Analysis of Observational Data
Lucas Noldus1, Aurelia Kwint1, Wibo ten Hove1 and Ruud Derix2
1Noldus Information Technology b.v., Wageningen, The Netherlands 2Noldus Information Technology GmbH, Freiburg, Germany
User-centered design of computer systems involves a variety of methods and techniques to gather data about the quality of a design, prototype or product in relation to the requirements and desires of the user. Which method is used at any given moment depends on the phase of the development process. Interviews, questionnaires and focus group discussions are commonly used to gather user requirements or views on new design concepts. As soon as a prototype is available, walk-through sessions allow users to comment on details of the user interface and system functions relevant to the tasks to be carried out. Later on, user testing takes place in order to evaluate how well the system can be used to perform actual tasks. Provided they are well-designed and realistic, user trials show how users will respond to the future system and provide invaluable evidence concerning problems users may encounter. User tests also allow a quantitative measurement of productivity issues.
User testing involves the measurement of all relevant aspects of the interaction between the user and the system. Interaction involves the behavior of two interactants, so both user behavior and system behavior must be evaluated. Relevant aspects of the user's behavior include the use of input devices (keyboard, mouse, etc.), manipulation of objects (e.g. documentation materials), body posture, facial expression, gaze direction, verbal comments (e.g. thinking aloud, talking to colleague). System behavior is usually limited to screen display (menu, dialog or message being activated), auditory cues and triggering of peripheral apparatus.
Measuring human-computer interaction often results in a wealth of data to be analyzed. To increase the efficiency of user testing as a component of the design