Social Learning in Action: Predictors of Substance Refusal Efficacy in Middle School Youths

Background and Purpose: Peer norms play a powerful role in shaping youth behavior during early adolescence, a developmental period when youth are especially sensitive to social influence. Although substance use (SU) behavior trends remained stable through 2023 (NIDA, 2023), recent data show that 5.9% of U.S. students reported current e-cigarette use, including 3.5% of middle schoolers (CDC, 2024). Guided by Social Learning Theory (Bandura, 1977), this study examines how peer SU, peer SU disapproval, and perceived SU harms influence substance refusal efficacy. To account for individual-level factors, emotional regulation and parental education were also examined. This study builds on prior research by examining the combined effects of peer SU, peer disapproval, and perceived SU harms in a diverse middle school sample. Using a theory-driven, multi-model design, this study aims to inform prevention strategies and strengthen SU prevention efforts.

Methods: Baseline data, collected at the pre-test stage before any intervention was delivered, came from 1,457 students enrolled in public middle schools participating in a SU prevention program. Data were used to assess refusal efficacy, peer SU, peer SU disapproval, emotional regulation, and perceived SU harms related to SU, using validated Likert-scale items (α ≥ .88). Demographic variables included gender, race, and parental education. Four OLS (ordinary least squares) regression models examined the effects of peer, demographic, and contextual variables on refusal efficacy.

Results: The sample was racially diverse (48% Black, 22% White, 18% Latinx, Others <8%) and evenly split by gender (49.5% female, 50.5% male). Most students were in the 7th grade (66.4%). A majority of parents had at least some post-secondary education (mothers: 66.9%; fathers: 56.4%). Model 1 (R² = .30), which included peer influences only, significantly explained 30% of the variance in refusal efficacy. Parental education was added in Model 2 (R² = .30), which remained significant. In Model 3 (R² = .28), emotional regulation replaced parental education and was also significant. Model 4 (R² = .29), including all predictors, was significant, F (19, 977) = 21.05, p < .001, and explained 29% of the variance. Peer disapproval (b = 0.45, SE = 0.08, p < .001), SU harms perception (b = 0.50, SE = 0.04, p < .001), and peer SU (b = -0.18, SE = 0.04, p = .003) remained significant predictors. Emotional regulation was also a significant predictor (b = 0.21, SE = 0.09, p < .05), though parental education, gender, and race were not.

Conclusions and Implications: Findings support Social Learning Theory, with peer norms and perceived SU harms emerging as strong predictors of refusal efficacy. Youth who perceive higher peer disapproval and greater SU harm perception report stronger confidence in resisting use, whereas those exposed to peer use report significantly lower efficacy. While peer influence was the most salient factor, effects of emotional regulation suggest additional leverage points for intervention. Prevention strategies should prioritize reshaping peer norms, correcting SU related harm misperceptions, and strengthening emotional regulation skills. Longitudinal research is warranted to examine causal mechanisms and enhance peer-focused strategies in diverse school settings.