Age–period–cohort influences on trends in past year marijuana use in the US from the 1984, 1990, 1995 and 2000 National Alcohol Surveys

BackgroundPrevious studies have described trends in marijuana use in the US and examined age–period–cohort (APC) effects finding increased lifetime use among cohorts born after 1945. However, no studies have utilized data on current consumption in multiple cross-sectional surveys to estimate these factors.MethodsAge–period–cohort models including demographic factors are estimated using logistic regressions in four US National Alcohol Surveys (NAS) conducted between 1984 and 2000. Trends in past year marijuana use are also evaluated.ResultsMarijuana use declined over the study period from 10% to 7.2% of the population. Declines were mainly seen among men, resulting in a degree of gender convergence, particularly for those aged 18–25. Significant effects of age, period and cohort were found, with steep declines in use by age from the early 20s to the 40s. All male cohorts born after 1945 and female cohorts born between 1945 and 1960 showed elevated prevalence compared to earlier cohorts.ConclusionsTrend results from the NAS differ from those in other surveys and indicate decreased prevalence of past year marijuana use and gender convergence. APC results confirm past findings of age effects and cohort differences between those born before and after 1945. Marijuana use presents many measurement difficulties and future research to understand differences across surveys is needed.