Feedback Interventions: Toward the Understanding of a Double-Edged Sword

Feedback intervention (FI), that is, providing people with some information regarding their task performance, is one of the mostly widely applied psychological interventions. Yet there is a growing body of evidence that such interventions yield lughly variable effects on performance (Ilgen, Fisher, & Taylor, 1979; Kluger & DeNisi, 1996; Latham & Locke, 1991; Salmoni, Schmidt, & Walter, 1984). Indeed, in a meta-analysis, we found that although FIs improve performance on average, they reduce performance in more than one third of the cases (Kluger & DeNisi, 1996; see Fig. 1). The latter fact is contrary to the common belief that FIs most often improve performance. Furthermore, we (Kluger & DeNisi, 1996) found no evidence that information about failure (negative FIs) and information about success (positive FIs) have differential effects, on average, on performance. In summary, the data suggest that, at least under certain circumstances, FIs can impair performance and that the processes through which FIs affect performance require more than simple explanations. Although FIs are widely used (e.g., performance appraisals, grades, teaching evaluations), little is known about how they work. As a result, psychologists do not understand when and why FIs might have negative rather than positive (or no) effects on performance. In the present article, we offer an initial explanation of the effects produced by FIs, drawing upon three theoretical constructs that have been developed in connection with control theory: the regulation of feedback-standard discrepancies, locus of attention, and task complexity. These theoretical constructs pertain mostly to the motivational processes induced by FI. The learning processes induced by FI are beyond the scope of this review.2 We begin by tracing the development of the assumption that FIs are always highly effective interventions (for a more thorough review, see Kluger & DeNisi, 1996).