ABSTRACT:
I will give an introduction and overview of the technical side of my work in behavioral decision research. The core goal of my research program is to model invariance of hypothetical constructs (e.g. preference) and heterogeneity in behavior (choices) jointly. So-called random preference models of decision making can be characterized by convex polytopes. These, in turn, define order-constraints on choice probabilities that require specialized frequentist or Bayesian data analytics. In this talk, I review how to test transitivity of preferences empirically and how to generalize distribution-free random utility and random preference models so as to encompass context-dependence. Empirical application areas include Economics, Marketing, Psychology, and Zoology.
BIOSKETCH:
Prof. Regenwetter’s primary interests can be categorized as falling within three paradigms: probabilistic measurement, social choice, and preference evolution over time. Probabilistic measurement theory reformulates axiomatic measurement structures (e.g., in decision theory) in a probabilistic framework and thereby makes them empirically (and statistically) testable. Social choice theory is the theory of aggregating individual preferences or choices into a social ordering or choice. Dr. Regenwetter’s interest in social choice is behavioral. Using random utility models as measurement tools, he evaluates and compares competing social choice functions on empirical data of various kinds. Dr. Regenwetter studies preference change over time via stochastic process models in which random utilities are indexed by continuous time.
