Bayesian Inference for the Distribution of Grams of Marijuana in a Joint

WP 2016-1.0

Greg Ridgeway, Beau Kilmer

As debates about marijuana legalization intensify in the United States and abroad, there is increased focus on creating credible measures of marijuana consumption. This information is not only important for projecting tax revenues and implications for drug tracking organizations, but knowing how much marijuana users are consuming is useful for understanding health and other behavioral consequences. This paper advances the methodology for estimating marijuana consumption by using a large dataset of over 10,000 marijuana transactions spanning 11 years and 43 communities, adapting the Brown-Silverman drug pricing model to these data, and conducting a non-parametric Bayesian analysis to flexibly synthesize marijuana price and weight data.

PDF icon WP2016-01_Ridgeway_Joints_01.22.2016(1).pdf