Psychology's Theory Crisis, and Why Formal Modelling Cannot Solve It

Abstract

In light of psychology’s ‘theory crisis’, multiple authors have recently argued that adopting the formalization of theories and/or formal modelling is a necessary or useful step towards stronger psychological theory. In this article, I instead argue that formal modelling cannot solve the core problem the psychological ‘theory crisis’ refers to, which are the currently high degrees of contrastive and holistic underdetermination of our theories by our data. I do so by first introducing underdetermination as an explanatory framework for determining the evidential import of research findings for theories, and showing how both broader theoretical considerations and informal assumptions are key to this process. Then, I derive the aforementioned core problem from current theory crisis literature and tentatively explore its possible solutions. Lastly, I show that formal modelling is neither a necessary nor sufficient solution for either contrastive or holistic underdetermination, and that its uncritical adoption might instead worsen the crisis.

Are Small Effects the Indispensable Foundation for a Cumulative Psychological Science? A Reply to Götz et al. (2022)

“…Moreover, psychologists are typically less interested in prediction than in explanation (Shmueli, 2010), which provides an additional reason to design theory-driven experiments. This is also in line with recent calls to facilitate the generation of reliable knowledge about causal mechanisms by focussing on good description, conceptual analysis, theory building, and severe tests (Oude Maatman, 2021;Scheel et al, 2021b) and not the accumulation of small effects.…”

CRediT where Credit is Due: A Comment on Leising et al. (2021)

“…The target article is the latest to join in calling for more formal theory (e.g., Borsboom et al, 2021;Gray, 2017;Oberauer & Lewandowsky, 2019). While we agree that better theory can help move psychology forward (Maatman, 2021), psychology trainees aren't trained to think theoretically (Bosch, 2018). If psychology is to improve theory, then psychologists must be trained in theory (Smaldino, 2019).…”

“…Finally, students use software for simulations to create models that test theoretical propositions. This program could be split into separate courses where students learn to understand the difference between modeling and theory (Haslbeck et al, 2019), simulate models to test theories (Robinaugh et al, 2020), and investigate incompatibilities and underdetermination in theory (Maatman, 2021).…”

Eight Hypotheses on Technology Use and Psychosocial Wellbeing: A Bicultural Phenomenological Study of Gaming during the COVID-19 Pandemic

“…Currently, the vast heterogeneity of gaming behaviors severely complicates related correlational findings, which are often carried out deductively by instruments that measure constructs with predefined relevance. In other words, hypothesis testing in psychology is often theory driven, but as the ongoing theory crisis has made manifest (e.g., Eronen & Bringmann, 2021 ; Fried, 2020 ; Maatman, 2021 ), theories that are genuinely falsifiable, operationalizable, and useful do not exist in excess. Hence, it can be useful to pursue hypotheses development with explorative qualitative methods, which avoid many of the confirmatory biases attached to dominant paradigms (e.g., Feyerabend, 1993 ), especially in theoretically undeveloped psychological subfields such as that of gaming (Karhulahti, 2022 ).…”