Make a difference? – An experiment on the effectiveness of performance rating policies

Performance appraisals by supervisors are often used in organizations as a basis for bonus and promotion decisions. These subjective appraisals are often prone to supervisors’ biases (e.g., centrality and leniency bias) resulting in ratings that seem not sufficiently differentiated. Previous research shows that forcing supervisors to differentiate in subjective performance appraisals by implementing a forced distribution system may increase worker productivity. However, forced distribution systems pose a variety of other problems. In a laboratory experiment, we investigate two alternative rating policies which aim to foster differentiation and thus productivity. Firstly, supervisors are simply advised to differentiate across different appraisal levels. We find that this simple recommendation results in less lenient and less compressed performance ratings and tends to increase workers’ performance. Secondly, we analyze how supervisors evaluate workers’ performance when the rewards, which are tied to ratings, may not exceed a certain budget. We detect that specifying a financial budget mitigates the leniency bias and increases workers’ performance. Under both rating policies, the performance differences wear off over the course of the experiment.