Lilliard E. Richardson

American states have used different approaches in adoption of cannabis policies and continue to modify those policies after approval. States also differ in how long it takes to implement such policies, and this influences the availability of legal marijuana. Such policy differences and implementation timelines could influence usage of marijuana and other illicit drugs by adolescents, young adults, and older adults. We develop an original coding scheme for marijuana legalization policies by classifying policy bundles characterized by three views of marijuana: as a pharmaceutical; as a permissive drug, or as a state fiscal revenue source. We test the impact of state legal marijuana policy characteristics on age group rates of marijuana use with panel regression models including control variables and fixed effects for 2000–2019. This design moves beyond a dichotomous construct of marijuana legalization and accounts for the dynamic adaptation of policies beyond their initial adoption. States with a higher pharmaceutical score experienced lower marijuana usage rates for adolescents and young adults while states with a permissive approach or fiscal approach experienced higher rates of marijuana use for all age groups. We find no consistent spillover effect of the pharmaceutical or permissive marijuana policy bundles on other illicit drug use for any age group, but fiscal bundles show some association with greater illicit drug use for adults. These more nuanced measures better reflect state policies as implemented and provide more clarity of the policy impact on target populations’ marijuana usage.