Too good to be true? Are supervisor-perspective ratings a valid substitute for actual supervisor ratings?

Due to well-known problems with self-ratings of job performance (e.g., inflation, weak correlation with supervisor ratings) and the challenges of collecting supervisor ratings of job performance, researchers sometimes use supervisor-perspective ratings (e.g., "how do you think your supervisor would rate your job performance?") instead. The assumption is supervisor-perspective ratings are less affected by the noted issues with self-ratings and therefore are more similar to actual supervisor ratings than traditional self-ratings. In fact, a considerable number of researchers have used supervisor-perspective ratings as an alternative to actual supervisor ratings. The purpose of this study is to meta-analytically determine the degree to which supervisor-perspective ratings are a valid substitute for actual supervisor ratings and identify the boundary conditions for this substitution. Our meta-analyses demonstrate that supervisor-perspective ratings are generally not a viable substitute for actual supervisor ratings. This is especially the case when (a) citizenship performance is measured, (b) data are collected in collectivistic cultures, and (c) all study data are gathered from the same source. We recommend not using supervisor-perspective ratings as a substitute for actual supervisor ratings. (PsycInfo Database Record (c) 2023 APA, all rights reserved).