Sources of self-efficacy and outcome expectations in science, technology, engineering, and mathematics domains: A meta-analysis

Abstract

This meta-analysis examined the structure of the four theoretical sources of self-efficacy (mastery experience, vicarious learning, verbal persuasion, and affective state) and their relations to efficacy beliefs in science, technology, engineering, and mathematics (STEM) fields. Data were derived from 104 studies (including 141 independent samples) conducted across a 37-year period (1977–2013). Based on analysis of all samples, the omnibus test offered support for a two-source factor model of the efficacy sources: direct personal experiences (comprised of prior mastery experience, verbal persuasion, and affective state) and vicarious learning. Both sources were predictive of self-efficacy and outcome expectations. Moderator analyses supported the validity of the two-source model and accounted for substantial amounts of the variance in self-efficacy by gender, race/ethnicity, age group, and type of mastery experience (subjective vs. objective) measurement. Across all moderating conditions, the direct experiences factor was highly correlated with vicarious learning and produced large positive paths to self-efficacy. Vicarious learning yielded small yet negative paths to self-efficacy, which likely resulted from statistical suppression. A greater portion of the variance in self-efficacy was explained when mastery experience was assessed subjectively than objectively. We discuss practical implications of gender and racial/ethnic differences in the model tests along with the implications of the findings for theory and future research.