MODELING EASE OF LEARNING OF USER INTERFACES

This article presents a task-analytic methodology for evaluating user interfaces in terms of ease of learning and discusses the results of a n empirical study designed to test the validity of the methodology. Th e proposed methodology represents a tool that enables user interface d esigners and human factors practitioners to predict the ease of learni ng of a given interface by a particular user. Ease of learning is meas ured by the cognitive learning time which is modelled as a function of the cognitive complexity associated with learning a given interface a nd the learner's prior knowledge. Measures of cognitive complexity and prior knowledge are described and tested. The results of the empirica l study are promising. The methodology accurately predicted the relati ve ease of learning of 27 text-editing tasks and accounted for more th an 77% of the variance among the observed learning times.