DRAFT
Caltech Logo

T&C Chen Center for Social and Decision Neuroscience Seminar

Thursday, February 24, 2022
4:00pm to 5:00pm
Add to Cal
Baxter Lecture Hall
Imprecise Probabilistic Inference from Sequential Data
Michael Woodford, John Bates Clark Professor of Political Economy, Columbia University; Visiting Associate in Economics, Caltech,

Abstract: We document systematic departures from the predictions of Bayesian inference, even on average, in the estimates by experimental subjects of the probability of a binary event following observation of successive realizations of the event. In particular we find under-reaction of subjects' probability estimates to the presented evidence after only a few observations, and at the same time over-reaction to the evidence after a longer sequence of observations. Both these biases in average responses and the variability of subjects' responses across trials are instead consistent with a "noisy counting" model of probability estimation, which allows subjects to give relatively reasonable responses to the experimental task while economizing greatly on both the attention that a subject must pay to their current situation and the degree of control that they exert over their precise response.

For more information, please contact Sheryl Cobb by phone at 626-395-4220 or by email at [email protected].