentation of decomposition information. This method is particularly useful as a basis for the preparation of control or information requirements lists (see also Deliverable 3I : "Needs of travellers: an analysis based on the study of their tasks and activities" Infopolis 2 project ). Description of the Technique There are three steps which must be undertaken in any task decomposition: obtain a task description, choice of decomposition categories, collection and presentation of decomposition information. Obtain a task description The starting point for a task decomposition is a set of clear task descriptions for all the task elements which are associated with a particular task. Normally, these descriptions would be derived from a hierarchical task analysis, but for short task sequences, or certain highly proceduralised tasks, these task descriptions could be produced as a simple linear sequence of task elements representing the order in which they would be undertaken. Choice of Decomposition Categories This technique, developed by Miller (in "A method for man-machine task analysis"), proposes that each element is decomposed into the following categories: Description Sub-task Cues initiating action Controls used Decisions Typical errors Response Criterion of acceptable performance Feedback The above decomposition approach still has relevance to the analysis of tasks today. However, this categorisation does not cover all issues, and so it will be necessary to develop other decomposition categories. Collection and presentation of decomposition information The decomposition categories are listed on a set of information collection forms, using a set of clear sub-headings, and starting with the task description. Example from Infopolis 2 (Delivrable 3 - WP 3.2: Traveller Needs Not published yet) Task Purpose Cue (for Decision Action element element) travel steps: Optimum way with rapidity Frequency Route Networks Pre-trip as criteria Vehicles On-trip Display Control Likely errors Looking for Timetable information (screens, leaflets, PAT, … ) Vehicles To miss arrival and one’s travel time connection, To get lost, … Lines End-trip Schedule Advantages The task decomposition technique can be applied to a general or specific task descriptions. This technique assures that the issues of interest are systematically considered for every task situation or element. The decomposition information is particularly useful for deriving control and information requirement lists. Disadvantages The main difficulty is that it may take a considerable amount of time and effort to obtain all the necessary information and to evaluate eventual conflicts Task Simulation Methods: Walkthrough and Talkthrough Techniques Overview A walkthrough, as the name implies, involves a user walking through a task by demonstrating it. So a walkthrough is a step by step method that requires users to undertake and explain some demonstration of a task in a realistic environment, without necessarily undertaking that task. A talkthrough method is very similar, but it is undertaken more remotely from the normal task location, so that the tasks are verbalised rather than demonstrated. So, walkthroughs and talkthroughs have a close relationship to verbal protocols. Application Walkthroughs and talkthroughs are generally used to describe and verify the principal observable components of tasks. Walkthroughs can take place in real time with a minimum of comments from the user who is performing the task. Walkthrough and talkthrough methods can be considered as lying on a continuum which represents the degree of accuracy with which a task and its interfaces are represented. These techniques can be used in their own right to assess whether a task is feasible and can be accomplished in time. Advantages Although it is not necessary to perform a task, only to demonstrate it, a walkthrough is effectively an actual performance of a real task. This is also true of talkthroughs, since the user is describing observable events and behaviours rather than inferred mental processes. In the case of a talkthrough, or if the walkthrough is not executed in real-time, it is possible to interrupt the process with a period of questioning. Task description methods can be used to further describe the data collected in a more useful way. They can be constructed from sufficiently comprehensive walkthrough or talkthrough methods. Disadvantages The technique requires an experienced user. The exception to this is if the system is intended either for casual and untrained users. In such cases, the user's expectations of the way in which tasks should be performed will be a focus of interest. In the case of interrupted walkthrough, the analyst must to have a good knowledge of the system, in order to ask the right questions. The effort which must be devoted to data analysis is considerable. Types of Walkthrough and Talkthrough techniques Talkthroughs The main advantage of talkthroughs is that they do not require any special task surrounding, thought it is helpful to have some technical documentation available, such as procedure or control / display panel drawings. This gives the advantage that talkthroughs can be undertaken early during the design stage before any mock-ups, simulators or real plant have been produced. In such situations, it is more profitable to utilise the knowledge of several individuals and to make a group discussion. Another important use of talkthroughs is as a means of getting an orientation before undertaking a more extensive walkthrough. Detailed Walkthroughs These are walkthroughs in which the accent is placed upon gathering information about particular aspects of a task, so that is paced by the rate at which responses are given, which will not necessarily correspond to the normal rate of task performance. Real-time Walkthroughs In real-time walkthroughs the main aim for the user is to closely imitate the way in which a task would be undertaken. An essential element of this is that the walkthrough should be undertaken at a realistic pace. And so for most tasks the depth of the comments which the respondent is able to provide will be limited. In order to enable a person to undertake a real time walkthrough at a realistic progression, it will be necessary to undertake such a walkthrough in a relatively highly realistic task environment. The main reason for undertaking a real time walkthrough will be to determine whether the workload which is imposed by a particular task will be feasible for a user, and to assess the effects of dynamic factors upon task performance. Team Walkthroughs These are real time walkthroughs which are undertaken by a team of users who are either undertaking the same task, or a group of highly related tasks. Such walkthroughs are useful for investigating the interactions and communications between individuals in complex multiperson tasks. Task Simulation Methods: Mock-ups and Simulators Overview Under the names mock-up and simulator a range of techniques can be considered which involve the development and use of some form of simulation of the system. The main objective is to observe user activity and performance in a representation of the task environment. In an evaluation of the mock-ups real potential users of the system are involved in implementing typical tasks they are likely to encounter when the system is in operational use. Application Simulators and mock-ups will usually be used when the real system is unavailable or expensive for study, or in order to study systems prior to the real system being designed. The simulation would typically be used to finalise appropriate working methods, ergonomics of display / control layout and design, identification of possible sources of error, or to derive learning recommendations (see also Deliverable 4: "Mock-ups: different versions of different systems" - Infopolis 2 project. Not published yet) . Advantages The simulators and mock-ups offer the possibility to assess user activities in a situation which is not normally easily observable. Disadvantages The behaviour observed in simulation may not be fully realistic, because of the user's motivation toward the simulator or mock-up and because the simulation may not contain all the features of the real task.