Proceedings of the Annual Meeting of the Cognitive Science Society

Decision models are formal algorithms that are used to represent decision processes and predict choice across a wide rangeof disciplines. These models are often highly complex, which makes it difficult to understand the relationships betweendifferent models, the unique features of individual models and, in turn, the fundamental properties of choice behaviorcaptured by these models. We address this issue in a large-scale computational analysis that uses parameter bootstrappingcross-fitting techniques to derive pairwise measures of decision model distances. Our analysis includes over 80 prominentmodels of risky and intertemporal choice, and results in an ontology of decision models, with data-driven model clustersand hierarchies that synthesize over seven decades of quantitative research on human choice behavior.