Features of the School Environment That Moderate Adolescent Marijuana Use: An Application and Extension of the Theory of Planned Behavior

Introduction: Marijuana is the most commonly used illicit drug among adolescents in the United States (SAMHSA, 2014). Schools are key “host environments’ for the delivery of marijuana prevention programming (Bonell et al., 2013). Understanding the school environmental conditions that potentiate or inhibit the relationship between intent and actual substance use is of importance to prevention scientists working in school settings as they represent the potentially modifiable features of the ‘host environment’ that facilitate, or impair, the effectiveness of substance use prevention programming. The Theory of Planned Behavior (TPB; Ajzen, 1991) provides a well-validated framework for exploring the linkages between substance use beliefs and intentions that drive individual adolescent substance use decision-making (Armitage & Conner, 2001). According to this model, attitudes, subjective norms, and perceived behavioral control regarding substance use combine to predict the intention to use, which in turn predicts actual substance use. Although this model has been shown to predict adolescent marijuana use (Kam et al., 2009; McMillan & Connor, 2004), no studies to date have explored the independent contributions of the psychosocial context of the school environment on the substance use decision-making pathways implied by the TPB model.

Method: This study utilized a large statewide youth risk surveillance dataset of youth self-report of substance use and related risk and protective factors as measured by the CTC Youth Survey (Glaser et al, 2005). A total of 217, 276 adolescents in grades 6, 8, 10, and 12 were assessed. Structural Equation Models utilizing FIML followed by tests of moderation of implied pathways were used to test hypotheses. Latent variables representing youth attitudes, subjective norms, and perceived behavioral control regarding marijuana use were regressed on intentions to use and lifetime incidence.

Results: The overall model fit of the Theory of Planned behavior predicting adolescent marijuana use provided good fit to the data (c²(51) = 52848, p < 0.001, CFI = 0.96, TLI = 0.92, RMSEA = 0.06, NFI = 0.96). After controlling for core features of the TPB model, several features of the school environment continued to moderated the relationship between intention and use including academic performance (β = .013, p <.001), school commitment (β = -.02, p<.001), positive home-school communication (β = .09, p<.05), teacher praise (β =.02, p<.001), and rewards for prosocial involvement (β = .09, p<.001). The availability of organized prosocial activities (e.g., sports, clubs) did not significantly influence substance use decisions.

Conclusion: These findings replicate and extend the Theory of Planned Behavior as applied to adolescent marijuana use, and provide additional evidence of how features of the school context directly influence adolescent marijuana use decision-making. Greater attention to these organizational variables may help potentiate the effect of universal school-based drug prevention programming.