Convergence of multiple rating sources with objective measures of work performance: a meta-analysis

Given the common use of subjective-ratings (e.g., supervisor, peer or self ratings) for performance appraisal, the purpose of this study was to evaluate the extent to which subjective ratings converge with, and account for unique variance in, objective measures of work performance. It is important to determine the extent to which subjective measures (prone to rater biases) converge with measures that do not have the same vulnerabilities. Results demonstrated that peer-ratings had the highest convergence with objective measures (ρ = .31), self-ratings were the next highest (ρ = .20), while supervisor-ratings had the lowest convergence (ρ = .16), these correlations were statistically significantly different form each other. Regression analysis demonstrated that self-ratings accounted for significant variance in objective measures, beyond peer and supervisor-ratings. We also investigated possible moderators of the subjective-objective performance relationship. For strict measures of task performance, peer-ratings were found to have the greatest convergence with objective measures of task performance (ρ = .34). Moreover, the subjective-objective relationship was stronger when both the subjective-rating and the objective measure represented precisely the same construct and when both were at a similar level of specificity