We present a theoretical framework and empirical strategy to measure the causal effect of interim feedback on individuals’ performance. Our identification strategy exploits a natural experiment in a leading UK university where different departments have historically different
rules on the provision of feedback to their students. Our theoretical framework makes precise that if feedback provides students with a signal of their true ability, then the effect of feedback depends on the balance of standard substitution and income effects, and whether students are over or under confident with regards to their true ability. Empirically, we find the provision of feedback has a positive effect on student’s subsequent test scores throughout the ability distribution, with the effect being more pronounced for more able students. We find no evidence of any individuals being discouraged by feedback. The results are reconcilable with the theory if preferences and beliefs are correlated so that underconfident individuals exert more effort when their ability is revealed to be higher than they expected and vice versa. The results have implications for the interplay between the provision of feedback and incentives in organizations.