When the chips are down: Effects of attributional feedback on self-efficacy and task performance following initial and repeated failure

In two experiments, we manipulated the controllability and stability of causes of failure and explored the impact of these factors on self-efficacy and performance. In Experiment 1, participants (N=80; mean age 20.0 years, s=1.0) were provided with false negative feedback following performance on a blindfolded dart-throwing task. Consistent with theory and recent research, an induced belief that failure was beyond control and unlikely to change led to lower self-efficacy and poorer performance (all F_1,75>5.49, all P<0.05, all η²=0.01). A second experiment (N=80; mean age 22.0 years, s=2.1) demonstrated that following an induced belief that failure was beyond control and unlikely to change, only new perceptions that a repeated failure was within one's control and likely to change resulted in higher self-efficacy and improved performance (all F_1,75>4.53, all P<0.05, all η²>0.004). All effects were mediated by self-efficacy: Sobel's (1982) test, all z>1.97 (in absolute magnitude), all P<0.05, all r>0.22 (in absolute magnitude). These findings suggest that in novel circumstances individuals believe in the best for themselves unless possibilities to self-enhance are explicitly precluded, and only reinvest efforts when opportunities for self-enhancement become clearly admissible.