The study has several strengths. We used an annually-collected survey, which can generate state-level prevalence estimates, to provide timely data on marijuana-related variables in Colorado during a period when the Colorado commercial marijuana industry experienced rapid growth. Given that youths with more serious substance involvement are less likely to attend school and more likely to drop out (Younge et al., 1996), using the NSDUH reduces a potential school-based-survey selection bias. In addition, we focused on a single state where changes in medical marijuana registration are well-documented and examined multiple age groups and multiple use and disorder phenotypes. We also assessed change pre-post growth of medical marijuana commercialization, not pre-post medical marijuana law enactment.
