In this study, we examine the ability of SGOMS models to predict human behaviour on two different scales, in micro cognitive task performance and in high level problem solving roles to better understand strategy use and training. To do this, two experiments were designed to isolate the role of knowledge structures in task performance. The first experiment involves modelling an application-based game, played on mobile phones. Results were compared to two models: the SGOMS model that matched the knowledge structures the players had learned during training, and a model optimized for speed, resulting in the fastest game play possible using ACT-R. In the second experiment we examined SGOMS predictions in a high level problem space of an Emergency Operations Center (EOC) simulation, with many interruptions and communication demands, comparing professional EOC managers and undergraduate performance. By comparing results between tasks, HCI design can be augmented using predictive modeling to inform the design to produce efficient and effective training programs.

Additional Metadata
Keywords ACT-R, App, HCI, SGOMS, Training
Persistent URL dx.doi.org/10.1007/978-3-319-91122-9_41
Series Lecture Notes in Computer Science
Citation
West, R, Ward, L. (Lawrence), Dudzik, K. (Kate), Nagy, N. (Nathan), & Karimi, F. (Fraydon). (2018). Micro and macro predictions: Using SGOMS to predict phone app game playing and emergency operations centre responses. In Lecture Notes in Computer Science. doi:10.1007/978-3-319-91122-9_41