Abstract
Sexual assault is a pervasive problem on university and college campuses in the United States that has garnered growing national attention, particularly in the past year. This is the first study to systematically review and synthesize prevalence findings from studies on campus sexual assault (CSA) published since 2000 (n = 34). The range of prevalence findings for specific forms of sexual victimization on college campuses (i.e., forcible rape, unwanted sexual contact, incapacitated rape, sexual coercion, and studies’ broad definitions of CSA/rape) is provided, and methodological strengths and limitations in the empirical body of research on CSA are discussed. Prevalence findings, research design, methodology, sampling techniques, and measures, including the forms of sexual victimization measured, are presented and evaluated across studies. Findings suggest that unwanted sexual contact appears to be most prevalent on college campuses, including sexual coercion, followed by incapacitated rape, and completed or attempted forcible rape. Additionally, several studies measured broad constructs of sexual assault that typically include combined forms of college-based sexual victimization (i.e., forcible completed or attempted rape, unwanted sexual contact, and/or sexual coercion). Extensive variability exists within findings for each type of sexual victimization measured, including those that broadly measure sexual assault, which is largely explained by differences in sampling strategies and overall study designs as well as measures of sexual assault used in studies. Implications for findings and recommendations for future research on the prevalence of college-based sexual victimization are provided.
Keywords
Introduction
Sexual assault is a pervasive problem on university and college campuses in the United States. The health consequences of sexual violence are well documented and include both short-term and long-term health problems such as depression, anxiety, eating disorders, post-traumatic stress disorder, and suicidal ideation (Black et al., 2011; Campbell, Dworkin, & Cabral, 2009; Gidycz, Orchowski, King, & Rich, 2008; Kaura & Lohnman, 2007) as well as injury, sexually transmitted infections, and chronic illness (Campbell, Sefl, & Ahrens, 2003; Fisher, Cullen, & Turner, 2000). Also, college students who have been sexually assaulted are more likely to engage in risky behaviors, such as binge drinking and drug use, have lowered academic achievement and may be at greater risk for revictimization (Combs, Jordan, & Smith, 2014; Fisher et al., 2000; Gidycz, Hanson, & Layman, 1995; Jordan, Combs, & Smith, 2014).
Sexual assault on college campuses has garnered growing national attention, particularly in the past year. In September 2014, the White House Administration partnered with numerous stakeholders to launch the national public awareness campaigns entitled “It’s On Us” and “Not Alone.” The goals of these campaigns are focused on the prevention and response to sexual assault on college campuses in the United States. Despite increased attention to this issue, estimates of campus sexual assault (CSA) vary widely, which is largely due to definitional inconsistencies, including methodological and measurement differences used to estimate the prevalence of sexual assault (Catalano, Harmon, Beck, & Cantor, n.d.).
Current Study
To date, no studies have systematically reviewed prevalence findings in research on college-based sexual victimization. The purpose of the current study is to address this gap through the review of CSA prevalence rates published from January 2000 to February 2015, to propose research questions for future studies on CSA, and to present programmatic considerations and recommendations for campus prevention and response strategies. For the purposes of this systematic review, “prevalence” was defined as the reported percentages of participants within each study sample that reported sexual victimization since entering college or during a study follow-up period or time frame while attending college. The purpose and aims of studies included in the systematic review focused on topics such as health outcomes of college-based sexual assault, evaluations of campus intervention or prevention programs, and risk factors for sexual assault on college campuses. Prevalence findings, research design, methods, sampling techniques, and measures (including the types of sexual victimization measured) for each study included in the review are presented and synthesized. Implications for findings and recommendations for future research on sexual victimization among college students are also discussed.
Seminal Research on CSA Prevalence
Several seminal studies on the prevalence of CSA have been published within the last 15 years that ultimately called for improvements in campus policies on sexual assault (Fisher et al., 2000; Kilpatrick, Resnick, Ruggiero, Conoscenti, & McCauley, 2007; Krebs, Lindquist, & Barrick, 2011; Krebs, Lindquist, Warner, Fisher, & Martin, 2007). These particular studies have been widely cited in the media and by other sources, including the White House Administration. The methods used in these four studies to measure sexual victimization among college students have also been replicated in other studies. Although each of the four studies measured varying forms of sexual victimization on college campus (e.g., physically forced or forcible rape, incapacitated rape, sexual coercion, and unwanted sexual contact), prevalence findings in these studies, particularly for forcible rape rates among college women, are generally comparable.
In 2000, Fisher and colleagues published their findings from the National College Women Sexual Victimization (NCWSV) study, which was conducted with a national, random sample of 4,446 undergraduate women attending 2- or 4-year colleges and universities in Fall 1996. Fisher, Cullen, and Turner (2000) found that 1.7% of the sample experienced a completed physically forced rape and 1.1% experienced an attempted physically forced rape since the start of the fall 1996 semester, resulting in an overall victimization rate of 27.7 rapes per 1,000 female students (or 1 in 36 college women). The NCWSV also found that 15.5% of women experienced at least one form of sexual victimization during their college tenure. A second component to the NCWSV study was focused on evaluating methodological differences between the NCWSV design and the National Crime Victimization Survey (NCVS), which was one of the first nationally representative surveys (beginning in February 1997) to include sexual victimization experiences among college student populations. In the NCWSV, Fisher and colleagues (2000) found that using behaviorally specific measures, in addition to incident-based measures, led to considerably higher reporting of sexual victimization rates compared to NCVS, which only included incident-based measures: 1.7% of women reported completed rape and 1.1% reported attempted rape in the NCWSV, compared to 0.16% of women who reported completed rape and 0.18% who reported attempted rape in the NCVS.
In the CSA study, Krebs, Lindquist, Warner, Fisher, and Martin (2007) surveyed a random sample of 5,446 undergraduate women at two public universities in the United States and found much higher prevalence rates of sexual assault compared to Fisher et al. (2000). Specifically, 12.6% of women experienced an attempted sexual assault, 13.7% experienced a competed sexual assault, and 7.2% experienced both attempted and completed sexual assault since entering college (Krebs et al., 2007). In another study, Krebs, Lindquist, and Barrick (2011) examined the prevalence of sexual assault on Historically Black College and University (HBCU) campuses using the same measures in the 2007 CSA study and found comparable, though slightly lower, prevalence rates of sexual assault on HBCU campuses. Findings suggested that approximately 8% of women enrolled in the HBCUs included in the study experienced an attempted sexual assault and 9.6% experienced a completed sexual assault since entering college (Krebs et al., 2011).
It is important to note that completed sexual assault prevalence rates in the CSA studies (Krebs et al., 2007, 2011) include measures of both physically forced rape and incapacitated rape (i.e., drug- and/or alcohol-facilitated), whereas only physically forced rape was included in the NCWSV prevalence rates (Fisher et al., 2000). The 2007 CSA study provides a separate prevalence rate for “physically forced rape” (3.4%), which is higher than the physically forced rape prevalence rate in the NCWSV study (1.8%). The 2007 CSA study finding (3.4%) is also comparable to the past-year forcible rape prevalence rate (3.2%) found in Kilpatrick, Resnick, Ruggiero, Conoscenti, and McCauley (2007), which used a national sample of 2,000 college women. An important distinction, however, among these studies is the difference in survey response rates for each study, which can influence prevalence findings. The NCWSV study had an overall response rate of 85.6% compared to the CSA study (2007), which had response rates of approximately 42% at each of the two universities.
Methodological Challenges
There are a number of challenges to measuring the prevalence of sexual assault on college campuses. First, widespread underreporting of sexual assault limits the utility and accuracy of estimating prevalence from reports of sexual assault to campus officials or law enforcement (Sinozich & Langton, 2014). Second, Rennison and Addington (2014) suggest that data collection methods and the approach to the survey (e.g., crime, health, or safety focused) lead to varying rates in reported victimization. Third, definitional issues and inconsistency in the types of victimization measured across studies (e.g., forcible rape, completed rape, attempted rape, sexual coercion, unwanted sexual contact, incapacitated rape, and alcohol- and drug-facilitated rape) contribute to differences in prevalence findings (Rennison & Addington, 2014). Krebs (2014) concludes that although greater consistency in definitions and measures of sexual victimization may be beneficial for future research on CSA, some methodological variations may also be appropriate, especially if prevalence findings are comparable among studies using different measures.
Finally, the composition of the study sample is also important to consider when determining the prevalence of CSA in the United States. In their recent report, Sinozich and Langton (2014) did not limit their college sample to those attending residential colleges but included those attending trade or vocational schools and nonresidential campuses. This is an important distinction since the situational context and campus environment likely vary between residential and nonresidential schools. Finally, many studies have measured lifetime victimization among students and not sexual violence experienced since entering college. Therefore, these types of studies do not accurately represent the extent to which sexual assault occurs on college and university campuses (Combs-Lane & Smith, 2002; Vanderwoerd, 2009).
Method
A systematic search was conducted to identify studies on CSA published between January 2000 and February 2015. Databases searched included National Criminal Justice Reference Service (NCJRS) Abstracts, ERIC, Sociological Abstracts, Ebsco Academic Complete, and Lexis Law Reviews using the following key words: campus/college/university*, sexual assault/rape/unwanted sexual contact/sexual violence, violation/investigation/victimization/cases/inquire*, and incidence/prevalence/rate*. The search strategy resulted in a total of 91 documents including peer-reviewed articles, dissertations, and reports. After eliminating duplicates (n = 6), the search strategy resulted in a final yield of 85 documents. Using the same key terms listed above, a second search was conducted through ProQuest Dissertation Abstracts and Google Scholar to identify additional sources of gray literature such as dissertations and reports that might not have been identified in the first search (Littell, Corcoran, & Pillai, 2008). Librarians at the NCJRS assisted in the search process.
A total of 136 additional documents were identified through the gray literature search; however, 87 of those were identified in the original search leaving 49 new records, which combined with the results in the first search (n = 85) resulted in a total of 134 records eligible to be screened. Eligible studies for inclusion in this systematic review were empirical studies published after 2000, available in English, and that reported prevalence findings (i.e., percentages, counts, and/or rate) on sexual assault, rape, or other type of unwanted sexual contact among college students in the United States. Studies that only measured lifetime sexual victimization among college students were not included in the review, since it could not be discerned whether sexual assaults occurred since students had entered college. Therefore, only studies that included measures to identify whether students had been sexually assaulted since entering college were included in the review. Studies that included measures of both lifetime and sexual victimization since entering college were included in the review. Finally, studies were limited to self-reported accounts of sexual violence and, therefore, studies reporting exclusively on sexual harassment, perpetration of sexual violence, or disclosure of sexual victimization or perpetration to friends/peers were not included. Additionally, studies on dating violence among college students were not included in the review and, therefore, studies that exclusively focused on sexual violence perpetrated by intimate partners were not included, but studies that reported on this in addition to other types of sexual victimization perpetrated by strangers, acquaintances, peers, and friends were included.
Titles and abstracts of the 134 records were individually reviewed to identify studies that met the inclusion criteria for review. A total of 19 irrelevant records were eliminated at this stage in the process because the document did not focus on college-based sexual victimization but on sexual assault in the general population. After removing studies that did not focus on CSA as defined by the inclusion criteria (n = 19), 115 studies remained (see Figure 1 for search and screening results). Among the 115 studies, 34 documents were eligible for inclusion in the review. The remaining 81 documents were excluded from the review for the following reasons: were literature reviews, editorials/commentaries, or other nonempirical documents (n = 14); were studies of non-U.S. colleges (n = 3); did not report any prevalence findings of CSA (n = 26); sexual assault prevalence finding was not eligible for inclusion (i.e., perpetration rates, rates of disclosure to friends, sexual violence perpetrated by intimate partners, and sexual harassment) or college-based prevalence of sexual assault was not discernable from the reported findings (n = 12); measured lifetime sexual victimization or prior to entering college only and not victimization since entering college (n = 17); and were documents that published repeated findings from a study that was already included in the systematic review (n = 9). Studies that had two or more publications on findings from the same data collection study were reviewed a second time and the study that presented the most complete findings on CSA was included in the review. Some articles were vetted by a second and/or third reviewer to ensure consistency in inclusion and exclusion processes.
Systematic review process
Limitations
Despite the comprehensiveness of this review, there are limitations that might be addressed in future studies. First, the inclusion of additional electronic databases and individual websites or institutional research sources (i.e., data collected for The Clery Act requirements) might have provided additional articles and reports, particularly within the gray literature. Additionally, no hand search of gray literature sources was conducted, which might have resulted in additional documents eligible for inclusion (e.g., practitioner, nongovernmental organization, and governmental agency data/reports). Finally, the majority of studies identified by this systematic review reported findings based on primarily White, heterosexual student samples, and thus findings are primarily limited to this population.
Results
Table 1 presents the research design, sampling, form of sexual victimization, measures, and prevalence findings for each of the 34 studies included in this review. A total of 22 studies used cross-sectional observational research designs, 11 used longitudinal prospective designs, and 1 study conducted a randomized control trial with a 2-month follow-up period. Longitudinal study time frames ranged from 2-month follow-up periods to follow-ups spanning participants’ 4-year college careers. All 34 studies used behaviorally specific measures and 16 used versions of the Sexual Experiences Survey (SES; Koss & Gidycz, 1985; Koss, Gidycz, & Wisniewski, 1987; Koss & Oros, 1982).
Studies Included in Systematic Review.
Note. ACHA/NCHA = American College Health Association National College Health Assessment; ASL = American Student List; CSA = Campus Sexual Assault Study (Krebs et al., 2007); HBCU’s = Historically Black Colleges and Universities; NCVS = National Crime Victimization Survey; SA = Sexual Assault; SES = Sexual Experiences Survey; SFV = Short-Form Victimization; YAAPST = Young Adult Alcohol Problem Severity Test.
aSome studies measured additional forms of unwanted sexual experiences including sexual harassment and attempted sexual contact, as well as perpetration of various forms of sexual victimization, which are not noted here since these forms of victimization did not meet review criteria. bFindings are reported using the language used in each study and breakdowns for specific forms of victimization are reported if provided in the study. cThese findings are based on the sample from the CSA study (Krebs et al., 2007) and were included in this review since data on the prevalence of sexual assault among sexual minority women were not reported in Krebs et al. (2007). dThis study reports on findings from a college student sample recruited in a National Institute of Justice (NIJ)-funded study by Kilpatrick et al. (2007) on the prevalence of incapacitated, drug-facilitated, and forcible rape among the general population in the U.S. McCauley, Ruggiero, Resnick, Conoscenti, and Kilpatrick (2009) reported on findings among the college student sample only. eParticipant response rate is based on an instructor response rate where 63% of invited instructors agreed to have the survey administered to the class.
Sampling strategies largely consisted of convenience samples recruited through campus advertisements, courses, or specific academic departments (n = 16). Other studies randomly selected students to participate using the university’s registrar’s list, residence halls, or specific departments (n = 15). Three studies used nationally representative samples of college students, including one that published estimates from the NCVS and two that used random sampling through the American Student List, which lists over 6 million students currently enrolled in college in the United States. Sample sizes widely varied across studies (see Table 1); however, sample characteristics among studies were largely similar. All studies included in the review sampled undergraduate 4-year college women (n = 34); however, some studies sampled both college women and men (n = 12). The vast majority of studies sampled predominantly White students and few studies examined college-based sexual assault among racial and ethnic minority student populations (n = 2). Several studies examined sexual assault among at-risk, marginalized, or underserved student groups, such as lesbian and bisexual women (n = 2), deaf and hard of hearing students (n = 1), and sorority members (n = 1). Finally, nearly all studies included in the review defined “college” students as undergraduates attending residential, private or public, and 4-year colleges, though some studies also included students attending 2-year colleges (i.e., studies using the American Student List as a sampling frame) and a few included graduate students within the sample (including Andar, 2014; Hines, Armstrong, Reed, & Cameron, 2012). One study reporting on findings from the NCVS appeared to sample students attending alternative postsecondary institutions, including vocational or trade schools (Baum & Klaus, 2005).
Findings on the prevalence of sexual victimization on college campuses varied significantly among studies, making it difficult to synthesize results across the 34 studies. However, prevalence findings among studies that measured similar, if not identical, forms of sexual victimization were more easily compared. Therefore, the ranges of prevalence findings among studies measuring the same form of sexual victimization are presented below. 1 When present, outliers within these ranges are noted and corresponding study numbers (S#) from Table 1 are cited. Studies that used definitions or measures that were not standardized or comparable to other studies are excluded from the ranges presented below and discussed later.
Among studies measuring completed rape, defined as forcible vaginal, anal, or oral intercourse using physical force or threat of force (n = 9), prevalence findings ranged from 0.5% (S12) to 8.4% (S21) of college women, including 0.7% (S28), 1.7% (S6), 1.9% (S10), 3.4% (S18), 3.8% (S19), 6.4% (S14), and 7% (S16), with findings of 0.6% (S12) to 0.7% (S18) for college men. A notable outlier within this range included S12 (Hines et al., 2012) that used 3 years of cross-sectional data, where participants from years 2009 and 2010 were significantly older than participants from 2008 (i.e., participants were less likely to be first-year students and more likely to be graduate students and students living off campus). Findings for studies measuring attempted rape, defined as attempted vaginal, anal, or oral intercourse using physical force or threat of force (n = 3), were comparable and ranged from 1.1% to 3.8% (S6, S10, and S14) of college women.
Findings among studies measuring unwanted sexual contact, defined as attempted or completed unwanted kissing, fondling, petting, or other sexual touching using physical force, threat of physical force, verbal coercion, or a combination of these but excluding vaginal, anal, and oral intercourse (n = 11) ranged from 1.8% to 34% of college women. The majority of these studies (n = 7) found rates over 20% (S1, S8, S24, S28, S29, S33, and S34). Findings for college males (n = 3) varied from 4.8% to 31% of unwanted sexual contact reported with two studies having rates at 6% or less (S8 and S33). The primary outlier for findings on unwanted sexual contact is S6 (Fisher et al., 2000) that found a rate of 1.8%. Fisher and colleagues used a national random sample of 4,446 college women, compared to the remaining 11 studies that had sample sizes less than 1,000 and used either convenience or random sampling at one university. Notably, Fisher et al. (2000) measured unwanted sexual contact over a 7-month period (“since Fall 1996”), whereas other studies measuring unwanted sexual contact included reports of victimization since entering college.
Findings from studies measuring incapacitated rape, defined as completed vaginal, anal, or oral intercourse while intoxicated or while on drugs (n = 7), ranged from 1.8% (S12) to 14.2% (S28) of women, including 3.4% (S26), 4.1% (S21), 4.5% (S19), 8.5% (S18), and 13% (S16), with findings of 1.9% (S12) of men. Notable outliers for studies measuring incapacitated rape include S12 (Hines et al., 2012) and S28 (Orchowski, Creech, Reddy, Capezza, & Ratcliff, 2012). Similar to the low prevalence findings of forcible rape in Hines, Armstrong, Reed, and Cameron (2012), the low prevalence of incapacitated rape found in Hines et al. (2012) may also be due to the sample characteristics (i.e., generally older students, graduate students, and students living off campus). The potential underlying explanation for the high prevalence of incapacitated rape (14.2% in the control group) is less clear in Orchowski, Creech, Reddy, Capezza, and Ratcliff (2012), especially since this was a longitudinal study that measured sexual victimization over a 2-month follow-up period using the control group from a randomized control trial of a CSA prevention program. Study participants were recruited from one university’s psychology department (n = 143). Students were primarily in their first or second year of college (93.7%) and between 18 and 19 years old (90%), which might explain higher prevalence rates of incapacitated rape since research suggests that women are most at risk of sexual assault early on in their college tenure (Krebs et al., 2007). It is also possible that treatment diffusion effects occurred in this randomized control trial in which case the control group may have been more likely to identify and/or disclose their experiences with sexual assault over the 2-month period.
Among studies measuring sexual coercion defined as completed unwanted sexual contact (kissing, fondling, or other sexual touching) or completed vaginal, anal, or oral intercourse through nonviolent means such as intimidation, pressure, lies, threats to end a relationship, or continual arguments (n = 4), findings ranged from 1.7% (S6) to 32% (S27) of women, including findings of 3.7% (S10) and 20% (S16). Among these studies, three measured coercion used to obtain sexual intercourse only (S6, S10, and S27) and one study measured coercion used to obtain “sexual activity” including intercourse and other forms of unwanted sexual contact (S16). Findings from S6 and S10 suggest comparable rates of sexual coercion used to obtain intercourse. Higher prevalence rates of sexual coercion in S16 (Nasta et al., 2005) are likely explained by the fact that this figure includes both intercourse and other forms of unwanted sexual activity facilitated through coercion. The high prevalence rate of sexual coercion among college women in S27 (Kalof, 2000) is questionable, given that the sample size of this study was only 54 compared to much larger sample sizes in other studies.
Finally, several studies measured broad definitions of “sexual assault” and/or “rape” (n = 10) defined as a range of sexual victimization experiences including attempted and completed forcible vaginal, anal, or oral rape (physical force or threat of force), physically forced or coerced unwanted sexual contact (fondling, kissing, or other sexual touching), incapacitated rape (completed or attempted vaginal, anal, or oral intercourse while unable to consent due to intoxication or influence of drugs), and sexual coercion (completed or attempted unwanted sexual contact or completed or attempted vaginal, anal, or oral intercourse through the means of pressure, lies, intimidations, arguments, etc.). These studies typically presented findings on a combination of two or more of these forms of sexual victimization experiences, labeled them as sexual assault or rape experiences among college students, and did not provide breakdowns of the different forms of victimization experienced among participants in the study. Studies with generalized or composite prevalence findings ranged from 6% (S3) to 44.2% (SS1) of women, including findings of 6.6% (S12), 9.5% (S23), 12.16% (S5), 16.27% (S22), 20.6% (S7), 24.3% (S15), 27% (S11), and 31.5% (S32). Two studies reported sexual victimization rates of 1.4% (S3) to 3.2% (S12) among male students.
Findings from this systematic review suggest that unwanted sexual contact appears to be the most prevalent form of sexual assault on college campuses (including sexual coercion), followed by incapacitated rape and completed or attempted forcible rape. Additionally, several studies measured broader constructs of sexual assault that typically included combined forms of college-based sexual victimization (i.e., forcible completed or attempted rape, unwanted sexual contact, and/or sexual coercion) and findings from these studies typically suggest higher prevalence rates of sexual assault compared to studies measuring specific forms of sexual victimization separately. However, extensive variability exists within findings for each type of sexual victimization measured.
Discussion
Variability in Prevalence Findings
Overall, many factors influence the range of prevalence findings reported among studies on CSA. Research design (longitudinal vs. cross-sectional studies, reporting time frame measured), differences in sampling strategies (i.e., small samples vs. larger samples, convenience vs. random sampling and nationally representative samples), different sample characteristics (i.e., underclassmen vs. upperclassmen), the measures used (SES and other behavior-based measures), and variability in the constructs and definitions for sexual victimization may explain the inconsistencies in prevalence findings among studies on CSA.
As part of the inclusion criteria for this review, all studies had to include college-based measures of victimization; that is, victimization experienced among students since entering college, in the past academic year, or in longitudinal studies, victimization experienced during the study follow-up time period or academic career. Therefore, all cross-sectional studies included in this review measured sexual victimization among students since entering college, since the beginning of an academic year or quarter, or during the past 6 months. Follow-up periods in studies using longitudinal and prospective designs ranged from several months to several years. The differences in time points of data collection for both cross-sectional and longitudinal studies may explain some of the variance in prevalence findings. For example, Wiscombe (2012) and Meyerson (2002) used versions of the SES (Koss & Gidycz, 1985; Koss & Oros, 1982) to measure sexual victimization in comparable convenience samples of college women (n = 157 and 299, respectively). Wiscombe’s cross-sectional study asked participants to report victimization since entering college and found that 36.3% experienced at least one form of sexual victimization compared to Meyerson’s findings, which suggested that only 3.68% of women reported an unwanted sexual experience after a 9-week follow-up period. In this example, the variance in prevalence findings is likely due to differences in study designs and time frames. It is also possible that in any longitudinal study on CSA, participating or not participating in a second, third, or fourth wave of data collection can be associated with the reporting of a sexual assault experience to the study; that is, some sexual assault victims may be more likely to report their experience in follow-up periods if they were exposed to information on sexual assault at the start of the study and others may be less likely to participate in subsequent waves of a study, particularly if attrition is due to victims dropping out or transferring from the university. However, not all longitudinal studies report comparisons between the characteristics of respondents at each wave of data collection, so the extent to which experiences may be correlated with participation in follow-up reporting periods is unclear.
While all studies used behavior-based measures, some provided additional definitions of sexual victimization to students. For example, Banyard et al. (2007) measured unwanted sexual contact among undergraduate men and women and found that 19.6% of females and 8.2% of males experienced some form of unwanted sexual contact while in college. Banyard and colleagues (2007) used behaviorally specific measures for unwanted sexual contact (i.e., “attempting or actually kissing, fondling, or touching someone in a sexual or intimate way, excluding sexual intercourse”) and students were given a definition of what “unwanted” experiences entail; that is, experiences where participants were “certain at the time that you did not want to engage in the sexual experience and you either communicated this in some way or you were intimidated or forced by someone or you were incapacitated” (pp. 57–58). Students were given examples of possible communications that could have indicated that the advance was unwanted, “you said no, you said you didn’t want to, you cried, you protested, you physically struggled” (p. 58). Providing students with specific examples of possible behaviors or responses to unwanted sexual advances can measure the range of unwanted sexual experiences more closely, leading to enhanced recall of events and more accurate estimates of CSA.
Most studies measured different forms of sexual victimization separately (e.g., forcible rape, incapacitated rape, unwanted sexual contact, etc.); however, some studies measured multiple forms of victimization in 1 item. For example, Shenoy and colleagues (2010) surveyed a diverse sample of college women (White women = 319; Asian American women = 310; Latina women = 133; and Black women = 47) and found that 35% of women experienced a sexual assault in their lifetime with 63% of those women indicating the most recent incident occurred in college. Shenoy et al. (2010) used a very broad definition of sexual assault and asked women whether someone had ever “pressured, forced or frightened you into doing something sexually you did not want to do.” In this measure, sexual coercion, physical force, and threatened physical force for a range of possible sexual acts were measured in 1 item.
Other confounding factors, such as participants’ class rank (freshmen, sophomores, juniors, or seniors) and prior histories of sexual victimization, may also influence prevalence rates and explain some of the variability in findings (Fisher, Daigle, & Cullen, 2010; Rennison & Addington, 2014). Specifically, some research suggests that freshmen or college women early on in their tenure are at highest risk of sexual assault compared to upper-class students (Krebs et al., 2007). Research also strongly suggests that women with histories of sexual assault, including college women with childhood or prior histories of victimization, are at increased risk of revictimization in adulthood (Classen, Palesh, & Aggarwal, 2005; Desai, Arias, Thompson, & Basile, 2002). Many studies included in the review collected information both on childhood and college experiences of sexual assault. In studies where participants reported higher rates of childhood sexual assault, the higher likelihood of revictimization as an adult might explain the higher college prevalence rates found. For example, Krebs et al. (2007) found that one in five college women experienced college-based sexual assault at two universities and among participants in the study, 35% experienced moderate childhood sexual victimization, 6.4% experienced severe childhood sexual victimization, 26% experienced moderate adolescent sexual assaults, and 13.5% experienced severe adolescent sexual assaults.
Lower prevalence rates of sexual assault might be expected in longitudinal studies with shorter follow-up periods compared to cross-sectional studies measuring victimization since entering college, since less time has typically elapsed during longitudinal follow-up periods. However, this is not always true in studies included in the review. For example, Fite (2006) surveyed a convenience sample of predominately freshmen (43%) and sophomore (35%) psychology students and found that 20% of women (n = 236) reported a sexual assault (12% experienced moderate victimization or intercourse by sexual coercion or pressure and 8% experienced severe victimization or forcible rape or attempted rape) at the end of a 13-week follow-up period. This finding is comparable to and, in some cases, higher than both cross-sectional and longitudinal studies measuring similar forms of sexual assault (i.e., sexual coercion and completed or attempted forcible rape) over longer periods of time. In this example, class rank might explain the higher prevalence findings in this particular study; however, findings might also be attributed to participants’ experiences with childhood and adolescent sexual assault. Fite (2006) found statistically significant associations between childhood victimization and past semester (during 13-week follow-up period) adult sexual victimization and significant associations between adolescent victimization and past semester adult victimization. Therefore, it is feasible that the high prevalence of women in the sample with prior histories of victimization, as well as the percentage of the sample that were college freshmen and sophomores compared to juniors and seniors, impacted this study’s overall sexual assault prevalence findings.
Finally, a total of 17 studies on CSA were not included in the systematic review since they measured lifetime sexual victimization only among college students. Measuring lifetime victimization among college students is important to examine the extent of poly-victimization experiences among college students and also to develop appropriate treatment and intervention responses. However, these types of studies do not provide information on the extent of sexual victimization that college students experience while attending college/university. This distinction is important since the context of sexual assault among college-aged students is often different than adolescent sexual victimization (e.g., victim’s age, relationship to perpetrator, and risk factors), which is needed to inform prevention and intervention treatment approaches for both college-based sexual assault and adolescent victimization.
Future studies on CSA should clearly define what is meant by “college-based sexual victimization,” including the range of victimization experiences that may fall under “unwanted sexual contact,” as well as forcible rape, incapacitated rape, sexual coercion, and drug- and/or alcohol-facilitated rape. Using standardized definitions for each of these forms of victimization, as well as measuring them separately (vs. measuring multiple forms of victimization such as forcible rape and unwanted sexual contact in one item), may improve our understanding of the prevalence on various forms of sexual victimization experienced by students on college campuses. Understanding the extent of forcible rape, sexual coercion, unwanted sexual contact, and incapacitated rape could better facilitate the development of appropriate prevention and intervention strategies for these various types of victimization.
Sexual Coercion and Unwanted Sexual Contact
Based on findings from the studies reviewed, the prevalence of unwanted sexual contact and sexual coercion is the highest among all forms of sexual victimization reported by college students. This suggests a need for future prevention programming focused on reducing common elements of unwanted sexual contact and sexual coercion on college campuses. For example, Orchowski and colleagues (2012) found that 25.4% of women had experienced unwanted sexual contact through the use of arguments or pressure at a 2-month follow-up period. Orchowski also found that 14.2% of women experienced unwanted intercourse by administration of alcohol and/or drugs and 9.7% experienced unwanted intercourse by sexual coercion (i.e., arguments and/or pressure). Flack and colleagues (2007) examined both unwanted sexual intercourse and unwanted fondling within the context of the college “hookup” culture. Hookups were “casual encounters … with no expectation of future commitment” in which “pressure to engage in unwanted sexual behavior” was associated with the worst reported hookup experiences. The authors found that 78% of the unwanted intercourse (vaginal, anal, and oral) in their study occurred in the context of a student-defined hookup. Alcohol use by the victim was also noted to occur alongside unwanted sexual experiences within this context, as defined by participants reporting that either their judgment was impaired by alcohol or drugs or they were taken advantage of because they were “wasted.” In their 2008 study, Flack and colleagues noted that temporal increases in reported sexual victimization occurred simultaneously with fraternity and sorority pledging seasons, which are characterized by more frequent social gatherings and parties that combine a social expectation of causal intimacy, alcohol, and under- and upperclassmen.
A number of studies included in the review defined unwanted sexual victimization in the context of the victim saying “no” or resisting, or that the perpetrator used force, intimidation, or pressure, or the victim was unable to consent due to incapacitation. However, the lack of physical or verbal resistance (e.g., saying no or protesting) does not automatically imply consent. For example, in a discussion on the neurobiology of sexual assault, Campbell (2012) notes that students and law enforcement might be unaware of the “freeze” response to an unwanted sexual experience, in which a victim is physically incapable of resistance due to an autonomic survival response called “tonic immobility” or “rape-induced paralysis.” Future studies might consider the inclusion of this information for clarification purposes to obtain more accurate estimates of victimization. Social norms relegating “real” sexual violence to being between strangers or that where a victim physically fights off an attacker need to be addressed in study designs where participants may be internalizing these social norms and therefore not fully recognizing or reporting their own experiences as a sexual victimization (Corrigan, 2013; Thapar-Björkert & Morgan, 2010).
In a review of CSA prevention efforts, Banyard (2014) calls for developmentally appropriate programming that addresses co-occurring and key risk factors for sexual victimization. First semester programming might include messaging that defines sexual coercion, consent (including the requirement of an affirmative consent rather than a lack of resistance, the impact of alcohol on one’s capability to consent, and increased sexual victimization risks associated with substance use), and healthy relationship factors. Additional and ongoing sexual assault prevention programming might focus on educating students to identify and intervene in situations where peers might be perceived as at risk for sexual assault, such as those addressed in bystander intervention programs (Banyard, 2014). Furthermore, campuses should attempt to identify incidents of sexual victimization by promoting and supporting disclosure in a variety of manners, reducing victim blaming throughout disclosure, investigation, and adjudication processes, and providing real-time intervention and services to students troubled by experiencing or witnessing sexual victimization.
Unacknowledged Sexual Victimization
The use of behaviorally specific measures (vs. incident-based measures that use terms like rape) improves upon some of the methodological challenges to measuring sexual assault. Similarly, measuring the extent of unacknowledged sexual assault among college students may also address methodological barriers to measuring CSA prevalence. Some research has examined unacknowledged rape experiences since many individuals are unlikely to identify or label their experiences as rape or sexual victimization (Gavey, 1999; Littleton, Rhatigan, & Axsom, 2007). Although it was not specific to college-based sexual assault, Wilson and Miller (2015) conducted a meta-analysis on the prevalence of unacknowledged rape among survivors of sexual assault. Wilson and Miller found that the prevalence of unacknowledged rape appears to be highest among college students and explained that sexual victimization experiences are often cognitively labeled among victims as “bad sex” or “miscommunications.”
Very few studies, however, have focused on unacknowledged sexual victimization experiences on college campuses, including rape and other forms of sexual victimization among college students. Among a sample of 302 college women, Cleere and Lynn (2013) found significantly higher prevalence rates of unacknowledged sexual assault among college women compared to acknowledged victimization experiences. Approximately 39% were nonvictims, 15% were acknowledged victims, and 46% were unacknowledged victims of attempted rape, completed rape, unwanted sexual contact, and sexual coercion. However, Cleere and Lynn (2013) did not explicitly measure victimization experiences since entering college, though they did ask participants about recent victimization experiences in the past 6 months, past year, and past 2 years. Although the majority of women reported sexual victimization 3 or more years ago (47.8% of acknowledged victims and 17.3% of unacknowledged victims), 15.3% of acknowledged victims and 30.8% of unacknowledged victims reported a sexual assault in the past 6 months. An additional 10.9% of acknowledged victims and 26.9% of unacknowledged victims reported an assault in the past year and 26.1% of acknowledged victims and 25% of unacknowledged victims reported an assault in the past 1–2 years. Also, the majority of women with unacknowledged sexual victimization experiences were more likely to be intoxicated at the time of the assault compared to women with acknowledged victimization. Future studies on CSA might include measures of acknowledged versus unacknowledged sexual assault in addition to behavior-specific measures to examine the prevalence of specific forms of sexual violence among unacknowledged victims versus acknowledged victims. Outreach to unacknowledged victim populations to provide supportive services and resources should also be considered.
Conclusion
Unfortunately, college students experience a wide range of sexual victimization while attending college. The range of prevalence findings for forcible rape, incapacitated rape, unwanted sexual contact, and sexual coercion on college campuses varies widely, which is largely related to how these victimization experiences are measured and defined in studies as well as overall differences in research design and methodology. Despite the variability in the rates of sexual victimization reported among the studies discussed in this article, there does appear to be a trend toward higher rates of victimization when examining overall sexual victimization experiences (i.e., composite or generalized prevalence) versus individual types of sexual victimization. Therefore, it is important to interpret rates of sexual victimization with caution and properly attribute it to the specific or generalized type of victimization measured.
Nearly all studies included in this review sampled predominately White, heterosexual, female students and, therefore, the majority of prevalence research on CSA is limited to this population of students. There is a dearth of research on sexual victimization among minority college students. Some studies included in this review explored populations that might be at higher risk of sexual victimization on college campuses, including lesbian and bisexual women, students with disabilities, racial and ethnic minority women, sorority women, and students with prior histories of sexual victimization. All of these studies found higher rates of sexual victimization among these student populations. These findings are consistent with research on sexual victimization among nonstudent samples and within the broader lesbian, gay, and bisexual population (Rothman, Exner, & Baughman, 2011), among women with physical and intellectual disabilities (Martin et al., 2011), racial and ethnic minority women (Black et al., 2011), and in populations with polyvictimization experience in adolescence and adulthood (Finkelhor, Ormrod, & Turner, 2007). No studies included in the review focused on sexual victimization among transgendered students; however, research suggests that one in two transgender individuals experience sexual abuse or sexual assault in their lifetime (Kenagy, 2005; Kenagy & Bostwick, 2005). Disproportionate rates of victimization among students from these populations highlight the need for further prevalence research among at-risk special populations, culturally specific responses to victims, and efforts to make prevention programming inclusive on college campuses.
Finally, most studies included in the review defined and surveyed college student samples who were undergraduates attending residential 4-year colleges and universities. Few studies have sampled students attending vocational or trade schools or explored how the prevalence of sexual assault among nontraditional college student populations may differ from traditional full-time students attending 4-year colleges. A recent Bureau of Justice Statistics report (2014) examining data from the NCVS did sample vocational and trade school students in addition to part-time and full-time college students attending 4-year and 2-year colleges; however, the prevalence of sexual victimization was not distinguished between the different types of post-secondary enrollment (Sinozich & Langton, 2014). Future research is needed on how the experiences of sexual assault among students attending alternative college education programs may be similar or different to traditional college students in order to develop appropriate intervention and prevention responses for nontraditional college students. Additionally, contextual and cultural differences between public and private universities (e.g., small liberal arts colleges vs. large public state universities) as well as 4-year colleges and vocational or trade schools should be considered when measuring the prevalence of sexual victimization on differing types of postsecondary education campuses.
The range of sexual victimization experiences among college students, as well as the variability in prevalence findings, has substantial implications for victim services, including the provision of appropriate treatment and legal services. Treatment responses, as well as campus responses to victims, will vary depending on the types of sexual victimization experienced by students. Campus prevention and intervention strategies should focus on the unique dynamics and needs of particular campuses and universities, including diverse student body populations. The prevalence of different types of sexual victimization (e.g., forcible rape, incapacitated rape, and drug-facilitated rape) may differ from campus to campus, and therefore, prevention, intervention, and victim services efforts should start with a detailed understanding of the specific needs of a campus population.
Key Findings
The prevalence of completed forcible rape, incapacitated rape, unwanted sexual contact, and sexual coercion measured on college campuses widely varies in the United States.
Measurement and definitional differences of sexual assault largely contribute to variability in findings.
Unwanted sexual contact and sexual coercion appear to be most prevalent on college campuses followed by incapacitated rape and attempted or completed forcible rape.
The majority of research has sampled predominately White, heterosexual, female students attending 4-year colleges and prevalence findings are limited to this population.
There is limited research on sexual assault among minority students or particularly at-risk student populations; however, some studies included the review found higher rates of sexual assault among these populations (e.g., lesbian and bisexual women, sorority women, and students with prior histories of victimization).
Recommendations for Practice, Policy, and Research
Practice
College students experience different forms of sexual victimization. Treatment responses, as well as campus responses to victims, should be related to the types of sexual victimization experienced by students.
High prevalence of unwanted sexual contact and coercion suggests that prevention efforts focus on the dynamics of these two forms of victimization.
Disproportionate rates of victimization among students from minority populations (LGBTQ [lesbian, gay, bisexual, transgender, queer], students with disabilities, racial, and ethnic minorities) highlight the need for inclusive and culturally specific responses to victims.
Policy
Despite discrepancies in prevalence findings, a significant proportion of college students experience sexual assault even in studies with lower estimates. Therefore, campus-wide prevention and response policies are necessary.
The prevalence of different forms of sexual assault may vary from campus to campus; therefore, prevention and intervention strategies should start with a detailed understanding of the specific needs of a campus population.
Research
The range of victimization experiences that may fall under “unwanted sexual contact,” as well as forcible rape, incapacitated rape, and drug- or alcohol-facilitated rape should be clearly defined and measured separately. Standardized definitions can enhance our understanding of how prevalence rates vary and inform the development of appropriate prevention and intervention strategies for various types of sexual victimization.
Behaviorally specific measures should continue to be used in studies, including presenting students with examples of unwanted sexual experiences, such as those provided in Banyard et al. (2007).
Future studies should measure sexual victimization among specific populations of students who may be at greater risk for sexual assault (e.g., LGBTQ students, students with past histories of sexual victimization, and students with disabilities).
Measuring sexual victimization “since entering college” is necessary to accurately estimate the extent of campus sexual assault and distinguish these experiences from lifetime, childhood, and adolescent sexual victimization.
Footnotes
Authors’ Note
Opinions or points of view expressed are those of the authors and do not necessarily reflect the official position or policies of the U.S. Department of Justice.
Acknowledgments
The authors thank the National Criminal Justice Reference Service (NCJRS) Librarians Scott Hertzberg and James Fort for their assistance in the systematic search process.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.