Objective: Confirm statistical significance of changing trends in marijuana depenence among heavy users. Study Population: Daily/Near-Daily marijuana users, according to self-report from the National Survey on Drug Use and Health (2012-2016). Interventions: none were used, as this paper is based on secondary data analysis. Methods: Plotting time trends, regressions. Statistical Analysis: + quasibinomial regression using R's survey::svyglm from the survey package. Analysis was performed in RStudio. This predicted self-reported dependence based on a DND user's race, gender, age, and educational attainment. The regression was weighted for NSDUH survey weights. + Another quasibinomial regression, this time predicting self-report of a DSM-IV symptom instead of the diagnosis itself. Ethical Considerations: As this used a de-anonymized public dataset, no ethical considerations were identified. Dissemination Plans: This study was submitted to Drug and Alcohol Dependence as a 2000 word piece.