Psychometric properties of multisource performance ratings: A meta-analysis of subordinate, supervisor, peer, and self-ratings

Abstract

The purpose of this investigation was to examine the psychometric properties (interrater reliabilities within source and correlations between sources) of subordinate, supervisor, peer, and self-ratings of job performance. Different job types and dimension types were compared. Using meta-analytic methodology, we found that subordinates showed the lowest mean reliability (.30) and supervisors showed the highest (.50), with peers in between (.37). Mean correlations between sources were low for subordinate ratings (.22 with supervisor, .22 with peer, and .14 with self-ratings) and for self-ratings (.22 with supervisor and .19 with peer ratings). The mean supervisor-peer correlation was higher at .34. Both reliabilities and correlations between sources tended to be higher for nonmanagerial and lower complexity jobs. Comparisons of between-source correlations with within-source reliabilities indicated that, with some qualifications, the different sources had somewhat different perspectives on performance. Dimension reliabilities differed somewhat for interpersonal and cognitive dimensions.