Introduction

Sexual assault on college campuses remains endemic throughout the USA; multiple surveys have established that 20–25% of college women are sexually assaulted during their time of enrollment [1••, 2]. College men also report perpetrating sexual assaults in high numbers, with 10–15% reporting “the use of any tactic to make someone engage in sexual activity when that individual is unwilling or unable to consent” annually [1••]. Negative outcomes of sexual assault that can have life-long impacts are numerous: psychological consequences (e.g., PTSD, depression, anxiety), physical consequences (e.g., STIs, unwanted pregnancy, genital and non-genital injuries), and behavioral consequences (e.g., drug and alcohol abuse, suicide) [3]. For students in college, these impacts can include an inability to complete their college education, drastically affecting their earnings and career trajectories [4,5,6]. Title IX, passed in 1972, mandates a prompt response to sexual assault and harassment, as well as other forms of sex discrimination, within federally funded educational institutions; it also requires the implementation of programs geared toward the prevention of sexual assault. The Clery Act of 1990 requires that campus crime statistics, including those of sexual assault, be made publicly available [7]. While the body of evidence on the impact of these policies on sexual assault outcomes is limited, there is evidence that some sexual assault interventions have been successful at reducing sexual assault victimization and perpetration, and at increasing bystander behavior to intervene to prevent sexual assault [8,9,10]. It is estimated that one-half of all sexual assaults (in college and other settings) involved alcohol consumption by one or both parties [1••, 2]. In an analysis of sexual assaults reported by the general population on the National Violence Against Women Survey, the odds of completed rape were nearly twice as high, and the odds of injury were nine times higher, for alcohol-using offenders compared to those offenders not using alcohol [11].

Alcohol use and abuse is common on college campuses, with some studies finding that 40% of student populations participate in “heavy episodic drinking,” with a strong relationship between alcohol use and sexual assault [12]. Therefore, studies of college sexual assault prevention have recommended that sexual assault interventions specifically address the role of alcohol in sexual assault prevention [13, 14]. There is some evidence of a direct correlation between colleges’ alcohol policies and their sexual assault rates: in a study of 524 college campuses, Stotzer and MacCartney [15] found that those with weaker alcohol policies had higher reports of sexual assaults. Although some college-based sexual assault programs educate about alcohol as a risk factor for sexual assault, specific alcohol reduction interventions may be needed to actually reduce the incidence of alcohol-related sexual assault.

The purpose of this review is to examine the literature on interventions that address the reduction of alcohol use on college campuses and measure subsequent sexual assault-related outcomes. Two recent reviews have asked similar questions. Lippy and DeGue [16] analyzed available literature on the impact of population-based alcohol policies, such as those regulating pricing and outlet density, on sexual assault, mediators of sexual assault, and other potentially related behavioral outcomes. Tait and Lenton [17] completed a literature review examining the impact of web-based alcohol interventions on sexual and intimate partner violence in both college and non-college settings [17]. While we review two studies that were also reviewed in the Tait and Lenton paper [17], we have expanded our review to include non-web-based interventions and focused our review to only include college settings, and only include sexual assault-related outcomes. Our study serves to answer the following question: “What does this evidence say about whether and to what extent alcohol use interventions on college campuses are associated with reductions in sexual assault?”

Methods

Search Strategy

A review of existing literature was conducted using the following inclusion criteria:

  1. 1.

    A controlled intervention of alcohol use related attitudes and behaviors. Because of the limited number of potential studies, we did not restrict our search to randomized controlled trials.

  2. 2.

    Conducted with English-speaking undergraduate students at colleges and universities in the USA.

  3. 3.

    Collected outcomes on factors related to sexual assault victimization, perpetration, or bystander behaviors.

  4. 4.

    Published in a peer-reviewed journal from January 1, 2005, to September 15, 2019.

Searches were conducted in three major public health databases: Scopus, PubMed, and PsycInfo. The search strategy included indexed terms and keywords that limited our findings to universities and colleges, alcohol drinking, health education, prevention, or intervention, and sexual assault.

We identified 281 articles. Ninety-two duplicates were detected using Covidence software, and 157 non-interventions were removed during title/abstract screening. An additional 15 articles were removed because they did not evaluate an intervention conducted in the USA, and one was excluded because it did not target college students.

A full-text review was conducted for the remaining 16 articles. Upon closer review, only five met our inclusion criteria and were selected for extraction. Three additional articles were identified for abstract review through the reference list of Tait and Lenton’s [17] literature review; two of which met our inclusion criteria. The PRISMA chart depicting the screening and review process can be seen in Fig. 1; a detailed search strategy can be found in the Appendix.

Fig. 1
figure 1

PRISMA flow diagram

Full size image

Data Extraction

A total of seven articles were selected for extraction. All articles studied the impact of an alcohol use intervention on sexual assault outcomes, though there was variation in content and delivery of the interventions. A coding sheet was developed by our study team. Each article was coded separately by two members of the team; the coding sheets were compared, a consensus was reached on any differences, and a single coding sheet was produced for each article.

Results

Search Results

The search yielded a total of seven papers that met our criteria. One of the papers [22••] was a secondary analysis of a randomized controlled trial described in another of the included publications [21]. All papers were published between 2010 and 2019. Table 1 describes the interventions, study design and measures, hypotheses, and outcomes. Details on the effect sizes of these results can be found in Table 2.

Table 1 Aims, interventions, and measures for studies of alcohol use interventions on college campuses with sexual assault outcomes, 2010–2019
Full size table
Table 2 Effect sizes for studies of alcohol use interventions on college campuses with sexual assault outcomes, 2010–2019
Full size table

Intervention Characteristics

Interventions for three studies included information on alcohol use only (i.e., there was no sexual assault-related content). AlcoholEDU for College, a commonly used web-based alcohol education tool, was the focus of two papers [19, 20]. In both papers, AlcoholEDU was administered to incoming freshmen. This self-guided web-based intervention leads students through information about blood alcohol content and the dangerous effects of alcohol, as well as about local alcohol-related laws and policies. The remaining modules guide students through goal setting and harm reduction strategies.

In the intervention studied by Clinton-Sherrod et al. [18], all participants completed an online self-report of their own alcohol use and attitudes. The participants were then randomized to one of three intervention groups, or to a control group with no intervention. The three intervention groups received some combination of in-person motivational interviewing and/or personalized feedback; one group received only personalized feedback on their own drinking behaviors, one group received only motivational interviewing regarding alcohol use, and one group received both interventions.

The remaining four studies incorporated sexual assault-related content into at least one intervention group. In the Gilmore et al. papers [21, 22], there were three intervention groups: an alcohol-only intervention with modules similar to those in AlcoholEDU, a web-based sexual assault education and risk reduction intervention, and a combined intervention. There was an assessment-only control group as well.

The Testa et al. paper [24] focused on a parent-based intervention targeted at incoming female college freshmen. Two intervention groups of mothers received a handbook focused on college drinking and effective mother-daughter communication strategies, with one group receiving an “enhanced” handbook which also included information on sexual partner selection and sex refusal assertiveness. Mothers were encouraged to review and discuss the handbooks with their daughters.

Only one study was targeted specifically at men to reduce sexual assault perpetration by enhancing bystander intervention behaviors [23••]. There was no control group, as this was a pilot study. The intervention consisted of three in-person sessions: a 1:1 session including motivational interviewing and personalized feedback with regard to drinking behaviors, sexual activity, and bystander behaviors; a group session for all participants, focusing on masculinity, empathy, and bystander intervention skill-building; and a group booster session.

Study Design and Measures of Sexual Assault

Four of the six studies discussed in the seven papers were randomized controlled trials (RCTs) [18, 19, 21, 22••, 24]. Paschall et al. [19] employed a multi-site randomized design, in which 15 campuses implemented AlcoholEDU and 15 campuses matched on similar characteristics did not; the campus was the unit of analysis. In the remaining three RCTs [18, 21, 22••, 24], participants were randomized at the individual level, and individuals were the unit of analysis. Orchowski et al. [23••] described a pilot study, which did not include a comparison group, and used a pre-/post-test design. Wyatt et al. [20] used a time-series ARIMA analysis to explore changes in reported alcohol use and alcohol-related behaviors (including being taken advantage of sexually), in the years prior to and during the implementation of the AlcoholEDU on a university campus.

Samples and sample sizes varied across studies. All but two of the studies included only women as participants in the interventions, and three of the women-only studies only included women with a history of heavy episodic drinking [18, 21, 22••]. Only one study included only men, and all men were those with a history of heavy drinking [23••]. Nearly all studies had sample sizes of at least 200 participants. The outliers include one study that had a sample size of 25 [23••], and another with a sample size of 2400 [19].

Every study included pre-/post-test measures of alcohol use behaviors and sexual assault victimization or perpetration; more details on the alcohol measures can be found in Table 2. Both of the interventions centered around AlcoholEDU [19, 20] used “being taken advantage of sexually” as a measure of sexual assault victimization. Clinton-Sherrod’s [18] study of an alcohol-only intervention determined sexual assault victimization with 4 yes/no questions asking about experiences of pressured, attempted, forced, and incapacitated sexual assault. The interventions studied by Gilmore [21, 22••] measured the occurrence of alcohol-related sexual assault and severity of sexual assault using the Sexual Experiences Survey (SES) [25]. The parent-based intervention studied by Testa et al. [24] measured sexual victimization between follow-up periods using a revised version of the SES updated by the author. The Orchowski study [23••], undertaken with men as participants, included self-reported sexual aggression using the SES Short Form for Perpetration [25], as well as self-reported intent to use bystander intervention strategies to assist friends and strangers in avoiding sexual assault using two behavioral intention scales developed by Banyard et al. [26].

Findings

The fact that not all hypotheses were supported does not necessarily indicate that the intervention was not effective. Rather, it reflects the various types of hypotheses presented, including hypotheses regarding mediating factors, and the impact of interventions on subsets of participants. For example, the hypotheses regarding the impact of mediating factors explored by Clinton-Sherrod et al. [18] were mostly not supported; even so, among women with a history of sexual victimization, the motivational interviewing with personalized feedback intervention was associated with a significant decrease in rates of unwanted sexual activity compared to the control group (β = − .261; p < .01). Reduction in alcohol use ambivalence was a predictor of decreased sexual victimization, but it was not associated with any specific intervention. Gilmore et al. [22••] hypothesized that reductions in sexual assault would be indirectly associated with reductions in heavy drinking, drinking-related incapacitation, and alcohol-induced blackouts, although this hypothesis was not supported.

However, all five interventions reviewed in these seven studies led to a decline in sexual assault outcomes in at least one paper; AlcoholEDU led to a decrease in one paper [19], but not another [20]. Although Paschall et al.’s [19] study of the AlcoholEDU intervention found a significant decrease in the event rate ratio of victimization in colleges that received the intervention as compared to controls (ERR = .38; p < .05), this significance dissipated by the second follow-up period.

In addition, two studies found that outcomes differed depending on the prior sexual assault or perpetration history of the participants. Women with prior sexual assault histories showed steeper declines in sexual assault victimization relative to women without such histories as a result of motivational interviewing with personalized feedback (β = − .144; p < .05) [18] and combined sexual assault and alcohol risk reduction web-based interventions (β = − 0.206; p < .05) [21]. While the intervention in the Orchowski et al. [23••] study led to a reduction in sexual coercion perpetration, 25% of their small sample of men (N = 5) did perpetrate some form of sexual aggression during the follow-up period, and all but one of them had reported a history of sexual aggression at baseline.

Conclusion

This study explored the evidence addressing whether and to what extent alcohol use interventions on college campuses are associated with reductions in sexual assault. The first noteworthy finding is that despite the widespread implementation of alcohol use interventions on college campuses, we could only locate seven peer-reviewed papers, discussing five interventions that have explored the relationship between such interventions and reductions in sexual assault. Given that alcohol use has been unequivocally identified as a risk factor for sexual assaults in the general population and among college students [1••, 2], this lack of information to guide college administrators is indeed troubling.

Despite the limited number of available studies, our review suggests that there is reason to be optimistic about the potential for alcohol use interventions to reduce sexual assault on college campuses. As mentioned earlier, five of six interventions reviewed in these seven studies led to a decline in sexual assault outcomes. Tait and Lenton’s [17] prior review focused on the impact of web-based interventions, which have been adopted by higher education institutions across the country, and concluded that there was insufficient evidence, in part due to lack of research and under-powered existing studies, to determine that web-based alcohol-related interventions would lead to a decline in sexual assault outcomes on college campuses. This review ultimately recommended that web-based alcohol interventions be tailored to include specific information on sexual assault prevention. This was done in the Gilmore et al. study [21, 22••], published after the Tait and Lenton review, which found that sexual assault risk reduction interventions were in fact associated with a significant reduction in alcohol-related sexual assault compared to the control group [22••]. Moreover, our review demonstrates the potential and feasibility of alternative interventions that would have fallen outside the scope of the Tait and Lenton review [17], such as the mother-daughter communication intervention [24] and the motivational interviewing intervention [18]. It is noteworthy that Gilmore et al. [22••] found that adding sexual assault content to their alcohol use intervention not only reduced sexual assault outcomes but independently reduced alcohol use as well.

Taken together, we find compelling evidence that alcohol use interventions can be an effective strategy for reducing sexual assault outcomes on college campuses. We are not suggesting that such interventions replace existing sexual assault prevention initiatives, quite the contrary. We are recommending augmenting alcohol use interventions, by ensuring such interventions directly address the link between alcohol and sexual assault. Doing so would likely reduce the number of sexual assaults on campuses and may further reduce alcohol use as well.

Despite these encouraging findings, it is clear that very little is known about the pathways with which these programs are having an impact on sexual assault. Many of the hypotheses about mediators and pathways were not supported. For example, although Clinton Sherrod et al. [18] found that an alcohol use intervention was related to reduced alcohol use and reduced sexual assault outcomes, reductions in alcohol use were not related to reductions in sexual assault outcomes. One intervention that studied the effects of mothers talking to their daughters (after receiving a booklet to guide their discussion) was promising, yet due to the nature of the study, we do not know the content and quality of the discussion between the mother and daughter, in order to understand what components of the intervention were the key to its success [24]. Without further understanding of the essential components of these programs, we will be hampered in our ability to design the most effective program for wider distribution.

It is also noteworthy that some of the studies only included students who were heavy drinkers [18, 21,22,23], and thus findings related to such interventions cannot be applied to general college populations. For overall sexual assault outcomes (not restricted to alcohol-related sexual assaults), programs seem to have the most effect on women with histories of prior sexual assault. Four of the studies enrolled only women who reported some heavy drinking, and the single study focused on men similarly enrolled men who were heavy drinkers. The two studies that evaluated AlcoholEDU along with other intervention components targeting entire campuses had mixed results regarding sexual assault outcomes. We also know little about implementation features that effect program success.

There are several important limitations that should be considered when evaluating the results of the studies we reviewed. Although randomization was a strong feature in most of the studies, generally the follow-up periods were quite short (e.g., 2–3 months), and loss to follow-up was a limitation in some. The one study that included an 18-month follow-up showed diminishing effects, underscoring the need to follow participants over time, and suggests the need for including (and evaluating) booster sessions over time. None of the studies examined process measures that would help to understand implementation issues. There was considerable variety in measures that were used, and the field would benefit from some agreement on an overarching conceptual framework and standardized variables used for outcomes, mediators, and moderators.

We would be remiss if we did not address an important consideration in our recommendation to expand the scope of alcohol use intervention goals to include the reduction of sexual assault. We are not advocating an approach to sexual assault prevention in which perpetrators are excused for their behavior due to alcohol use. We are not suggesting that reducing alcohol use will have any effect on the root causes of sexual assault—including toxic masculinity, sexism, and unhealthy relationship norms. We are similarly not suggesting that a victim is at fault for being victimized if he or she has consumed alcohol. The onus of the responsibility for sexual assault is always on the perpetrator, and efforts to eliminate sexual assault need to focus on interrupting perpetrator behaviors. However, given the well-established link between alcohol use and both victimization and perpetration of sexual assault, if we wish to reduce sexual assault, reducing alcohol use is an appropriate and important strategy that appears to hold promise.

As a note on gender and sexuality, our search terms were designed to be inclusive of all interventions within the scope of this question, without making any assumptions as to the gender of the victim or perpetrator. Perhaps unsurprisingly, and certainly reflective of a needed shift in future research, all articles were written with an implicit (or explicit) assumption of the perpetrator being male and the victim being female, and none of the articles we found discussed the experience of LGBT students on college campuses specifically. Finally, more attention should be paid to the relationship between drinking and bystander behavior; in the sexual assault prevention literature, bystander interventions have good evidence for reducing perpetration and victimization [9, 27]. Such an approach puts the responsibility for preventing sexual assaults on the entire campus and facilitates a social environment where all students, of all genders and sexual orientations, help to protect one another.

Given the large scope of the problem of sexual assault on college campuses, we recognize the urgent need for more research into the effect of alcohol-related interventions on sexual assault outcomes, so that successful programs can be implemented across colleges and universities in the USA.