Chill Out or Get Active?: The Relationship Between Marijuana Use and Exercise

Marijuana use is often associated with a sedentary lifestyle, including the consumption of unhealthy foods, lack of motivation to engage in social activities, and disinterest in exercise. This notion aligns with the view that (un)healthy behaviors generally cluster. However, little research has investigated these relationships using large datasets and advanced statistical methods. Specifically, previous studies employ limited measures for exercise and marijuana use in their analyses, which leads to an incomplete understanding of these relationships. The present study partially fills this gap by analyzing the relationships between several measures for current and lifetime use of marijuana and various forms of exercise. The core dataset is Wave 4 of the National Longitudinal Study of Adolescent to Adult Health (Add Health). Numerous gender-specific models are estimated with a large set of control variables, binary and continuous measures of exercise and marijuana use, and alternative estimation techniques. Regardless of gender, results consistently show that marijuana use is not significantly related to exercise, which is counter to conventional wisdom that marijuana users are less likely to be active. Indeed, the only significant relationships appear for occasional marijuana users (especially among females), who are more likely to exercise relative to non-users. For example, females who report having smoked marijuana at some point in their lives are 1.12 times more likely to exercise than non users (p = 0.012). These findings endure when subject to various robustness checks and sensitivity analyses. In particular, we estimate similar models using another large and nationally-representative dataset—Behavioral Risk Factor Surveillance System (BRFSS)—and obtain very similar results. In addition to cross-sectional analysis of Wave 4 of Add Health, we also conduct fixed-effects analysis using Waves 3 and 4 to partially address endogeneity concerns. Again, results are similar to the core cross-sectional findings. As additional states legalize medicinal and recreational use of marijuana, this topic holds important social and policy importance when we consider possible negative consequences of expanding the intensive and extensive margins. This research is particularly relevant to the health promotion and risky behaviors literature as it offers evidence that health-related behaviors may be independent of each other. While endogeneity bias remains a possible concern with these analyses and the results are not necessarily causal, a first look at these relationships suggest that marijuana users do not engage in less exercise than non-users.