Substance Use-Related Perceptions and Adolescent Substance Use: A Multi-Level Model Approach to Understanding Alcohol, Tobacco and Marijuana Use Among Alaska High School Students

BACKGROUND:  Declines seen over the past decade in lifetime use of tobacco, alcohol, and marijuana by Alaska adolescents have recently leveled off. To continue to reduce adolescent use of these substances, and the related negative health outcomes, we need to identify the health beliefs that are most highly associated with use, and thus may become targets for intervention. Research suggests individual beliefs, such as perceived harm of a substance and perceived parental permissiveness related to that substance, may be predictive of adolescent substance use. We sought to examine the contribution of community-level (aggregate) beliefs in addition to individual-level beliefs when predicting adolescent use of alcohol, cigarettes, and marijuana.

METHODS:  We obtained data from the Youth Risk Behavior Survey (YRBS), which includes responses from students in grades 9-12 across the state of Alaska.  We modeled alcohol, tobacco, and marijuana use as a function of individual-level variables (e.g., age, gender, and race) and both individual- and community-level (school district) measures of parental permissiveness, “coolness”, and perceived harm of use of each substance across a representative statewide sample of Alaska school districts. We utilized multilevel logistic regression to determine the impact of community-level beliefs on alcohol, tobacco, and marijuana prevalence among Alaska adolescents, controlling for individual-level variables.     

RESULTS: Preliminary results indicate community-level beliefs are modestly associated with substance use among Alaska youth, beyond any association with individual-level beliefs; however, the type of belief that increases the odds of use varies depending on the substance. In our model of smoking prevalence, community-level parental acceptance of smoking was associated with greater individual odds of smoking (OR=1.05; p<.05). In our model of marijuana use, lack of community-level perception of harm was associated with greater odds of adolescent marijuana use (OR=1.03; p<.05). Finally, community-level perceptions related to both coolness (OR=1.03; p<.05) and lack of harm (OR=1.03; p<.05). were associated with adolescent alcohol prevalence. 

CONCLUSIONS: Both individual-level and community-level beliefs about adolescent substance use are associated with adolescent substance use. Furthermore, the pattern of associations varies across substances (e.g., parental permissiveness appears to be more important for smoking compared to alcohol or marijuana use). The findings underscore the need to consider community norms when developing interventions to address adolescent substance use.