The predictive validity of self-efficacy in training performance: little more than past performance.

Past research on the influence of self-efficacy in training has provided mixed results. Key differences between studies pertain to whether past performance is operationalized as a residual variable or as an unadjusted variable and to the type of task used. In this study, the authors conducted and performed a reanalysis to examine the influence of self-efficacy using both operationalizations of past performance in 2 experimental tasks. Results indicate that, regardless of task version or type, self-efficacy predicted performance only when a residual measure of past performance was used, but not when past performance was unadjusted. However, when past performance was adjusted, the findings for self-efficacy were likely a statistical artifact. These results suggest that self-efficacy is a consequence rather than a cause of performance in training. (c) 2005 APA