Main and Interactive Effects 0 Running head: MAIN AND INTERACTIVE EFFECTS Main and Interactive Effects of Controllability and Generalisability Attributions upon Self- efficacy

Objective: This study examined main effects of controllability and interactive effects of controllability and generalisability attributions upon self-efficacy. Method: Participants (N = 360; mean age 21.64, SD = 6.96 years) completed measures of precompetition self-efficacy (one hour prior to competition one), attributions (one hour after competition one), and subsequent self-efficacy (at least one week following competition one and one hour prior to competition two). All measures were completed in reference to sport competitions. Results: Demographic variables and pre-competition self-efficacy were entered as control variables in moderated hierarchical regression analyses. Results demonstrated that individuals who perceived performance as more successful, had higher subsequent self-efficacy when they generalised (ΔR 2 = .38, p < .01) causes of performance across time (stability: b = .45, p < .01), and/or across situations (globality: b = .49, p < .01), and/or perceived causes to be unique to themselves (universality: b = -.46, p < .01). Individuals who perceived performance as less successful, had higher subsequent self-efficacy when they viewed causes of performance as controllable (ΔR 2 = .10, b = .25, p < .01); an interaction (ΔR 2 = .07, p < .01) for controllability and globality (b = .22, p < .01) demonstrated that if causes were perceived to be global, higher levels of controllability were associated with higher levels of subsequent self-efficacy. Conclusion: This study provides evidence, following more successful performances, that attributions to generalisability (stability, globality, and universality) affect self-efficacy; following less successful performances, globality (a generalisability dimension) moderates the effect of controllability upon self-efficacy.