WITHDRAWN: Changes in Rates of Marijuana and Other Substance Use Among College Students before and after Oregon Recreational Marijuana Legalization

Introduction: There is a growing U.S. trend toward recreational marijuana legalization (RML), but there are few controlled studies of RML effects. We hypothesized that rates of marijuana use would increase more among young adults attending college in Oregon than in comparison states following Oregon’s RML (July, 2015 and start of retail sales in October, 2015). Given the potential for substitution and facilitation effects we also examined changes in other substance use rates. Method: Data were drawn from repeated cross-sectional National College Health Assessment-II surveys administered to random samples of students within participating four-year colleges in the United States from 2008-2016. Data on 18-26 year old undergraduates from institutions that participated both prior to and after Oregon RML were selected. The final sample included two institutions in Oregon (n=7,412) and 123 institutions (n=274,340) in non-RML states. The sample was 66.2% female, and Black, Hispanic, Asian, White, or other race/ethnicity (5.0%, 8.4%, 12.7%, 69.8%, and 4.0%, respectively), 89.2% heterosexual, 45.5% age ≥ 21 years, and attending college in the Midwest, Northeast, South, or West (15.0%, 22.5%, 30.7%, and 31.7%, respectively). Participants self-reported their use (and frequency of use) of marijuana, tobacco, alcohol, and other drugs in the past 30 days. Results: Mixed effects logistic or ordinal regression models accounted for clustering of participants in institutions and controlled for numerous individual-, contextual-, and institution-level covariates (e.g., age, off-campus residence, public/private institution, respectively), as well as secular changes in substance use rates over time. RML was associated with relative increases in marijuana use rates [Odds Ratio (95% Confidence Interval) = 1.29 (1.13-1.48), p = .0002] and frequencies [OR (95% CI) = 1.25 (1.10-1.43), p = .0008], and decreases in tobacco use rates [OR (95% CI) = 0.71 (0.60-0.85), p < .0001], but no significant changes in alcohol or other drug use rates. RML effects for marijuana use did not differ significantly for participants under and over age 21 years. Conclusions: RML is not a single policy or event, but rather a package of policy changes that unfold over time and in the context of other local, state, regional (including neighboring RML states), and national policies. Still, a number of changes in marijuana and tobacco use rates coincided with Oregon's RML and may have important public health implications for young adults. Future studies should examine RML effects on substance-related outcomes of particular relevance to young adults including substance use disorders, motor vehicle accidents, and sexual assault.