– Educational level and cultural factors moderate individuals’ learning experiences and subsequently their self-efficacy beliefs,
particularly in science, technology, engineering, and mathematics (STEM) domains.
– The results suggest the importance of intervening early in STEM-related academic and career development and attending to
gender and racial/ethnic influences in the types of learning experiences to which individuals are exposed.
– The authors found that variance explained was significantly higher in non-STEM domains (R2 .37) relative to STEM domains (R2 .22), suggesting that other factors are also important in accounting for self-efficacy beliefs related to STEM subjects.
– In all of the subsamples studied, Performance Accompishment (PA) was the dominant influence on self-efficacy. Once PA was statistically controlled, the other three sources were weakly predictive of self-efficacy.
– The authors note that there was no evidence of race/ethnicity as
a moderator of the sources-to-efficacy relationships in STEM
college samples. This suggests that among college students, the
sources are similarly predictive of STEM self-efficacy for nonWhite
and White individuals.
2017 - Unique Effects and Moderators of Effects of Sources on Self-Efficacy: A Model-Based Meta-Analysis
Bandura (1986) postulated in his social–cognitive theory that
self-efficacy beliefs are constructed from and modified through
four primary sources of information: personal performance accomplishments (PA; e.g., past successes, mastery experiences), vicarious learning (VL; e.g., observing the explicit behaviors of others such as role models), socially persuasive communication (SP; e.g., verbal encouragement), and affective arousal experienced while completing a task (AA; e.g., anxiety or stress related to task performance).
The dataset for the meta-analysis included 28 research reports which provided data on 61 unique samples, which included 8,965 participants.
– The authors searched electronic literature databases (ERIC, PsycINFO, SocINDEX, Science Direct) during the academic year of 2013–2014 using the following search terms: source and efficacy, along with mastery experience, performance accomplishment, verbal or social persuasion, emotional or physical arousal, vicarious learning.
– To be included in the meta-analysis research reports had to provide quantitative data for one or more samples on the effects of one or more sources of self-efficacy (PA, VL, SP, and AA) in some academic domain.
– Samples included students ranging from kindergarten through doctoral studies, and the data sources included peer-reviewed articles, book chapters, dissertations, conference proceedings, and unpublished manuscripts.
– Search included relevant manuscripts between 1977 and spring 2014.
– The dataset for the meta-analysis included 28 research reports which provided data on 61 unique samples. Of the 28 research reports, 11 (39%) focused on the mathematics domain, and 10 (36%) on other STEM fields including biological sciences,
physics, technology, and engineering. The remaining 7 (25%)
studies focused on non-STEM domains; six on general academic
self-efficacy and one on efficacy for learning English as a Second
Language.