Examining the Causal Impact of Recreational Marijuana Legalization on Psychotic Disorder-Related Hospitalizations: Evidence from Logistic Regression and Propensity Score Matching
This study investigates the causal effect of recreational marijuana legalization (RML) on psychotic disorder-related hospitalizations in the United States between 2013 and 2017—a period marked by staggered state-level legalization. Utilizing individual-level clinical data from the MH-CLD database and integrating state-level socioeconomic indicators, I apply both logistic regression and propensity score matching to estimate the effect of legalization while controlling for confounding variables such as age, gender, ethnicity, education, income, urbanization, and baseline rates of mental illness and marijuana use. In the full-sample logistic regression model, RML is associated with a statistically significant reduction in the odds of psychotic disorder-related hospitalizations. However, this association disappears after matching, suggesting that the initial effect may have been driven by pre-existing differences between states. The matched analysis reveals no statistically significant relationship between legalization and hospitalization outcomes. These findings underscore the importance of addressing confounding in observational research and contribute to the ongoing policy debate on the mental health implications of cannabis legalization.