State Dependent Stochastic Choice Data

By definition, private information is hard to observe. Data engineers dream up revealing data sets. One such is “State Dependent Stochastic Choice” data records the impact of the underlying state of the world on patterns of choice.

A Testable Theory of Imperfect Perception (with Daniel Martin) Economic Journal 2015 uses this data set to characterize Bayesian expected utility maximization

Rational Inattention, Revealed Preference, and Costly Information Processing (with Mark Dean), uses it to characterize rational inattention

Social Learning and Selective Attention (with John Leahy and Filip Matejka) uses it to identify individual preferences from market data.

A second engineered data set is “Choice process” data, which records provisional choices in the period before a decision is finalized.

In “Search, Choice and Revealed Preference”, Mark Dean characterize sequential search and reservation utility stopping rules.

In “Search and Satisficing,” Mark, Daniel Martin and I gather data on the choice process. In our experiments most subjects search sequentially and stop search when a “satisficing” level of reservation utility is realized.