Narratives in Subjective Performance Evaluations: Do Ratings Change the Narrative?

Abstract

Several high-profile companies have removed numerical subjective performance ratings from their performance-management processes in favor of only using narrative evaluations. Using an experiment, I examine whether requiring supervisors to provide a numerical subjective performance rating in addition to a narrative evaluation moderates the effects of supervisors’ directional evaluation incentives (i.e., reasons to evaluate an employee more or less favorably) on the favorability of their narrative evaluations. As predicted, I find the favorability of supervisors’ narrative evaluations reflects their directional evaluation incentives, but to a lesser degree when they also provide a numerical performance rating. Theory suggests the moderating effects of providing a numerical rating occur because (1) numerical ratings are biased toward the middle of the scale, and (2) supervisors strive for “consistency” between their numerical ratings and narrative evaluations. I demonstrate that the rating process affects narratives, and removing subjective numerical ratings may have unintended consequences in practice.