The Gateway Effect and Predicting Hard Drug Use

University Hall, Rm 134

This paper uses a linear probability model to predict the use of drugs based upon demographic variables as well as information on the use of other substances collected from the National Survey on Drug Use and Health. In addition to analysis of various demographic variables, a two stage regression is used to isolate the marginal effect of tobacco use on marijuana use, and marijuana use on the use of illicit drugs other than marijuana. This use pattern is commonly referred to as a \"gateway drug\" theory. The findings are consistent with a traditional gateway drug theory, finding that the dominant determinant of marijuana use is tobacco use, and that the dominant determinant of illicit drug use is marijuana use. Additionally, several demographic variables are found to have a strong effect on likelihood of using marijunana and illicit drugs.