Learning under Ambiguity: An Experiment in Gradual Information Processing

This experiment studies belief updating under ambiguity, using subjects' bid and ask prices for an asset with ambiguous payoff distribution. Bid and ask quotes allow for distinguishing between the two main paradigms of updating under ambiguity--full Bayesian updating and maximum likelihood updating. We find substantial heterogeneity in subjects' reaction to information. The majority of subjects (54\%) chose quotes that were in line with full Bayesian updating, while another, non-negligible, group (35\%) behaved like maximum likelihood updaters. (JEL: G11, C91, D81)