First, we plotted and summarized the prevalence of past-month marijuana use by year and state and MML status (see footnote of Figure 1). Second, for each year, we performed 2-sample t-tests comparing the outcomes in states which had passed MML by that year to states that had not passed MML by the same year. For example, we compared the prevalence of past-month use in 8 states with MML in 2002 to the remaining 42 states that had not passed MML by 2002. Third, we performed a longitudinal analysis including a random intercept for state and a fixed linear trend for year to compare the prevalence of marijuana use and perceived riskiness in the years prior to MML passage (data available for 8 states prior to MML) to that of: 1) post-MML years in states that passed MML and 2) all years for states that did not pass MML by 2011.
