Uncertainty, Information, and Procedures

This dissertation contains three chapters that explore decision-making under uncertainty using three different approaches. Chapters 1 and 2 examine belief updating experimentally and empirically, while Chapter 3 investigates theories of choice under risk within a procedural framework. Chapter 1, titled "Confidence in Inference," considers a decision-maker who chooses between objects, each associated with a sample of signals. I axiomatically characterize the set of choices consistent with established belief updating models. A simple thought experiment reveals a natural choice pattern that lies outside this set. Specifically, these models cannot rationalize the effect of increasing sample size on choice. In a controlled experiment, 95% of subjects' choices violate models of belief updating. Using a novel incentive-compatible confidence elicitation mechanism, I find confidence in correctly interpreting samples influences choice. As suggested by the thought experiment, many subjects neglect the sample size between large samples. This behavior is positively associated with higher confidence. Chapter 2, titled "Updating and Misspecification: Evidence from the Classroom," is a joint work with my PhD colleagues Marc-Antoine Chatelain, Paul Han, and Xiner Xu. We empirically investigate and quantify the extent to which misspecification leads to failures in Bayesian updating. In an incentive-compatible manner, we collect a rich, high-frequency dataset on students' beliefs within the context of first-year courses. By examining students' beliefs about their grades for each test and the noisiness of testing, we recover, for the first time outside the lab, measures of misspecification, updating biases, and their evolution over time. Initially, students are overconfident and fail to correct their beliefs due to overestimating testing noise and non-Bayesian updating. We conducted a randomized controlled trial (RCT) in which treated students received non-personal information about the testing noise. The treatment corrects students' perceptions of testing noise and improves their updated beliefs. Our findings suggest that misspecification contributes significantly to the failure of updating but can be alleviated through relatively simple interventions. Chapter 3, titled "Procedural Choice under Risk," studies various models of non-expected utility (non-EU) from a procedural perspective. I examine the process of merging a binary sublottery into what the decision-maker (DM) considers equivalent; by repeatedly performing this process, the lottery simplifies to a degenerate lottery. I show that a DM who uses this procedure costlessly and consistently is necessarily an expected utility maximizer. This framework allows for the direct connection of various non-EU theories with their procedural mechanisms. I demonstrate that complexity cost models, rank-dependent expected utility, betweenness models, and quadratic utility are all nested via different relaxations of the procedural regularities exhibited by the DM. This work provides a procedural foundation for these models, and by modeling the procedures explicitly, I offer insights into the psychological mechanisms underlying these models.

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.