CS252. User Interface Design

Lecture Outline for Feb. 16
Empirical Methods in UCD


1.     Overview of Evaluation Methods
     We've studied the basic strategy for UCD
          (1)

          (2)


          (3)

          (4)


     What is usability evaluation?





     "Empirical" and "non-empirical" methods



2.     Empirical Methods
     13 basic empirical methods:

     Private camera conversation:



     Co-discovery:



     Focus group:



     User workshop:



     Think-aloud protocol:



     Incident diary:



     Feature checklist:



     Software logs ("logging use"):



     Field observation (also called contextual study):



     Questionnaire:



     Interview:



     Valuation method:



     Controlled experiment:


     Designing empirical questions/tasks
          Style influences ease of analysis
          Unstructured (open-ended) questions/tasks:




          Structured questions/tasks:




3.     Questionnaires versus Surveys
     "User surveys" are everyone's first thought for getting user input


     What makes a good survey? 





     Cross-validation:


     Consistency checking:


     What a user survey really involves:





     Difference between questionnaires and surveys



4.     Designing Questionnaires and Feature Checklists
     Question formats
          see handout

     What goes wrong in questionnaires (example from [Shneiderman 1992])
           length:


          wordiness/reading time:


          number of terms/choices: 


          ambiguity:


          double-takes:


          gratuitous items: 


     Improving questions
          see handout





     Case Study: questionnaire re-design (for IBM)
          Original version of electronic survey



           Revised version



          Example of reformulation