Practicing Theory Building in a Many Modelers Hackathon

• DOI: 10.15626/mp.2023.3688
• Publication Date: 2025-03-04T12:36:21Z
• Authors: Noah Van Dongen; Adam Finnemann; Jill De Ron; Leonid Tiokhin; Shirley Wang; Johannes Algermissen; Elena C. Altmann; Štěpán Bahník; Li-Ching Chuang; Andrei Dumbravă; Jens H. Fünderich; Sandra J. Geiger; Daria Gerasimova; Aidai Golan; Judith Herbers; Marc Jekel; Anton Kunnari; Yih-Shiuan Lin; David Moreau; Yvonne Oberholzer; Hannah K. Peetz; Julia Rohrer; Adrian Rothers; Felix Schönbrodt; Yashvin Seetahul; Anna Szabelska; Natasha Tonge; Nicole Walasek; Marlene Werner; Denny Borsboom

Scientific theories reflect some of humanity's greatest epistemic achievements. The best theories motivate us to search for discoveries, guide us towards successful interventions, and help us to explain and organize knowledge. Such theories require a high degree of specificity, which in turn requires formal modeling. Yet, in psychological science, many theories are not precise and psychological scientists often lack the technical skills to formally specify existing theories. This problem raises the question: How can we promote formal theory development in psychology, where there are many content experts but few modelers? In this paper, we discuss one strategy for addressing this issue: a Many Modelers approach. Many Modelers consists of mixed teams of modelers and non-modelers that collaborate to create a formal theory of a phenomenon. Here, we report a proof of concept of this approach, which we piloted as a three-hour hackathon at the Society for the Improvement of Psychological Science conference in 2021. After surveying the participants, results suggest that (a) psychologists who have never developed a formal model can become (more) excited about formal modeling + and theorizing; (b) a division of labor in formal theorizing is possible where only one or a few team members possess the prerequisite modeling expertise; and (c) first working prototypes of a theoretical model can be created in a short period of time. These results show some promise for the many modelers approach as a team science tool for theory development.