Peer Evaluation and Team Performance: An Experiment on Complex Problem Solving

Abstract

Today's employees often work in teams on complex problems. Yet, we know very little about how to incentivize such work. Using a laboratory experiment, we study how groups performed on guesstimation problems under the following incentive schemes: fixed pay, pay that depends on group performance, and pay that depends on group performance as well as on the outcome of a peer evaluation. Overall, group performance was not affected by either kind of incentive. Yet, groups behaved significantly differently under peer evaluation. Subjects reported higher motivation and groups worked harder and longer in the group phase. Our results suggest that subjects based their peer evaluation not on the actual correctness of an individual's answer but instead used more easily observable proxies of effort. Exploratory analyses suggest that female majority groups performed significantly worse under peer evaluation, while the performance of male majority teams was not affected. We conclude that peer evaluations are an imperfect incentive mechanism when group members cannot easily assess the quality of individual performance.