Department of Economics

How does overconfidence arise and how does it persist in the face of experience and feedback? In an experimental setting, we examine how individuals’ beliefs about their own performance on a quiz react to noisy, but unbiased feedback. In a control treatment, each participant expresses her beliefs about another participant’s performance, rather than her own. On average, they express accurate posteriors about others’ scores, but they overestimate their own score, believing themselves to have received ‘unlucky’ feedback. However, this driven by overconfident priors, as opposed to biased information processing. We also find that, while feedback improves estimates about the performance on which it is based, this learning does not translate into improved estimates of related performances. This suggests that people use performance feedback to update their beliefs about their ability differently than they do to update their beliefs about their performance, which may contribute to the persistence of overconfidence.