On Utilizing Published Prevalence Estimates to Perform Difference-In-Difference Tests: Testing the Impact of Recreational Marijuana Laws

• Activity Number: 119 - SPEED: Government and Health Policy
• Type: Contributed
• Date/Time: Monday, July 30, 2018 : 8:30 AM to 10:20 AM
• Sponsor: Health Policy Statistics Section
• Author(s): Christine Mauro* and Chen Chen and Silvia Martins and Magda Cerdá and Melanie M. Wall
• Companies: Columbia University and New York State Psychiatric Institute and Columbia University and University of California, Davis and Columbia University
• Keywords: difference-in-difference; recreational marijuana laws; policy impact

Abstract

Colorado and Washington passed recreational marijuana laws (RML) in 2012. Two-year aggregated prevalence estimates of past-month marijuana use (MU) were extracted from the National Survey on Drug Use and Health, a cross-sectional survey of the US population. Difference-in-difference (DID) analyses were used to compare changes in MU in these states from 2010-2011 (pre-enactment) to 2014-2015 (post-enactment) to changes in the total US. A naïve estimator treats the survey prevalence estimates without error. To account properly for uncertainty, we develop a method where for each state and year, 10,000 observations are simulated from a normal distribution with mean and standard deviation set to the published prevalence and standard error derived from the 95% confidence interval. This provides 10,000 DID estimates of the change in Colorado and Washington compared to the change in the total US. The 2.5th and 97.5th percentiles are then used to obtain a 95% confidence interval for the observed DID estimate and assess statistical significance. Analyses showed a significant increase in MU among those 26+ in Colorado, but not in other age groups or in Washington.