A Latent Class Analysis of Substance Use and Risky Sexual Behavior for College Students

High tolerance of peer substance use (Neighbors et al., 2007) and normative perceptions of substance use (Elek et al., 2006) may increase substance use for young adults attending college. Further, the transactional associations between substance use and sexual activity may lead to increases in both risky sexual activity and substance use (King et al., 2012; Parks et al., 2012). Thus, prevention and intervention programs that target only one of these risk behaviors may not result in significant reductions in risk behavior for college students engaging in both risky sex and substance use. The goal of the present study was to identify groups of college students who engage in distinct and meaningfully different patterns of substance use and risky sexual activity.  

College students (N=656; M=20.52 years; 70% female) completed an online survey for extra credit in their psychology course. Risky sexual behavior was assessed with items that included having had an opposite or same sex casual sex partner, pregnancy, or STD diagnosis. Substance use was assessed with items including tobacco, drunkenness, marijuana, alcohol and marijuana use in one sitting, and use of illicit drugs (i.e., inhalants, hallucinogens, narcotics, amphetamines, cocaine, prescription stimulant misuse).

Latent class analyses were conducted with substance use and risky sex items using Mplus v.6.1 (Muthén & Muthén, 2008). Results suggested that four classes provided the best fit. The Lo-Mendell-Rubin test was no longer significant for more than four classes, and the Vuong-Lo-Mendell-Rubin difference test indicated that adding another class provided no significant advantage in model fit after four classes. Class separation and the homogeneity of classes were the strongest for the four class model. The low risk class (27% of sample) had only low probabilities for having had a casual sex partner and having been drunk. The moderate risk class (27% of sample) had high probabilities of having had a casual sex partner, drunkenness, and moderate probability of engagement in smoking. The intermediate risk class (41% of class) had high probabilities for smoking, drunkenness, marijuana use, simultaneous alcohol and marijuana use, and having had a casual sex partner. The extreme risk class (5% of sample) had high probabilities for all substance use and risky sexual activity items.

These results demonstrate strong correspondence between substance use and risky sexual activity and suggest several distinct patterns of engagement in risk behaviors for college students. Important implications for prevention and intervention programming for college students engaging in risky sexual activity and substance use will be discussed.