Comparing the Cognitive Walkthrough to Other Walkthroughs
 
    The primary characteristic that links all types of walkthroughs are the reliance on the tasks that the users would perform as a basis for the evaluation.  That which differs amongst the various walkthroughs is the way in which these tasks are used to inspect the interface and provide useful information.  The cognitive walkthrough (CW) as mentioned focuses on a cognitively based problem solving model that focuses on how users develop goals, chooses and execute actions, and integrate their background knowledge into their goal structure.  This approach is different from the pluralistic approach that relies on real users to develop action sequences and provide feedback about interface design.
 
    The pluralistic walkthrough uses a different composition of evaluators that tends to blend some actual users, designers, and usability experts.   In contrast, the CW team is composed of actual designers, implementors, and occasionally usability specialists into the analyst team.  Also, the pluralistic walkthrough has the participants write down their anticipated action strategy for a given task as they are presented with each screen of the system.  After this is done, all the representative users voice their concerns and ideas while the other team members (mostly the designers) note the responses.  The designers then subsequently raise other important issues that were perhaps not discussed by the users themselves. Only at the end, if some important usability issues have not been addresses does the usability specialist provide input into the process.  The format for the pluralistic walkthrough is very similar to that of a focus group whereby issues are raised, discussed, and presumably solved by drawing on the various experiences of all the group members.   Although this walkthrough views the entire team as evaluators, the users are responsible for developing the action sequences and for bringing out usability problems.   It does however require a basic understanding of the cognitive problem solving process but this can be learned and implemented by anyone interested in evaluating the system.  The pluralistic walkthrough often leads to unique problems and solutions as it draws from the strengths of all the potential groups of people involved in the design, including users, designers, and usability personnel.  This walkthrough method is often regarded as an early concept test or a software test.  Refer to Bias (1994) for a detailed description of the pluralistic walkthrough.
 
    The advantage of the CW over the pluralistic walkthrough is that there exists no need for actual users in the process.  The CW relies on the detailed user profile provided in the preparation phase as a basis for understanding how the tasks will interact with the user's background knowledge.  This approach is conceptually consistent with the premise of the CW because by having the team develop a user profile, it forces the analysis to focus on what the design team believes the user population to be and therefore emphasizes appropriately what the designers have wanted to focus on.  The hazard with this approach is that, what if the user profile is wrong?  This would lead to an ineffective CW and would not draw out the usability problems that should be addressed for the real target group.  This is where the designers should draw on the experience of outside resources, such as the marketing personnel, in order to develop an accurate user profile and understand what the users are really like.

    The CW also stems from the same roots as another type of walkthrough, the programming walkthrough.  The objective of the programming walkthrough is to design a programming language using the general attributes of the language as a starting point and elaborating the goals, subgoals, and actions required to accomplish a specific programming task.  Although both the CW and the programming walkthrough focus on goals, subgoals, and actions as the fundamental units of analysis, their applications are very different.  Nevertheless, each walkthrough stems from the same basic premise, that is focusing on the tasks to inspect the target object (interface / programming language).  For more information on programming walkthroughs, refer to Bell et al. (1994).
 

Return to main page