This paper describes a model-based analysis of errors in expert use of graphical user interfaces. Two mechanisms, speed-accuracy tradeoffs and attention failures, were simulated by a model consisting of processes for display elaboration, selection of objects to be operated on, and selection of a next action. Both selection processes are performed in a context defined by the elaboration process, simulated by a sampling mechanism which uses knowledge in long-term memory for display comprehension. The model provides a well motivated account of errors made by expert users. Finally, the use of the model in interface evaluation is discussed.
Muneo Kitajima, Peter G. Polson