Toward a Psychologist's Guide to Computational Modeling: An Interdisciplinary Scoping Review of Methods for Mechanistic Model Construction

Psychological systems are complex, exhibiting nonlinear behavior, emergence, adaptation, and feedback loops. To better capture this complexity, researchers are increasingly using mechanistic computational models (i.e., computer code designed to represent explanatory principles or rules of a target system from which data can be simulated). Despite this growing interest, there is little guidance on how to most effectively develop and use computational models in psychological research. Other scientific disciplines, such as ecology and chemistry, have used computational modeling extensively for decades and may offer valuable insights from which psychologists can learn. This scoping review synthesizes methodological steps used in developing and evaluating computational models across disciplines, spanning physical and biopsychosocial phenomena. We identified 53 computational modeling roadmaps and conducted a narrative synthesis with frequency analyses. The narrative synthesis is organized according to three research questions: i) what is the general purpose of computational modeling?; ii) what are the proposed modeling steps?; iii) how should researchers move through the modeling steps? For each of these research questions, we also examined whether roadmaps aimed at modeling physical phenomena differ from those that aim at modeling biopsychosocial phenomena. Our narrative synthesis identifies critical gaps in the existing modeling roadmaps and offers a foundation for developing more comprehensive how-to guides for psychologists seeking to integrate computational modeling into their research.