This investigation was designed to determine the effects of preexisting and manipulated self-efficacy on competitive motor performance. Male (n = 46) and female (n = 46) subjects were classified as being high or low in preexisting self-efficacy before the experiment began and were randomly assigned to either a high- or low-manipulated self-efficacy condition in a 2 × 2 × 2 (sex by self-efficacy by manipulated efficacy) design. Efficacy was manipulated by having subjects compete against a confederate on a muscular leg-endurance task where the confederate was said to be either a varsity track athlete (low-manipulated self-efficacy) or an individual who had had knee surgery (high-manipulated self-efficacy). To create aversive consequences, the experiment was rigged so that subjects lost in competition on the two muscular leg endurance task trials they performed. Both preexisting and manipulated self-efficacy were found to significantly influence performance, with preexisting self-efficacy influencing performance only on Trial 1 and manipulated self-efficacy only on Trial 2. The findings support Bandura's (1977) theory of self-efficacy and are discussed in terms of the permeability of initial efficacy states.