CS252. User Interface Design

Solutions to Homework 3 - Chapter 5 in Jordan book


2. Which of the methods discussed in the chapter require a working product (or at least a good interactive or Wizard-of-Oz prototype)?

The following methods rely on having some kind of product (or close facsimile):

The following methods probably need a product, but could conceivably be based on paper prototypes:


4. Why do think-aloud protocols generally provide more information that is relevant to upgrading a software product than less structured methods, such as co-discovery?

Think-aloud protocols are the only way to determine why a user acts as he/she does. For example, if a user goes to the Edit menu instead of the File menu when he/she wants to save a file, only think-aloud protocols indicate whether it was a slip (user accidentally positioned the mouse incorrectly) or a mistake (user thought "Save" was on the Edit menu). Designers need to know about mistakes, since they indicate a mismatch between the user's mental model and the conceptual model presented by the software.


7. What is it about the valuation method that distinguishes it from other ways of having users indicate their preferences about planned features (focus groups, feature checklists, questionnaires, and interviews)?

The strength of this method is that it forces users to prioritize their needs and preferences. Typically, the user is told he/she only has a certain amount of money (or development time) to allocate, so the user must choose among many possibilities to indicate the one that appears to be the "best" tradeoff.


10. Compare and contrast the four non-empirical methods in terms of the following:

Task analyses require expertise with the particular modeling technique being used. Problems associated with consistency are the most likely to be found (e.g., inconsistent number/type of operations to perform similar operations); compatibility issues may also be uncovered. The analyses can be very time-consuming for a complex interface, resulting in involve significant salary costs.

Property checklists require familiarity with design guidelines and expertise in how to apply them to interface design, but also depend indirectly on the expertise of the person who originally created the design checklists. Problems associated with visual clarity and prioritization of functionality/information (e.g., legibility of text, menu organization issues)are the most likely to be found . Costs are relatively low, since checklists can be applied fairly quickly. (This doesn't count the initial cost of developing or acquiring the checklists.)

Expert appraisals require expertise in usability engineering or human-computer interaction. They are likely to uncover problems related to visual clarity, consistency, compability, prioritization of functionality/information, and consideration of user resources. The costs come from salary for the expert; depending on how good he/she/they is, it could vary from inexpensive to moderately expensive.

Cognitive walkthroughs require not just expertise in usability engineering or human-computer interaction, but also familiarity with the target audience. They are likely to uncover problems related to compatibility, feedback, error prevention and recovery, and user control The costs come from salary for the expert; since this is a time-consuming method and there aren't likely to be very many such experts to choose from, it can be relatively expensive.