Prospective associations between alcohol and drug consumption and risky sex among female college students.

ABSTRACT

Females who had vaginal sex were interviewed annually in their first two years of college (n=386, ages 17 to 20 at study outset) and asked about risky sex behaviors and substance use. In year one, [60.2%.sub.wt] had intoxicated sex, [31.4%.sub.wt] had multiple sex partners, and [48.9%.sub.wt] had unprotected sex (i.e., without a condom). At follow-up, high rates of persistence (86.0%, 52.7%, 78.8% respectively) and initiation (36.0%, 23.9%, 41.8%) were observed. In multiple logistic regression analyses, drug use and drinking were independently associated with having multiple sex partners. Intoxicated sex independently predicted condom non-use and multiple sex partners, and appeared to mediate the relationship between substance use and multiple sex partners.

Implications for prevention and future research are discussed.

INTRODUCTION

Sexual health and substance use are two of the chief concerns facing college health professionals in the U.S., and there is considerable overlap between them. "Risky sex" is a broad term encompassing a number of behaviors that place a person at risk for unintended pregnancy, sexually transmitted infection (STI), sexual violence, or other adverse outcomes. Having sex without a condom, having numerous sex partners, having sex with a stranger, and having sex while under the influence of alcohol are a few examples of risky sex behaviors that have been studied in college students. Cross-sectional studies of college students have repeatedly demonstrated the high prevalence of risky sex. For example, the most recent national data indicate that nearly half of sexually active students did not use a condom during their most recent sexual encounter, and 13.9% of all students have had unprotected sex during the past year as a result of their drinking (American College Health Association, 2007). Empirical evidence points to a strong link between binge drinking and both unplanned sex and unprotected sex among college students (e.g., Wechsler, Lee, Kuo, & Lee, 2000). Moreover, serial patterns of casual sex, sometimes known as "hooking up," are a common part of the cultural landscape for many college students and are linked both situationally to alcohol and drug use (Grello, Welsh, & Harper, 2006) and to more severe overall patterns of alcohol intoxication (Paul, McManus, & Hayes, 2000).

A great deal of research has explored the psychological, social, and cognitive mechanisms by which alcohol use might influence risky sex. First, heavy alcohol use is related to riskier patterns of decision-making about sex partners, namely, in having more sex partners (Graves, 1995; Wechsler, Dowdall, Davenport, & Castillo, 1995), having sex that was not planned (O'Hare, 1998; Wechsler et al., 2000), and having sex with casual and new partners (Bon, Hittner, & Lawandales, 2001; Goldstein, Barnett, Pedlow, & Murphy, 2007; Graves, 1995; Testa & Collins, 1997). Risky partner choices are thought to be attributable, in part, to alcohol's acute disinhibiting effects and lowered risk perceptions with new sex partners (Fromme, D'amico, & Katz, 1999). Second, alcohol use can have a deleterious effect on safer-sex behaviors once a sexual encounter has begun. Several studies have demonstrated that alcohol intoxication appears to reduce the likelihood of using condoms and other contraceptives (Goldstein et al., 2007; Hingson, Heeren, Zakocs, Kopstein, & Wechsler, 2002; Meilman, 1993), thereby putting students who drink more heavily at greater risk for adverse outcomes such as unwanted pregnancy (Perkins, 2002) and STI (Chesson, Harrison, & Stall, 2003).

In contrast to the wealth of information available on alcohol-related risky sex, fewer studies have examined the association between illicit drug use and risky sex. One cross-sectional study of college students found an association between marijuana use and condom non-use while drunk or high, independent of alcohol use (Bon et al., 2001). Similarly, in a large national survey of high school students, illicit drug use was associated with having more sexual partners and condom non-use, independent of alcohol use (Richter, Valois, Mckeown, & Vincent, 1993).

Few prior studies of risky sex have focused on female college students; yet the health and social consequences of risky sex are uniquely burdensome for women. An estimated 2% of sexually active female college students become pregnant annually (American College Health Association, 2007). HPV infection is widespread among sexually active female college students (Winer et al., 2003) and increases their risk of cervical cancer. As the receptive partners in heterosexual encounters, females are subject to higher HIV transmissibility and therefore more likely than males to contract HIV even though they engage in risky sex less often (Anderson & Mathieu, 1998).

Although prior studies have examined college students' risky sex behaviors cross-sectionally, very little information is available about how risky sex behaviors in general, and alcohol and drug-related risky sex behaviors in particular, change over time, and the extent to which these behaviors are isolated or indicative of more chronic behavior patterns. The present study uses longitudinal data from a cohort study to fill critical gaps in our knowledge about female college students' sexual behaviors, by examining the temporal patterning of risky sex behaviors and how drug use and drinking might influence changes in risky sex over time. This study has three aims: 1) Describe the prevalence, initiation, and persistence of three risky sexual behaviors in the past six months (multiple sex partners, condom non-use, and intoxicated sex); 2) describe the association between drug use, alcohol consumption, and risky sex during the first year of college; and, 3) prospectively model the association between changes in risky sex and drug use and alcohol consumption.

METHODS

The College Life Study is a longitudinal, prospective study of health behaviors among a college student cohort. The sample was ascertained in two stages. First, during summer orientation in 2004, a brief screening survey was administered to 3,401 first-time incoming freshman students ages 17 to 19, representing 81.8% of the incoming class. From the pool of screener participants, illicit drug users were purposively oversampled for the longitudinal study. During their freshman year, students were invited by telephone and email to participate in a two-hour face-to-face assessment (of which 86.4% of contacted students completed), consisting of a personal interview and self-administered questions on sexual behavior and other highly sensitive topics. The resulting sample of 1,253 students (645 females) was re-contacted twelve months later for a second assessment (91.1% response rate). Students received $5 for completing the screener and $50 for each interview. Informed consent for participation in all phases of the longitudinal study was acquired under a University IRB-approved protocol. A federal Certificate of Confidentiality was obtained. Additional details on the sampling and recruitment methods are presented elsewhere (Arria et al., 2008).

Participants

The analysis sample for the present study consists of the 386 unmarried female college students who: a) participated in both interviews; b) had vaginal sex at least once in their lives by their first interview; and, c) provided complete data on all variables for the present analyses. Sample characteristics are presented in Table 1. Demographics were similar to the general population of first-year female students at the university (Arria et al., 2008). Females excluded from the present analyses because they never had sex (n=203) or had incomplete data (n=56) were not significantly different from the analysis sample with respect to race, age, or mother's education; however, they were slightly more likely to identify religion as "extremely important" (29% vs. 17%, p<.002), they drank less frequently (31.4 vs. 51.2 days in the past year, p<.0001), and they used fewer illicit drugs (0.7 vs. 1.3 drugs used in the past year, p<.0001).

Measures

Age and race were obtained from university administrative data. Religiosity was measured via one self-administered item in year one: "How important is religion in your life?" Responses were dichotomized as "extremely important" vs. less (i.e., "not," "slightly," or "moderately") important (Bachman et al., 2001; Jessor, Donovan, & Costa, 1991). The rationale for including religiosity as a control variable in the present study stems from prior evidence that religiosity is an important correlate of both risky sex and substance use in college students, especially for women (Templin & Martin, 1999; Poulson, Eppler, Satterwhite, Wuensch, & Bass, 1998).

In both years, questions on vaginal sex were modeled after the Youth Risk Behavior Survey (Centers for Disease Control and Prevention, 1997). To enhance participants' feeling of privacy and reduce the occurrence of underreporting for these sensitive questions (Gribble, Miller, Rogers, & Turner, 1999), self-administered forms were used, and participants were encouraged to place their completed questionnaires directly into a security envelope. Time since sexual debut was computed as the age at interview minus the age at first vaginal sex. Questions referencing the past six months asked the number of vaginal sex partners, how often they engaged in vaginal sex after alcohol or drug use, and how often they had used a condom. Number of sex partners (range: 0 to 30), was recoded into a binary variable for having multiple sex partners (2 or more vs. 0 or 1). A binary variable for intoxicated sex was coded as "no" for individuals who responded "never" to both questions on alcohol and drug use before sex, and "yes" for all responses of "rarely," "sometimes," "often," or "every time." A binary variable for condom non-use was coded as "yes" for responses of "never," "rarely," "sometimes," and "often;" individuals who used a condom "every time" or who did not have vaginal sex in the past six months were coded as "no" because they did not have unprotected sex. We included individuals with no recent vaginal sex partners (i.e., in the past six months) because refraining from vaginal sex was conceptualized as a relevant strategy for sexual risk reduction, similar to using condoms consistently and limiting one's number of partners. Because we had no information about the quantity or duration of alcohol or drug use on the occasions of "intoxicated sex," this variable was understood to encompass all possible levels of substance-induced impairment.

Participants were personally interviewed about their alcohol and drug use, using procedures similar to the National Survey on Drug Use and Health (NSDUH) (Substance Abuse and Mental Health Services Administration, 2003). Frequency of alcohol use was captured as the number of days they had any drink containing alcohol during the past year. Recency of use was recorded for marijuana, inhalants, hallucinogens, cocaine, amphetamine/methamphetamine, heroin, ecstasy, and nonmedical use of prescription stimulants, analgesics and tranquilizers. An index of overall drug involvement was derived by summing the total number of drugs used in the past year, which ranged from 0 to 10.

For both years, weighted prevalence was computed for three risky sex behaviors: multiple sex partners, condom non-use, and intoxicated sex. Because drug users were oversampled at study outset, sampling weights were applied to ensure prevalence estimates more accurately reflected the general population of female students who ever had sex in the target population. To illustrate changes in risky sex across the two years, initiation and persistence rates were computed (without weighting). For example, for multiple sex partners, initiation was computed as the number of individuals who had multiple sex partners at their year-two assessment but not at year one, divided by the total number of individuals who did not have multiple sex partners at year one. Persistence was computed as the number of individuals who had multiple sex partners at both interviews, divided by the total number who had multiple sex partners at year one. Initiation and persistence were computed similarly for condom non-use and intoxicated sex.

Statistical Analyses

The association between drug and alcohol use and risky sex was tested in two series of multiple regression models. First, to examine the cross-sectional associations in year one, multiple sex partners and condom non-use were modeled as binary dependent variables. Our analytic strategy was to test the possible association of overall substance use patterns (i.e., drinking frequency, number of illicit drugs) and acute alcohol and drug effects (intoxicated sex) with sexual decisions regarding partner choice (multiple sex partners) and condom use. Therefore, alcohol use frequency and number of drugs used were entered as the hypothesized main effects of primary interest, while age, race, religiosity, and number of years since sexual debut were held constant. Models were tested with and without intoxicated sex as an additional explanatory variable, using a model comparisons approach (MacKinnon, Lockwood, Hoffman, West, & Sheets, 2002) to test for evidence of a possible intervening effect. Second, to examine the possible relationship between drug and alcohol use and changes in risky sex over time, a series of prospective models was tested, with multiple sex partners and condom non-use as the dependent variables. For these models, a four-level polytomous dependent variable was modeled based on the temporal pattern of when that behavior was reported: refrained (behavior reported in neither year one nor year two), initiated (behavior reported in year two but not year one), desisted (behavior reported in year one but not year two), and persisted (behavior reported in both year one and year two). The same explanatory variables were entered as above. Each hypothesized effect was evaluated with respect to the risk for initiating relative to refraining, and continuing relative to desisting, the behavior.

RESULTS

Table 2 describes the sample with respect to risky sex behaviors. As can be seen, nearly one-third had multiple sex partners in the past six months, and this proportion remained stable across the two years of data collection. The overall prevalence of condom non-use was 48.9% in year one, and increased substantially by year two ([61.1%.sub.wt]). Intoxicated sex was reported by almost two. thirds of the sample each year and primarily revolved drinking before sex. In most cases intoxicated sex occurred infrequently; however, a small but growing minority said they drank alcohol often or every time they had sex ([7.8%.sub.wt] in year one, [11.7%.sub.wt] in year two).

With respect to changes in risky sex over time, all three behaviors had high rates of initiation and persistence. Among individuals who had fewer than two sex partners in year one, 23.9% had multiple sex partners in year two. The corresponding initiation rates were even higher for condom non-use (41.8%) and intoxicated sex (36.0%). With respect to persistence, individuals continuing the behavior from year one to year two were in the majority for having multiple sex partners (52.7%), condom non-use (78.8%), and intoxicated sex (86.0%).

Table 3 presents the results of the multiple regression models on multiple sex partners and condom non-use, using cross-sectional data from year one. As can be seen in Model 1, both the number of drugs used (AOR=1.5; 95% CI=1.1, 1.7; p=.01) and the frequency of drinking (AOR=1.4; 95% CI=1.1, 1.8; p<.01) were independently associated with an increased risk for having multiple sex partners. However, neither of these effects was robust to the inclusion of intoxicated sex, which was strongly predictive of having multiple sex partners (AOR=5.5; 95% CI=2.8, 10.5; p<.01) in Model 2. This combination of results suggests an intervening effect in which drug use and drinking increase the risk for having multiple sex partners primarily via intoxicated sex. Interestingly, in Models 3 and 4, neither drinking nor drug use was associated with condom non-use, although intoxicated sex was (AOR=3.4; 95% CI=2.0, 5.8; p<.01). In addition, as expected, the longer an individual had been sexually active, the more likely she was during the most recent six months to have had multiple partners (AOR=1.5; 95% CI=1.2, 1.8; p<.01) and sex without a condom (AOR=1.9; 95% CI=1.6, 2.4; p<.01).

The models predicting changes in risky sex revealed very few significant effects in either the hypothesized predictors or the control variables (results available upon request). In the model predicting initiation of having multiple sex partners in year two, intoxicated sex was a marginally significant predictor (AOR=1.8; 95% CI=0.9, 3.7; p=.09); none of the other hypothesized predictors was significant. In the comparison of continuing to have multiple sex partners relative to desisting, the number of illicit drugs used was the only significant predictor (AOR=1.6; 95% CI=1.1, 2.4; p=.01). By contrast, in the model on condom non-use none of the hypothesized effects was even marginally significant at p<.15.

Because use of alcohol or drugs before sex was collapsed into one variable for this study, it was of interest to examine the relative impact of each behavior separately (i.e., "drugged sex" as distinct from "drunk sex.") Therefore, we replicated the cross-sectional models in Table 3, substituting the intoxicated sex variable with the drugged sex variable. Each model was tested with and without drunk sex, to determine whether drugged sex might contribute any unique variance to having multiple sex partners or condom non-use. Results indicated drugged sex was not significantly associated with having multiple sex partners, regardless of whether drunk sex was held constant. Having drugged sex was associated with a two-fold increase in the risk for condom nonuse (AOR=2.0; 95% CI=1.0, 3.9; p=.04), but this effect was not robust to the inclusion of drunk sex. These results suggest that, when it occurs in sexual situations, drug use tends to have a negative effect on condom use, similar to alcohol.

DISCUSSION

Consistent with earlier studies revealing high rates of risky sex among college students (Roberts & Kennedy, 2006), this study showed that, among first-year female students who ever had vaginal sex, nearly half had sex without condoms ([48.9%.sub.wt]), one-third had multiple sex partners ([31.4%.sub.wt]), and a majority had sex under the influence of alcohol or other drugs ([60.2%.sub.wt]) during the past six months. Condom non-use became even more prevalent in year two ([61.1%.sub.wt]). This study provides strong evidence that college is a period of rapid escalation in risky sex among females as well as a time of rapid acquisition of new risky sex behaviors. This finding is particularly troubling considering that once adopted, risky behaviors usually persisted from one year to the next. To our knowledge, this is the first prospective evidence of the persistence of certain risky sex behaviors among female college students.

Females in this study who drank more frequently were at greater risk for having multiple sex partners, but not for condom non-use. Although we did not ask participants how long they had known their partners, the findings are compatible with prior evidence that frequent or heavy drinking is related to having casual sex (Bon et al., 2001; Goldstein et al., 2007; Graves, 1995; Paul et al., 2000; Testa & Collins, 1997) and multiple partners (Graves, 1995; Wechsler et al., 1995). However, as others have noted, the relationship between drinking and condom use is complex and difficult to observe (Cooper, 2002), and our finding that intoxicated sex, but not drinking frequency, was related to condom non-use comports with that view. Intoxication in a sexual context appears to be a key determinant of risky sex regardless of drinking frequency. Drinking in sexual situations might indicate underlying personality factors that convey a propensity for risky sex, such as sensation seeking (Velez-Blasini, 2008) or sexual sensation seeking (Hendershot, Stoner, George, & Norris, 2007).

The finding that illicit drug use, including nonmedical prescription drug use and drinking, are independently associated with risky sex is consistent with prior research with adolescents (Richter et al., 1993). To our knowledge, only one study reported this association in college students (Bon et al., 2001), but it did not provide any information about drugs other than alcohol and marijuana. The present study observed that drug use confers an additional risk for having multiple sex partners, regardless of drinking frequency, particularly when drug use occurs in sexual situations. The association was also observed prospectively: among women who had multiple sex partners, only drug use was significantly predictive of continuing to have multiple sex partners. Although these prospective findings were not always statistically significant, the general picture that emerges is that female college students who use drugs and alcohol in sexual situations may be at greater risk for initiating other risky sexual behaviors, and continuing the risky sex behaviors they have already adopted.

A noteworthy limitation of this study is that it was restricted to questions about heterosexual vaginal intercourse. Because oral sex is prevalent among college students (American College Health Association, 2007; Goldstein et al., 2007), it is possible that some individuals not having unprotected vaginal sex were having unprotected oral sex and were at risk for contracting STIs orally. Limited evidence suggests that the factors affecting sexual decision-making are similar for both oral and vaginal sex (Goldstein et al., 2007). Moreover, recent evidence suggests that casual non-coital sexual encounters constitute a significant proportion of college students' sexual activity and should be considered within a continuum of risky sex behaviors (Paul et al., 2000; Velez-Blasini, 2008).

Several other limitations of this study also warrant mention. First, although we did distinguish between alcohol and other drug use in sexual situations, we cannot determine which drugs were used at those times. Also, because we did not capture the specific quantity or duration of substance use in sexual situations, we acknowledge the possibility that some sexual encounters in which drinking or drug use was minimal might have been erroneously coded as "intoxicated sex." This concern is mitigated, however, by the fact that in most cases (90%), participants coded as positive for intoxicated sex also characterized their typical alcohol consumption at a level that could reasonably be expected to produce some level of impairment, i.e., at least three drinks per drinking day. Data not shown in a table. Also, we have no information on current or past STIs, which would be a useful indicator of sexual risk. Our models did not account for several factors relevant for understanding sexual risk such as the frequency of sex, relationship status (i.e., casual vs. steady), other contraceptive methods, and attitudes and perceptions regarding susceptibility and safer sex practices, or for underlying trait variables such as sensation-seeking that could account for both substance use and risky sex. Moreover, we cannot estimate how many students may have underreported their risky sex behaviors; however, in collecting self-report data on sexual behavior, self-administered questionnaires are regarded as advantageous over interviewer administration with respect to social desirability bias (Gribble et al., 1999).

Notwithstanding these limitations, this study fills an important gap in our knowledge about how young women's sexual risk behaviors change during college. One prior cross-sectional study of the sexual behaviors of freshman, sophomore,junior, and senior year college students (Siegel, Klein, & Roghmann, 1999) found that seniors had increased levels of partner trust and oral contraceptive use relative to freshmen, but condom use remained stable across all four groups. Our prospective study will be able to shed additional light on the temporal patterning of risky sex, substance use, and their correlates throughout the entire college career.

This study provides evidence that illicit drug use may influence condom use independent of alcohol, but the present data illustrate only a global association based on individuals' overall behavior patterns, and therefore cannot contribute information about possible causal effects of drug and alcohol use within a given sexual event. Prior studies focusing on alcohol use in sexual contexts have demonstrated the usefulness of event-level data for explaining the cognitive and situational factors involved in sexual decision-making (Anderson & Mathieu, 1996; Dermen & Cooper, 2000; Goldstein et al., 2007). Future event-level studies should include questions about use of drugs other than alcohol.

The present findings draw attention to the need for safer-sex educational programs to continue throughout the college years. Many women in this sample had relatively little sexual experience. Based on their ages at interview and ages at sexual debut, 46.7% had been sexually active for fewer than two years. In light of their relative inexperience, many female college students might benefit from practical skill-building programs aimed at sexual risk reduction, such as rehearsing ways to talk about sex and condoms with their partners. Furthermore, in this study risky behaviors were highly persistent over time; therefore, educating and empowering college women about their ability to change past risky behaviors may be another important prevention strategy.

Safer-sex programs emphasizing responsible use of alcohol (Clapper & Lipsitt, 1991; Goldstein et al., 2007; Smith & Brown, 1998) should simultaneously address drug use in the context of sexual risk reduction. In addition to encouraging women to limit their overall involvement in substance use, prevention initiatives should specifically encourage women to avoid using alcohol and other drugs in potentially sexual situations. Other promising strategies include increasing the visibility of cues promoting condom use in locations where substance use and sexual situations are likely to occur (Kaly, Heesacker, & Frost, 2002) such as in campus housing and bars catering to a college-student clientele, and skill-building interventions on how to make appropriate risk appraisals in sexual situations while intoxicated (Abbey, Saenz, Buck, Parkhill, & Hayman, 2006). Students engaging in risky sex might benefit from prevention programs that acknowledge the personal goals and priorities that motivate them to have sex rather than focusing solely on the reasons why they should refrain (Patrick, Maggs, & Abar, 2007). Given the strong interrelationships between substance use and risky sex observed in this population, and evidence that sexual motives often influence substance use behaviors (i.e., drinking to facilitate a sexual encounter; Anderson & Mathieu, 1998), it makes sense to address both sexual and substance use motives simultaneously.

Campus health professionals have a unique opportunity to positively influence the sexual health of female college students during a critical period in their sexual development. National data indicate that health center staff and health educators enjoy high credibility amongst college students, yet only a minority of students receives information from their university regarding pregnancy prevention (22.5%), HIV (28.0%), and other STIs (35.8%) (American College Health Association, 2007). Resources should be directed toward strengthening outreach programs and funding health initiatives that address substance use and sexual health simultaneously, such as including sexual risk as a component of interventions for students who are referred for substance use problems. College students are a highly educated population with a high level of sexual risk (Rothenberger & Buck, 1998), and therefore present health professionals with an important opportunity to make a substantial impact on public health.

Correspondence concerning this article should be addressed to: Amelia M. Arria, Associate Director, Center for Substance Abuse Research (CESAR), University of Maryland, 4321 Hartwick Rd, Suite 501, College Park, MD 20740; Phone: [phone]; Fax: [phone]; [e-mail]

AUTHOR NOTE

No conflicts of interest exist. The investigators would like to acknowledge funding from the National Institute on Drug Abuse (R01DA14845). Special thanks are given to Sarah Kasperski, Laura Gamier, Gillian Pinchevsky, the interviewing team, and the participants. Results from this paper were presented at the 2007 American Public Health Association Annual Meeting.

REFERENCES

Abbey, A., Saenz, C., Buck, P. O., Parkhill, M. R. & Hayman, L. W., Jr. (2006). The effects of acute alcohol consumption, cognitive reserve, partner risk, and gender on sexual decision making. Journal of Studies on Alcohol, 67, 113-121.

American College Health Association. (2007). National College Health Assessment: Spring 2006 reference group data report (abridged). Journal of American College Health, 55, 195-206.

Anderson, P. B. & Mathieu, D. A. (1996). College students' high-risk sexual behavior following alcohol consumption. Journal of Sex & Marital Therapy, 22, 259-264.

Anderson, P. B. & Mathieu, D. A. (1998). Drinking and safer sex: Are college students at risk? Journal of Sex Education & Therapy, 23, 297-301.

Arria, A. M., Caldeira, K. M., O'Grady, K. E., Vincent, K. B., Fitzelle, D. B., Johnson, E. P. & Wish, E. D. (2008). Drug exposure opportunities and use patterns among college students: Results of a longitudinal prospective cohort study. Substance Abuse, 29, 19-38.

Bachman, J. G., O'Malley, P. M., Schulenberg, J., Johnston, L. D., Bryant, A. L., Merline, A., Freedman-Doan, P., Ridenour, N. J. & Hart, T. C. (2001). Analyses showing how religiosity, social activities, and drug-related beliefs mediate relationships between post-high school experiences and substance use, Institute for Social Research.

Bon, S. R., Hittner, J. B. & Lawandales, J. P. (2001). Normative perceptions in relation to substance use and HIV-risky sexual behaviors of college students. The Journal of Psychology, 135, 165-178.

Centers for Disease Control and Prevention. (1997). CDC surveillance summaries, youth risk behavior surveillance: National College Health Risk Behavior Survey, United States, 1995. Morbidity and Mortality Weekly Report, 46.

Chesson, H. W., Harrison, P. & Stall, R. (2003). Changes in alcohol consumption and in sexually transmitted disease incidence rates in the United States: 1983-1998. Journal of Studies on Alcohol, 64, 623-630.

Clapper, R. L. & Lipsitt, L. P. (1991). A retrospective study of risk-taking and alcohol-mediated unprotected intercourse. Journal of Substance Abuse, 3, 91-96.

Cooper, M. L. (2002). Alcohol use and risky sexual behavior among college students and youth: Evaluating the evidence. Journal of Studies on Alcohol Supplement, 63, 101-117.

Dermen, K. H. & Cooper, M. L. (2000). Inhibition conflict and alcohol expectancy as moderators of alcohol's relationship to condom use. Experimental and Clinical Psychopharmacology, 8, 198-206.

Fromme, K., D'amico, E. J. & Katz, E. C. (1999). Intoxicated sexual risk taking: an expectancy or cognitive impairment explanation? Journal of Studies on Alcohol, 60, 54-63.

Goldstein, A. L., Barnett, N. P., Pedlow, C. T. & Murphy, J. G. (2007). Drinking in conjunction with sexual experiences among at-risk college student drinkers. Journal of Studies on Alcohol and Drugs, 68, 697-705.

Graves, K. L. (1995). Risky sexual behavior and alcohol use among young adults: Results from a national survey. American Journal of Health Promotion, 10, 27-36.

Grello, C. M., Welsh, D. P., & Harper, M. S. (2006). No strings attached: The nature of casual sex in college students. Journal of Sex Research, 43(3), 255-267.

Gribble, J. N., Miller, H. G., Rogers, S. M., & Turner, C. F. (1999). Interview mode and measurement of sexual behaviors: Methodological issues. Journal of Sex Research, 36(1), 16-24.

Hendershot, C. S., Stoner, S. A., George, W. H. & Norris, J. (2007). Alcohol use, expectancies, and sexual sensation seeking as correlates of HIV risk behavior in heterosexual young adults. Psychology of Addictive Behaviors, 21, 365-372.

Hingson, R. W., Heeren, T., Zakocs, R. C., Kopstein, A. & Wechsler, H. (2002). Magnitude of alcohol-related mortality and morbidity among U.S. college students ages 18-24. Journal of Studies on Alcohol, 63, 136-144.

Jessor, R., Donovan, J. E. & Costa, F. M. (1991). Beyond adolescence: Problem behavior and young adult development, Cambridge: Cambridge University Press.

Kaly, P. W., Heesacker, M. & Frost, H. M. (2002). Collegiate alcohol use and high-risk sexual behavior: A literature review. Journal of College Student Development, 43, 838-850.

MacKinnon, D. P., Lockwood, C. M., Hoffman, J. M., West, S. G. & Sheets, V. L. (2002). A comparison of methods to test mediation and other intervening variable effects. Psychological Methods, 7, 83-104.

Meilman, P. W. (1993). Alcohol-induced sexual behavior on campus. Journal of American College Health, 42, 27-31.

O'Hare, T. (1998). Drinking and risky sexual behavior in young women and men: A covalidation study. Journal of Alcohol and Drug Education, 43, 66-77.

Patrick, M. E., Maggs, J. L., & Abar, C. C. (2007). Reasons to have sex, personal goals, and sexual behavior during the transition to college. Journal of Sex Research, 44(3), 240-249.

Paul, E. L., McManus, B., & Hayes, A. (2000). "Hookups": Characteristics and correlates of college students' spontaneous and anonymous sexual experiences. Journal of Sex Research, 37(1), 76-88.

Perkins, H. W. (2002). Surveying the damage: A review of research on consequences of alcohol misuse in college populations. Journal of Studies on Alcohol Supplement, 63, 91-100.

Poulson, R. L., Eppler, M. A., Satterwhite, T. N., Wuensch, K. L. & Bass, L. A. (1998). Alcohol consumption, strength of religious beliefs, and risky sexual behavior in college students. Journal of American College Health, 46, 227-232.

Richter, D. L., Valois, R. F., Mckeown, R. E. & Vincent, M. L. (1993). Correlates of condom use and number of sexual partners among high school adolescents. The Journal of School Health, 63, 91-96.

Roberts, S. T. & Kennedy, B. L. (2006). Why are young college women not using condoms? Their perceived risk, drug use, and developmental vulnerability may provide important clues to sexual risk. Archives of Psychiatric Nursing, 20, 32-40.

Rothenberger, J. H. & Buck, H. (1998). College health needs to participate in the national STD debate. Journal of American College Health, 47, 140-142.

Siegel, D. M., Klein, D. I. & Roghmann, K. J. (1999). Sexual behavior, contraception, and risk among college students. Journal of Adolescent Health, 25, 336-343.

Smith, C. D. & Brown, J. M. (1998). Sexual behaviors, extroversion, and alcohol use among college students. Journal of Alcohol and Drug Education, 44, 70-79.

Substance Abuse and Mental Health Services Administration. (2003). 2002 National Survey on Drug Use and Health Questionnaire, Rockville, MD.

Templin, D. P. & Martin, M. J. (1999). The relationship between religious orientation, gender, and drinking patterns among Catholic college students. College Student Journal, 33, 488-495.

Testa, M. & Collins, R. L. (1997). Alcohol and risky sexual behavior: Event-based analyses among a sample of high-risk women. Psychology of Addictive Behaviors, 11, 190-201.

Velez-Blasini, C. J. (2008). Evidence against alcohol as a proximal cause of sexual risk taking among college students. Journal of Sex Research, 45(2), 118-128.

Wechsler, H., Dowdall, G. W., Davenport, A. & Castillo, S. (1995). Correlates of college student binge drinking. American Journal of Public Health, 85, 921-926.

Wechsler, H., Lee, J. E., Kuo, M. & Lee, H. (2000). College binge drinking in the 1990s--A continuing problem: Results of the Harvard School of Public Health 1999 College Alcohol Study. Journal of American College Health, 48, 199-210.

Winer, R. L., Lee, S.-K., Hughes, J. P., Adam, D. E., Kiviat, N. B. & Koutsky, L. A. (2003). Genital human papillomavirus infection: Incidence and risk factors in a cohort of female university students. American Journal of Epidemiology, 157, 218-226.

Kimberly M. Caldeira, Amelia M. Arria, Elizabeth M. Zarate, Kathryn B. Vincent & Eric D. Wish Center of Substance Abuse Research (CESAR) University of Maryland

Kevin E. O'Grady University of Maryland

TABLE 1
Sample Characteristics for N=386 Female College Students Who
Ever Had Vaginal Sex

                                                n (%)     Mean (SD)
Race
  White                                    272 (70.5)
  Black or African American                 45 (11.7)
  Asian                                     26 ( 6.7)
  Other                                     43 (11.1)
Age at interview                                         18.2 (0.5)
Mother's education
  High school, GED or less                  72 (20.2)
  Technical or associate's degree           38 (10.6)
  Bachelor's degree                        138 (38.7)
  Graduate or professional degree          109 (30.5)
How important is religion in your life?
Not important                               87 (22.5)
Slightly important                         106 (27.5)
Moderately important                       127 (32.9)
Extremely important                         66 (17.1)
Number of years since first vaginal sex                   1.8 (1.2)
0                                           50 (13.0)
1                                          130 (33.7)
2                                          103 (26.7)
3                                           73 (18.9)
4 or more                                    30 (7.8)
Number of illicit drugs used, past year                   1.3 (1.5)
Used 1 or more illicit drugs               269 (69.7)
Number of drinking days, past year                       51.2 (51.9)

TABLE 2
Prevalence, Initiation, and Persistence of Risky Sexual Behaviors
during the Past Six Months, as Measured in Years One anal Two of
College, among N = 386 Women Ever Had Vaginal Sex

                                       Year One         Year Two
                                       Weighted         Weighted
Sexual Behaviors                    Prevalence (a)   Prevalence (a)
During the Past Six Months            [%.sub.wt]       [%.sub.wt]

Number of sex partners
  0                                            7.6             10.7
  1                                           61.0             57.0
  2                                           18.9             18.7
  3                                            5.6              8.0
  4 or more                                    6.9              5.7
Multiple sex partners                         31.4             32.4
(had 2 or more in the
past 6 months)

Condom use (d)
Never                                          8.3             12.2
Rarely                                         7.1              8.8
Sometimes                                     12.0             13.9
Often                                         21.7             26.1
Every time                                    43.5             28.1
Abstained from vaginal sex
during the past six months                    48.9             61.1
Condom non-use
did not use condoms eve time

Used drugs before having sex
  Never                                       80.2             71.2
  Rarely                                       8.3             10.2
  Sometimes                                    2.7              6.7
  Often                                        0.9              1.3
  Every time                                   0.4              0.0
  Abstained from vaginal sex
  during the past six months                   7.6             10.7
Drank alcohol before having sex
  Never                                       32.4             25.7
  Rarely                                      30.7             32.0
  Sometimes                                   21.6             19.5
  Often                                        4.0              7.3
  Every time                                   3.8              4.0
  Abstained from vaginal sex
  during the cast six months                   7.6             10.7
Intoxicated Sex
(used drugs or
alcohol before sex)                           60.2             63.0

                                      Initiation       Initiation
Sexual Behaviors                         (b)              (c)
During the Past Six Months                %                %

Number of sex partners
  0
  1
  2
  3
  4 or more
Multiple sex partners                         23.9
(had 2 or more in the                                          52.7
past 6 months)

Condom use (d)
Never
Rarely
Sometimes
Often
Every time
Abstained from vaginal sex
during the past six months                    41.8             78.8
Condom non-use
did not use condoms eve time

Used drugs before having sex
  Never
  Rarely
  Sometimes
  Often
  Every time
  Abstained from vaginal sex
  during the past six months
Drank alcohol before having sex
  Never
  Rarely
  Sometimes
  Often
  Every time
  Abstained from vaginal sex
  during the cast six months
Intoxicated Sex
(used drugs or
alcohol before sex)                           36.0             86.0

Results are shown in bold for variables that were dichotomized based
on the original response data.

(a) Prevalence refers to the past six months for each behavior, and is
based on a weighted sample size of 870 students.

(b) Initiation is computed as the proportion of non-cases in year one
who became cases in year two.

(c) Persistence is computed as the proportion of cases in year one who
continued to be cases in year two.

(d) Data on condom use were missing for 6 cases in year two;
frequencies were computed based on the 380 cases with non-missing
data.

TABLE 3
Results of Logistic Regression for Having Multiple Sex Partners and
Condom Non-Use, among N 386 Female First-Year-College Students Who
Ever Had Vaginal Sex

                                       Model 1:
                                  Multiple Sex Partners

                                   AOR (95% CI)        p

Age at interview               0.5 (0.3 to 0.8)     <.01
Race = White                   1.1 (0.6 to 1.8)       .82
Religiosity =
Extremely Important            1.0 (1.0 to 1.8)       .90
Number of years since
first vaginal sex              1.4 (1.2 to 1.7)     <.01
Number of days drank
alcohol in the past yeas       1.4 (1.1 to 1.8)     <.01
Number of drugs used
in the past yeas               1.5 (1.1 to 1.7)       .01
Had sex after drinking
or using drugs                               --
Model NaeelkerkeR2                          0.15

                                       Model 2:
                                  Multiple Sex Partners

                                   AOR (95% CI)        p

Age at interview               0.5 (0.3 to 0.8)     <.01
Race = White                   0.9 (0.5 to 1.5)       .59
Religiosity =
Extremely Important            1.3 (0.7 to 2.6)       .39
Number of years since
first vaginal sex              1.5 (1.2 to 1.8)     <.01
Number of days drank
alcohol in the past yeas       1.1 (0.9 to 1.5)       .36
Number of drugs used
in the past yeas               l.3 (1.0 to 1.6)       .07
Had sex after drinking
or using drugs                5.5 (2.8 to 10.5)     <.01
Model NaeelkerkeR2                           .25

                                         Model 3:
                                      Condom non-use

                                   AOR (95% CI)        p

Age at interview               0.7 (0.5 to 1.1)       .15
Race = White                   0.6 (0.3 to 0.9)       .02
Religiosity =
Extremely Important            0.5 (0.3 to 1.0)       .04
Number of years since
first vaginal sex              1.9 (1.5 to 2.3)     <.01
Number of days drank
alcohol in the past yeas       1.2 (0.9 to 1.6)       .11
Number of drugs used
in the past yeas               1.1 (0.8 to 1.4)       .56
Had sex after drinking
or using drugs                               --
Model NaeelkerkeR2                           .18

                                         Model 4:
                                        Condom non-use

                                   AOR (95% CI)        p

Age at interview               0.7 (0.5 to 1.2)       .20
Race = White                   0.5 (0.3 to 0.8)     <.01
Religiosity =
Extremely Important            0.6 (0.3 to 1.2)       .15
Number of years since
first vaginal sex              1.9 (1.6 to 2.4)    <0.01
Number of days drank
alcohol in the past yeas       1.0 (0.8 to 1.3)       .92
Number of drugs used
in the past yeas               1.0 (0.8 to 1.3)       .94
Had sex after drinking
or using drugs                 3.4 (2.0 to 5.8)     <.0l
Model NaeelkerkeR2                           .24

Results are shown in bold for variables that were statistically
significant at p<.05.

(a) The unit of analysis for computation of AORs was based on one
standard deviation, which was 51.9 for number of drinking days and 1.5
for number of illicit drugs. For all other variables, the unit of
analysis was 1.0.