The present experiment was designed to test the theory that psychological procedures achieve changes in behavior by altering the level and strength of self-efficacy. In this formulation, perceived self-efficacy. In this formulation, perceived self-efficacy influences level of performance by enhancing intensity and persistence of effort. Adult phobics were administered treatments based upon either performance mastery experiences, vicarious experiences., or they received no treatment. Their efficacy expectations and approach behavior toward threats differing on a similarity dimension were measured before and after treatment. In accord with our prediction, the mastery-based treatment produced higher, stronger, and more generalized expectations of personal efficacy than did the treatment relying solely upon vicarious experiences. Results of a microanalysis further confirm the hypothesized relationship between self-efficacy and behavioral change. Self-efficacy was a uniformly accurate predictor of performance on tasks of varying difficulty with different threats regardless of whether the changes in self-efficacy were produced through enactive mastery or by vicarious experience alone.