D. Harris (Ed.): Engin. Psychol. and Cog. Ergonomics, HCII 2007, LNAI 4562, pp. 477–486, 2007. © Springer-Verlag Berlin Heidelberg 2007 Cognitive Model Data Analysis for the Evaluation of Human Computer Interaction Jeronimo Dzaack 1 and Leon Urbas 2 1 Berlin University of Technology, Center of Human-Machine Systems Franklinstraße 28-29, Sekretariat FR 2-7/2, D-10587 Berlin Jeronimo.Dzaack@zmms.tu-berlin.de 2 Dresden University of Technology, Institute of Automation D-01062 Dresden Leon.Urbas@tu-dresden.de Abstract. In industry and consumer electronic, more and more operative tasks are changing to supervisory control and management tasks. This leads to more complex and dynamic user interfaces (e.g. integrated control interfaces, info- tainment systems in cars). Because of the integrated functionality and the com- plex data structures, these interfaces require more cognitive information processing. Usability of such interfaces can be evaluated by using cognitive modeling to investigate cognitive processes and their underlying structures. So far the explanatory power of cognitive models is limited due to the lack of fine- grained simulation data analysis. Having realized this drawback we developed SimTrA (Simulation Trace Analyzer) to simplify the analysis of cognitive mod- els. The tool automatically processes and analyzes data from cognitive models and allows the comparison of simulated data with empirical eye movement data. This paper describes the approach and its implementation. The practicabil- ity of SimTrA is demonstrated with an example in the domain of process control. Keywords: Cognitive Architectures, Eye Movement Data, Analysis, Human Computer Interaction. 1 Introduction Recent introductions of new information technologies in the field of dynamic human- machine systems (e.g. process control systems in the chemical industry or airplane cockpits) have led to increasing cognitive requirements caused by a shift from manual process operation to the management of complex automated processes. This calls for highly complex dynamic user interfaces. Because of the integrated functionality and the complex data structures, these interfaces require more cognitive information proc- essing. In a practical application cognitive models can be used to evaluate the usabil- ity of prototypes. This helps to detect errors in the interaction design of interfaces and gives indications about the cognitive demands of the future user. Cognitive models extend classical usability methods (e.g. questionnaires, observing) and expand the