How realistic should a model be

In this video, Wander Jager discusses the level of realism that is needed for a model.

As you experienced in the last exercises, adding additional behavioural processes to a model leads to an explosion of possible experiments you can run. If you want to model a specific behaviour, you may select a limited number of settings for a simulation model.

Could you use empirical data, e.g. from a questionnaire or interview to decide on which model settings are realistic for a specific case? Please share your thoughts with the other learners.