Abstract
A componential model is discussed for understanding the behavioral impact of user interface metaphors. The model expresses the effect of any metaphor in terms three generic components, each component corresponding to a way a metaphor can match or mismatch interface functionality. These component scores have to be estimated in some combination of task, user, and interface analysis. The main use of the model is for researchers who want to be better able to retrospectively explain or (generalizing past data into the future) predict the impact on human behavior of single and composite user interface metaphors. In order to validate the model, an experiment was conducted in which two interface metaphors were independently manipulated (present or absent) and their effect was studied on the performance of users performing several web navigation tasks using a simulated web site. The results showed that the predictions made by the model were generally confirmed by the data. These data also illustrate how the model may be used to describe unexpected negative effects of metaphors on user performance in terms of matches and mismatches. Finally, they suggest ways in which user interface metaphors can be redesigned in the future in order to better fit human cognition.
Get full access to this article
View all access options for this article.