Misperceiving bad news: the effect of feedback on task motivation and belief updating

with Sabine Fischer, Maria Kleshnina, and Manuel Staab

Utilising a Bayesian framework, we examine in a multi-stage lab experiment how individuals update beliefs about their abilities based on their own experiences as well as external feedback. Beyond the general issue of belief formation, we focus on how and to what extent external performance feedback can contribute to a more accurate self-assessment and improved outcomes. More specifically, participants are asked to solve a set of logical puzzles in two comparable rounds and are incentivised to accurately guess their performances. After the initial round, they receive feedback (treatment) regarding their performance, or not (control). We find that, on average, individuals in both groups adjust their guesses in the right direction after the initial round, providing evidence that they learn from experience. This holds for participants who are under- and overconfident in their ability (i.e. assess their performance initially as too low/high). Feedback does not improve the average accuracy of guesses, but this is driven by strongly overconfident individuals who seem to ignore the information from external feedback (but nevertheless adjust their performance estimates downward in the treatment and control group). Under- and moderately overconfident participants improve their guesses with feedback. Beyond the effect on self-assessment, negative feedback appears to crowd out task motivation, leading to a lower performance in the second round.

(The Working Paper is in Progress.)