Highlight
Does Using Marijuan Increase the Likelihood of More Serious Drug Abuse?
Achievement/Results
Does smoking marijuana increase the likelihood of cocaine or alcohol use later in life? Does early-life initiation in drinking alcohol lead to illegal drug use? Such questions fall under the hypothesis often referred to as the “gateway drug hypothesis.” The hypothesis posits that use of less serious substances (e.g., alcohol, marijuana) may increase the likelihood of using more serious drugs later in life (cocaine, crystal methane, heroine). A key methodological issue in empirical research on the gateway hypothesis is that those who use less serious drugs in the first place may be more likely to use all drugs, and hence it is difficult to to empirically disentangle the impact of gateway drugs from the effect of a higher propensity to use drugs in the first place.
IGERT trainee Monica Deza has addressed this question using a dynamic programming model of drug use calibrated to longitudinal data on youth drug use. The key innovation is to compare the impact of having used more serious drugs first on the likelihood of using less serious drugs (presumably, an empirical estimate driving by a higher fundamental propensity to use) to empirical estimates of moving from less serious to more serious drugs (which would be driven by both unobserved heterogeneity and a gateway effect). To the extent that the latter transition rate exceeds the former, there would be evidence of a gateway impact.
Ms. Deza had derived a detailed theoretical model of drug use that when taken to the data, reveal evidence of a small, yet statistically significant, gateway effect of marijuana use on more serious drug use.
Address Goals
This project applies dynamic programming methods to the literature on substance abuse for the first time. It has clearly advanced this epidemiological literature using standard methods from microeconomic and econometric modeling.
