Let's play pretend: Towards effective modelling in experimental psycho(patho)logy

Models are omnipresent in science. We introduce a novel framework to achieve more effective modelling practices in general psychology and experimental psychopathology.

We analyze three types of models that are common in psychology: laboratory models, computational models, and cognitive models. We then evaluate two common ways to assess the translational value of models (phenomenological similarity and deep similarity) as well as an arguably underappreciated way (functional similarity). Functional similarity is based on an assessment of whether variables (e.g., the administration of a pharmacological substance) have a similar effect (a) in the model (e.g., in a fear conditioning procedure) and (b) on the real-life target phenomenon (e.g., on real-life anxiety complaints).

We argue that the assessment of functional similarity is a powerful tool to assess the translational value of models in the field of experimental psychopathology and beyond.