Examining the Apparent Disconnect Between Risk, Disapproval, and Use of Marijuana Among U.S. 12th Grade Students from 1991-2016

Introduction. Historical trends in adolescent marijuana use have tracked closely with adolescent attitudes of perceived risk of harm and personal disapproval of marijuana use. Risk is hypothesized to associate with use directly, as well as indirectly through disapproval of marijuana use. These associations have been assumed to be generally constant in strength overall and across subgroups. Recent national studies have noted a disconnect between the trends of perceived risk and use of marijuana that raises the question of whether risk continues to act as a constraining force on use. This study examined change across time among adolescents in: (1) the strength of risk/use and disapproval/use associations; (2) interactions between disapproval and risk on use; and (3) differences in associations by gender and race/ethnicity.

Methods: Data were collected from 161,304 12^thgrade students participating in the annual, nationally-representative Monitoring the Future study from 1991-2016. Time-varying effect modeling was used to model change over continuous time in cross-sectional multivariable logistic regression associations between risk, disapproval, and use of marijuana.

Results: Among all 12^thgrade students, associations between risk of regular use and 30-day use, as well as associations between disapproval and use, remained strongly significant across all 26 years examined but the magnitude varied across time, weakening during the late 1990s and early 2000s, strengthening through 2013, and then weakening again. Significant time-varying risk/disapproval interactions on the probability of marijuana use were observed for the sample overall, as well as for females, Black students, and White students. The strength of the risk/use association in the absence of disapproval weakened over time, while the strength of the risk/use association with disapproval strengthened over time. Results indicated significant racial/ethnic differences in the strength of the risk/use association with disapproval overall and across time. Once the risk/disapproval interactions and racial/ethnic differences were accurately modeled, results indicated that perceived risk, when combined with disapproval, was a strong constraining force on marijuana use for all adolescents across time, but the association was strongest for White students and weakest for Black students.

Conclusions: When combined with disapproval, perceived risk is a strong constraining force on marijuana use among US adolescents. The association is dynamic over time and across population groups. Risk without disapproval appears to have an increasingly weaker association with use for the majority of US 12^th grade students. Overall societal trends mask important sub-group differences in these key associations.