Multisource Performance Management: Comparing Rater Convergence and Leniency Using the Graphic Rating Scale and the Relative Percentile Method

Abstract: Multisource performance management systems are common but often face issues with low rater convergence and inflated ratings. We studied the effectiveness of the relative percentile method (RPM) compared to the Graphic Rating Scale (GRS) in enhancing rater convergence and reducing inflated ratings. In our experiment with 263 participants, organized into groups, each participant developed sales pitches for evaluation by their peers, supervisors, and themselves. The findings revealed that GRS led to significantly greater self-supervisor rater convergence than the RPM. However, the RPM led to lower rater leniency across peer and self (but not supervisor) ratings. The implications for multisource performance evaluations are discussed.