Liking and similarity as predictors of multi‐source ratings

Abstract

This study explored the ability of liking and two forms of rater‐ratee similarity to predict multi‐source proficiency ratings of managerial competencies. Findings from the regression model that included all raters indicated that technical proficiency, rater‐ratee liking, demographic and attitudinal similarity about work all were significant predictors of proficiency ratings. For these ratings, attitudinal similarity was the strongest predictor of ratings after technical proficiency. Liking, attitudinal and demographic similarity combined increased explained variance by about one‐third over technical proficiency. Regression models, run for each rating source separately, suggested that supervisors, subordinates and peers may weigh performance and interpersonal elements differently, when making performance ratings.