Can we learn from wrong simulation models? A preliminary experimental study on user learning

A number of authors believe that wrong models can be useful, providing learning opportunities for their users. This paper details an experiment on model complexity, investigating differences in learning after using a simplified versus an adequate version of the same model. Undergraduate students were asked to solve a resource utilization task for an ambulance service. The treatment variables were defined as the model types used (complex, simple, and no model). Two questionnaires (before and after the process) and a presentation captured participants' attitudes towards the solution. Results suggest differences in learning were not significant, while simple model users demonstrated a better understanding of the problem. This paper consists of a preliminary behavioural operational research study that contributes towards identifying the value of wrong simulation models from the perspective of model users.