Substance Use and Risky Sex: A Multilevel Longitudinal Investigation

Substance use has been strongly linked with risky sexual behavior in adolescents and young adults. The relation has been theorized to stem from both being common behaviors for those with strong personality traits of impulsivity and sensation seeking. The vast majority of research linking substance use and high-risk sexual behavior has focused on situational overlap; that is, examining the increased probability of engaging in risky sex when engaging in substance use. The current study, however, focuses on the much less studied global overlap; the extent to which they are related over time. The purpose of this study is to examine the relationship between substance and risky sex among a nationally representative sample of adolescents and young adults longitudinally. Using data collected for the National Longitudinal Survey of Youth 1997, we examined the relationship between risky sex behavior, operationalized as engaging in sex with a stranger, with past-year use of tobacco, alcohol, and marijuana. We hypothesized a strong, positive relationship between risky sexual behavior and the use of each kind of substance, with marijuana expected to be the strongest predictor, followed by alcohol, then tobacco. With our NLSY97 sample (N=8,984), we used a multi-level logistic modeling procedure, modeling the odds of a late adolescent/young adult engaging in risky sex as a piecewise linear-growth function of each type of substance use. The growth function increased from age 16 to a peak at age 21, then decreased to age 29. This design is a particular strength of the study, because it examines this relationship over a much longer time period than previous studies, and also allowed us to look at variation both between and within individuals. We found all types of substance use to be significant predictors of risky sex behavior, as expected, but alcohol appeared  to be the strongest predictor, followed by marijuana and tobacco. These findings extend the literature by showing support for the global association of these behaviors in a longitudinal, prospective study. The relation of our findings to existing theories of unconventional behavior and sensation seeking is discussed.