Recreational cannabis laws had limited overall effect on drinking and smoking, but varied by group

Using law passage as the measure, recreational cannabis laws showed no association with alcohol or tobacco use. But using operational dispensary opening, these laws were associated with a 0.95 percentage point decrease in current drinking and 0.48 percentage point decrease in current smoking. Subgroup analysis revealed heterogeneous effects across demographics.

Researchers analyzed cross-sectional data from 4.8 million adults in the 2012-2022 Behavioral Risk Factor Surveillance System using a difference-in-differences approach, adjusting for individual characteristics and state-level factors. Two definitions of cannabis law implementation were tested: law passage and operational dispensary opening.

Whether cannabis legalization increases or decreases use of other substances has major public health implications. This study suggests the answer depends on how you measure legalization and which population groups you examine, adding important nuance to a polarized debate.

The distinction between law passage and dispensary operations is key. Laws on the books may have little effect until people can actually purchase cannabis legally. The heterogeneous effects across demographics suggest that cannabis substitution for alcohol and tobacco is not universal but depends on factors like age, sex, race, and socioeconomic status.

Cross-sectional survey data with self-reported substance use. The difference-in-differences approach relies on parallel trends assumptions that may not hold for all states. The small effect sizes, while statistically significant, may have limited clinical or public health significance. Cannot distinguish between substitution and complementary use.