A mathematical and computational formalization of the affect regulation hypothesis of alcohol use
Year: 2026
Type: preprint
Abstract: One of the most enduring ideas in alcohol research is that people drink to regulate negative affect and the stress that causes it. Despite decades of investment the field cannot say when, why, or for whom this occurs. An empirical contradiction illustrates the problem: experimentally induced stress reliably raises both the urge to drink and drinking in the laboratory, but in daily life negative affect predicts the urge without predicting drinking, while positive affect predicts drinking instead. We argue that this impasse is not a failure of data but of theory: the major affect regulation accounts are verbal, too loosely specified to generate testable predictions or to be revised when evidence contradicts them. We address this by formalizing the affect regulation hypothesis as a system of equations and implementing it as a computational model. In the model, affect reaches behavior through the urge to drink, a proximal motivational state; stress raises the value of alcohol, biasing choice; and context gates the path from urge to drink, discounting alcohol's value and, as the situation becomes less permissive for drinking, blocking the drink itself. Through simulation, we show that a single causal architecture reproduces both the laboratory and the daily-life pattern, rendering the contradiction an expected consequence rather than an anomaly. From the model we derive novel, falsifiable predictions and describe the empirical designs that would test or refute them. Most importantly, formalization makes the predictions checkable and revisable. We offer it as a falsifiable starting point for iterative theory development.
Source: PsyArXiv (OSF Preprints)
Authors: Jonas Dora, Kevin Michael King
Language: English
Subfield: Epidemiology
Field: Medicine
Domain: Health Sciences
Open Access status: green
OpenAlex ID: W7203805189