THE EFFECTS OF THE MANIPULATION OF PERCEIVED SELF-EFFICACY ON FEAR-AVOIDANT BEHAVIOR (PHOBIA, MODELING, FEAR-AROUSAL)

This study was designed to test Bandura's contention that perceived self-efficacy expectations are a chief determinant of, and bear a causal relationship to, fear-avoidant behavior. Previous attempts to test this proposed causal relationship have not manipulated self-efficacy independently of the duration or type of treatment administered. The effectiveness of the present self-efficacy manipulation was tested in a preliminary experiment using "math-anxious" students. Subjects rated their self-efficacy before and after viewing a modeling film. Four experimental groups received different degrees of "task-relevant" information about the model. The results indicated that the more a model descriptor was related to task-success, the less likely subjects were to increase their self-efficacy ratings following the film. "Low Relevant" information allowed for increases in perceived self-efficacy. "High-Relevant" information enabled subjects to discount the efficacy information value of the modeling, leading to posttest efficacy ratings that did not differ from those of control group subjects who did not view the film.;In the Main Experiment, spider-phobic subjects were assigned to either a "High Self-Efficacy" (HSE) group that received low-relevant information about the treatment film model, a "Low Self-Efficacy" (LSE) group that received high-relevant information, or a Control group that viewed a control film on spider habits. Varying the descriptions of the model led to significant differences in subjects' level and strength of perceived self-efficacy at posttest, even though HSE and LSE subjects received the same modeling treatment. This self-efficacy manipulation was associated with significant differences on most posttest measures of fear-avoidant behavior. MANCOVA's revealed that HSE subjects' anticipatory and performance fear-arousal decreased more after treatment than that of LSE subjects who did not differ from Controls. Univariate analyses revealed a similar pattern of results on four of six-fear-arousal measures. Differences in perceived self-efficacy were not accompanied by significant differences on the Behavior Avoidance Test. The data are largely supportive of the proposed mediational role ascribed to self-efficacy in the process of behavior change, and support Bandura's claim that self-efficacy percepts play a causative role in that process. The results are discussed with regard to the generative nature of self-efficacy, competing theories of operative mechanisms, and suggestions for future research.