Friends, Strangers, and Bystanders: Informal Practices of Sexual Assault Intervention

Abstract

Sexual assault is a part of many students’ experiences in higher education. In U.S. universities, one in four women and one in ten men report being sexually assaulted before graduation. Bystander training programs have been shown to modestly reduce campus sexual assault. Like all public health interventions, however, they have unintended social consequences; this research examines how undergraduate men on one campus understand bystander intervention and how those understandings shape their actual practices. We draw on ethnographic data collected between August 2015 and January 2017 at Columbia University and Barnard College. Our findings show that university training and an earnest desire to be responsible lead many men to intervene in possible sexual assaults. However, students’ gendered methods target more socially vulnerable and socially distant men while protecting popular men and those to whom they are socially connected. Students’ actual bystander practices thus reproduce social hierarchies in which low prestige may or may not be connected to actual risks of sexual assault. These results suggest that understanding intragroup dynamics and social hierarchies is essential to assault prevention in universities and that students’ actions as bystanders may be effective at preventing assaults in some circumstances but may lead to new risks of sexual assault.

Sexual assault is a significant public health problem for undergraduates worldwide (D’Abreu, Krahé, & Bazon, 2013; Hayden, Graham, & Lamaro, 2016; Krahé et al., 2015; Ogunwale & Oshiname, 2017; Schuster, Krahé, Ilabaca Baeza, & Muñoz-Reyes, 2016; Valls, Puigvert, Melgar, & Garcia-Yeste, 2016). There is a dearth of evidence-based sexual-assault prevention, and bystander programs are one of the few strategies with some demonstrated efficacy at US universities (Banyard, Moynihan, & Crossman, 2009; Mabry, A., & Turner, M. M., 2015). Most bystander intervention trainings teach students: (1) to confront social norms that promote or condone sexual violence, (2) how to intervene safely as bystanders, and (3) how to support survivors of sexual violence (Moynihan & Banyard, 2008). Little is known about how students interpret these trainings or how students actually act as bystanders.

Like their peers at institutions across the US, Columbia students receive bystander intervention training as part of a host of institutional efforts to prevent sexual assault. We found that—inspired to be bystanders out of concerns about individual and collective reputation, liability, and moral personhood—undergraduates at Columbia undertook a variety of bystander activities. We demonstrate however that these bystander practices were heavily and heteronormatively gendered and enmeshed in social relationships on campus; these practices may be effective in some circumstances, but in others may actually create new risks for sexual assault.

We draw here on 16 months of ethnographic research to describe the actions that some students take as bystanders to prevent sexual assaults on campus. This paper does not evaluate the efficacy of bystander interventions in achieving their stated goal. Rather, in the tradition of social scientific work examining the unintentional social and cultural effects of public health practice (Hirsch et al., 2010), we argue that the uptake of the bystander idea is shaped by the strong differentiation between ingroups and outgroups, so that students’ methods target less prestigious and socially distant men while protecting the reputations of men to whom they are socially or organizationally connected.

BACKGROUND

Research suggests that campus sexual assault is a problem in both the Global South (Ogunwale & Oshiname, 2017; Schuster et al., 2016; D’Abreu et al., 2013; Lehrer, Lehrer, & Koss, 2013) and the Global North (Hayden et al., 2016; Krahé et al., 2015; Valls et al., 2016). Bystander interventions are one of the few prevention methods with some demonstrated efficacy (Banyard et al., 2009; Mabry, A., & Turner, M. M., 2015), and they are being rolled out at many U.S. institutions of higher education. Bystander interventions train members of a community to develop awareness of situations where a person might need help (specifically, where someone is at risk of being sexually assaulted), and inculcate the skills and confidence to intervene in such contexts.

Students’ willingness to intervene varies by demographics, with traditionally college-aged (i.e. 18–22) white men from urban areas having the highest rates of rape myth acceptance and the lowest support for bystander attitudes, while traditionally-aged women of color from rural areas demonstrate the opposite beliefs (Diamond-Welch, Hetzel-Riggin, & Hemingway, 2016). Undergraduates report a range of behaviors they would use to intervene as bystanders (McMahon et al. 2013) that vary by gender (Koelsch, L. E., Brown, A. L., & Boisen, L., n.d.).

Evaluations have shown that bystander interventions can positively impact students’ beliefs, attitudes, and behavioral intentions on sexual assault prevention (Mabry, A., & Turner, M. M., 2015; McCauley et al., 2013; Senn, C. Y. & Forrest, A., 2016). Students who receive bystander intervention training, even if only through hearing lectures, report less rape myth acceptance and higher rates of active bystander behaviors (Coker et al., 2014). Immediately after bystander training, men in one study reported less sexual aggression and less support for sexual aggression, as well as fewer associations with sexually aggressive peers, though not all of these effects were not present at 4- and 7-month follow-ups (Gidycz, C. A., Orchowski, L. M., & Berkowitz, A. D., 2014).

Existing research on bystander attitudes and intentions focuses almost exclusively on individual-level factors; little is known about the social factors that might promote or deter bystander intervention behaviors or shape how bystanders intervene (Banyard, V., 2011; Jacuqes-Tiura et al., 2015). Social barriers (e.g. failure to notice a situation, failure to take responsibility for a situation, etc.) may make a bystander less likely to intervene to prevent a sexual assault; conversely, having a relationship with the victim or perpetrator makes intervention more likely (Burn, S. M., 2009). Research shows that men who drink heavily report less supportive attitudes towards bystander intervention than those who do not; regardless of gender, students using alcohol report difficulties figuring out if a woman is at risk and if they think she deserves help (Pugh, B., Ningard, H., Ven, T.V. & Butler, L., 2016). However, barriers to intervention are better predictors of bystanders’ helping behavior towards strangers than towards friends (Bennet, S., Banyard, V. L., & Garnhart, L, 2014), suggesting the need to consider whether individuals view their role as a bystander in socially mediated ways. Almost all of the work in this area draws upon self-reports, introducing the challenge of social desirability bias to the positive ‘observed’ effects of bystander training (i.e. those that receive training are more likely to report that they do what they’ve been told they should be doing). By directly observing students’ interventions (or lack thereof), participant observation offers an important corrective to accounts that rely solely upon self-reports.

Experiences of violence—both as vulnerability to and capacity to commit—underlie the power structures of gendered inequality (R. Connell, 1987). In wider US society, femininity is associated with a vulnerability (real or imagined) to violence (Dowling, 2002; Hollander, 2001), while masculinity is linked to a potential capacity for violence (Hollander, 2001). Violence is also used to demarcate hierarchical domination among masculinities (R. W. Connell, 2005). Individuals even ‘mobilize rape’ as a means of both marking men with prestigious forms of masculinity as ‘the good guys’ and emasculating their real or imagined ‘rapist’ peers (Pascoe & Hollander, 2016). Often, it is men of color who stand to lose from such deployments (Hollander, 2001; Messner, 2016; Pascoe & Hollander, 2016). These gendered ideas about physical power and violence are embedded within bystander prevention efforts themselves that call upon ‘good’ men to protect women from aberrant men who rape (McCaughey & Cermele, 2017; Messner, 2016). Furthermore, the premise in bystander interventions that a few bad men commit most campus assaults does not hold up to empirical scrutiny (Swartout, et al., 2015).

Drawing on how sexual violence figures into constructions of gendered prestige and using data from extended ethnographic research among undergraduates, we examine how college students act to prevent what they understand to be potential sexual assaults. Our descriptions of how students perform bystander interventions show that they actively interpret the educational information demonstrated to be effective by these earlier studies in unpredictable and innovative ways, resulting in actions that do seem to prevent certain forms of sexual assault, but may also generate new risks for other forms. We argue that efforts to prevent college sexual assault must take social prestige and intergroup dynamics into account.

DATA AND METHODS:

This paper presents findings from the ethnographic component of a large mixed-methods study of sexual health and sexual violence at Columbia University and Barnard College¹ (for a comprehensive outline of this study, see Hirsch et al., 2018). This paper draws exclusively on the ethnographic research, the methods of which are summarised in Table 1.

Table 1:

Qualitative Research Methods

Method of Research	Researcher	Total Materials Collected	Location	Question Addressed
In-Depth Interviews	Subjects chose a male or female interviewer and/or a faculty members Most subjects were interviewed by the person who first made contact with them about participating in the study	151 in-depth interviews 26 follow-up interviews	Private Offices on Columbia Campus	Collected information on subjects’ precollege life, orientation experiences, sexual intimacy, and social relationships to answer questions about the individual, interpersonal, and institutional level factors shaping vulnerability to sexual assault
Focus Groups	Conducted by faculty member at Columbia, with assistance from a doctoral or professional school student at Columbia	17 focus groups; averaging in size of 10 students Varied in composition. Some interest groups (e.g. LGBTQ+), some single sex, some just a mix of students.	Private Offices on Columbia Campus	Collected information on the normative contexts for sexuality, sexual relationships, socializing, and student behavior on campus.
Participant Observation	Participant observation conducted by six researchers Columbia faculty members only conducted participation observation at public events (e.g. sports games) Doctoral student and unaffiliated researchers conducted all participant observation in student-controlled spaces	Dorms 160.4 hours Special interest houses (residential) 9.5 hours Fraternity/sorority 18.75 hours Dining halls 8.5 hours Religious spaces 19 hours Ethnic and cultural spaces 27 hours Outdoor spaces 54.5 hours Campus spaces (including outdoor campus spaces) 185.5 hours Off-campus spaces 111.25 hours	Dorms 160.4 hours Special interest houses (residential) 9.5 hours Fraternity/sorority 18.75 hours Dining halls 8.5 hours Religious spaces 19 hours Ethnic and cultural spaces 27 hours Outdoor spaces 54.5 hours Campus spaces (including outdoor campus spaces) 185.5 hours Off-campus spaces 111.25 hours	Collected information on spatial dimensions of socializing, unarticulated but widely shared norms, and student interpersonal dynamics.

Data collection for this study ran from August 2015-January 2017. All subjects were current undergraduates at Columbia University or Barnard College. Ethnographic research participants were either recruited during participant observation by a member of the research team, or else participated as a result of having reached out to the research team after hearing of the study through campus-wide emails, fliers, tabling events, sponsored study breaks, or from a community member. Researchers gathered demographic information on all subjects to ensure that they represented a range of backgrounds. The two faculty investigators leading the SHIFT ethnography, Hirsch and Khan, developed all ethnographic research instruments.

The study relied upon two advisory boards. One board was a group of approximately 15 students, all of whom were ineligible to participate as subjects in the research. Members of the research team met with this advisory group every week for 2 hours to discuss research ideas. The purpose of this group was to develop a better understanding of student culture so that basic design features would be responsive to the local environment and to help introduce the research team to different student communities. These students were never shown any data, nor were they allowed to know the identity of any study participants. Advisory-board members were paid for each semester of participation. The second advisory board was comprised of key institutional actors such as deans or other student-facing senior administrators. The principal investigators met with this board approximately twice each semester. Because the research was intended to generate policy recommendations that touch many divisions’ areas of responsibility, administrators were integrated into the design from the start of research.

Seven team members—three men and four women—conducted ethnographic research, representing a range of racial and class backgrounds, as well as sexualities and religions. At any one time up to five researchers were in the field. Columbia University’s interpretation of federal Title IX regulations requires all University faculty and staff, except for those who interact with students as clinicians, providers of pastoral care, or confidential sexual assault resources, to be mandated reporters of any suspected sexual assault. The SHIFT research team was granted a research-specific exemption from mandatory reporting requirements. The ethnographic team met at least once a week for at least two hours to discuss all fieldnotes, observations, and interviews produced that week. The research team stripped all identifying information in the interviews from the data, and stored the data on a secure site that only investigators could access. Fieldnotes were written in a standardised format without identifying information, and all names were changed. Fieldnotes were coded using emergent thematic coding, and codes were standardised across all fieldnotes after data collection concluded in conversation with all team members (e.g. fieldnotes containing references to parties written before the code ‘party’ was used were retroactively coded as such).

Driven by the study’s framing examining the individual, interpersonal, and environmental-level factors shaping sexual health and sexual assault among undergraduates, the ethnographic data analysis used an inductive approach. After being transcribed and checked, two research team members independently coded interviews. Investigators used eleven themes for coding, reflecting the descriptive elements the research team judged to be most critical to capturing the breadth of students’ social and sexual lives based on an analysis of relevant literature on sexual assault and shared readings of fieldnotes: socializing, partner selection, relationships, sexual projects, stories of sexual assault, consent, telling someone about sexual assault, mental health experiences, alcohol and substance use, sexual experience (not assault), and other notes. The investigators also wrote up detailed narratives of each incident of sexual assault. This document included information on the victim’s personal, interactional, and institutional experiences, before the assault, leading up to, during, and shortly after the assault, and then some time after the assault.

This paper draws on incidents of bystander intervention witnessed in participant observation and recounted in IDIs, as well as all sexual assaults reported in research. Analysis examined when interventions were successful, not successful, or possible but not taken, as well as students’ explanations for their responses, with the goal of surfacing the shared social logic underlying how students apply bystander strategies. Data analysis was conducted by the lead author, Hirsch, and Khan.

RESULTS

Gendering bystanders and victims

Bystander intervention training given by the university during orientation emphasises that anyone can be a victim of or intervene as a bystander to sexual assault. Students, however, reported gendered expectations about who could be a victim and who could (and should) act as a bystander. Most students understood sexual assault within the gendered context of a heterosexual coupling, noting (as has been shown by larger national studies and our own survey research, Cantor et al., 2015; Mellins et al., 2017) that women were most at risk for sexual assault, although some heterosexual students and most queer students noted that trans, gender non-binary, and queer men could also be vulnerable. Most students were skeptical that heterosexual men could be victims of sexual assault; those who did acknowledge men’s possible victimisation emphasised that although men could be sexually assaulted in technical definitions of the term (namely, unwanted sexual contact) they did not undergo the psychological experiences of ‘real rape’ (see Hirsch, Khan, Wamboldt, & Mellins, n.d. for more details on students’ gendered understandings of consent).

Most students reported that heterosexual women were most likely to be assaulted, but only heterosexual men consistently reported acting as bystanders and were seen to be potential bystanders. In interviews and field observations, heterosexual men consistently spoke about how they had or would intervene and how their student organizations intervene in potential sexual assaults. In contrast, many women reported both procedures for avoiding individual danger, including socializing and drinking strategies, and criteria they wanted potential partners to demonstrate before having sex. They rarely spoke of intervening on behalf of a close friend. This gendered understanding reproduced the ideas that only men are potentially dangerous and that vulnerable women need protection by men.

Heterosexual men decide to intervene as bystanders for three reasons: liability, reputation, and moral commitment. Fraternities, athletic teams, and student organizations face organizational liability should an incident occur at an event they sponsor. One fraternity brother told us about how brothers need to purchase insurance that won’t cover students as alcohol is served at events: ‘[J]ust because the national organization makes us do it… [W]e have to be super careful because we have really no insurance if anything happens. So we always have to be very meticulous about how things happen…’ Students living in a building where an organization’s party took place may be held legally and financially liable for an assault.

Men also worried about the reputational consequences of not stopping an assault or of being associated with, in students’ words, ‘rapey’ people. During the course of the study, students routinely exchanged information with one another about known dangers, and diligently avoided and ostracised students or organizations known to be rapey. One man admitted that he pitied a known ‘rapey’ man as ‘honestly, in a way… he’s a fucking pariah… [A]s shitty as it sounds, I don’t fucking want to be seen sitting with him at the dining hall because then you’re like friends with the raper, you know… It could… fuck up your life too.’ Students saw proactive bystander intervention practices as the surest way to demonstrate their and their organizations’ moral status as ‘good men.’ For men, removing potential offenders from their events was a strategy to prevent their peers from perceiving them as responsible for or condoning of sexually aggressive behavior.

Moreover, the men in our study understood their organizational practices not just as liability reduction or reputation management (although they are), but also as what a moral human being would do. They overwhelmingly viewed assault as a moral violation, and understood themselves as moral, responsible people; they thereby expressed a duty to stop sexual assault. These expressions were often passionate, expressing a deeply-held moral duty. The president of one student group recounted having stopped a man at a recent event who repeatedly asked a woman out, only to hear later that a different woman that night had been sexually groped: ‘[T]he fact that it can… happen under the radar like that is scary…[T]he fact that… I saw that and was able to stop it, but couldn’t stop… the really bad things is… just sad.’

How Students Intervene

In interviews, casual discussions, and observations nearly every man described a wide range of tactics to intervene, including speaking with a person about their behavior, escorting an intoxicated student out of the party to their home or a safe space, stopping a student using physical force, and getting a student high on marijuana so that they would, in the words of one man, ‘fall asleep or get chill and calm down, or… just get the spins and start puking.’ Students’ criteria for determining if a man’s behavior merits intervention included vague feelings like ‘giving me bad vibes,’ showing that—in the absence of clear and agreed-upon criteria to identify a situation that requires intervention in trainings—students rely on their own personal judgments.

Men intervened differently depending on their relationship to and the gender of the person deemed to be creating a risky situation, either by acting aggressively or by putting themselves at risk. Table 2 shows how students’ self-described responses to a dangerous situation vary based on the gender of the person intervened with and their relationship to the student intervening. Students tried to remove intoxicated women from parties: if they knew them, by personally taking them to an empty, lockable room or to their home; if they did not, by finding the woman’s friends and having them do so. If the target of the bystander intervention was a man with whom a bystander shared a social connection, strategies were generally understated, indirect, and private; students tried to distract or incapacitate friends behaving inappropriately, but they did not embarrass them publicly or raise broader awareness of their problematic behavior by either forcibly removing or fighting them. However, students routinely demonstrated and endorsed public and violent interventions with men outside of their organization or friend groups.

Table 2:

What Interventions Look Like in Men’s Words

Gender of person acting inappropriately	Relationship to bystander	Intervention strategies
• Woman	• Friend	• Direct/Defuse: ‘Walk her home’ • Direct/Defuse: ‘[T]ake her upstairs, put her in bed on the couch or something so that she’s safe’
• Woman	• Unknown	• Delegate: ‘[F]ind one of her friends… and just be like, “take your friend home’”
• Man	• Friend • Club Member • Fraternity Brother Teammate	• Distract: ‘Let’s go play some [beer]pong’ • Distract [Incapacitate]: ‘Smoke him up’ • Defuse: ‘[D]efuse the situation’
• Man	• Not a friend • Not in a group with bystander	• Direct: ‘Kick him out’ • Direct [Violent]: ‘Grab the guy in a headlock and forcefully remove [him]’

Unintended Social Impacts

Stopping or being accused of potentially perpetrating a sexual assault is part of the construction of masculine prestige structures on campus. Men who are institutionally positioned to control social spaces (e.g. fraternities with houses, student groups whose members held an entire suite of dormitory rooms) earn reputations as good men by patrolling their events. A man self-described as ‘very renowned for being the kind of noble shit kind of guy’ in a prestigious fraternity noted that if he sees ‘a guy coming off too strong’ at parties he hosts his ‘first step is… getting them off the property as soon as possible.’ However, he contrasted this approach with a subtler, educational method used for fraternity members: ‘there’s guys… that come off too strong to some girls in our fraternity just because they’re younger kids that don’t exactly know how to talk to girls… And we try to help them… Be like, “You need to chill a little bit more.”‘

Being publicly asked to leave an event labels a man as rapey; other students may then avoid them out of fear both for their safety or for the appearance of condoning ‘such behavior.’ One woman spoke about avoiding a known rapey man in her dorm who had been asked to leave a party his first year as ‘I just worr[y] that it’s… sending some sort of signal… like if I hang out with him… it sends a signal to other people that… he’s trustworthy.’ As such, men who are institutionally positioned to control event spaces with their friends and to intervene against less connected men are able to accrue social prestige through interventions, whereas men without networks that give them control over social space are more likely to be targeted for intervention and thus perceived as rapey.

These practices do not protect against some forms of sexual assault. Most obviously, we learned of several sexual assaults over the course of our fieldwork that occurred immediately after the victim and the person who assaulted them had been in a public space where others observed their behavior and inferred mutual sexual desire. We saw no evidence of sexual assault facilitation; instead we would suggest that in the vast majority of these cases, students assumed that the person assaulted was ‘into’ the person who assaulted them and thus they did not intervene. It is critical to remember that all of these behaviors – perpetration and bystander intervention – occur in a social context in which students are also trying to facilitate sexual encounters their friends desire (Hirsch et al., n.d.). In other words, students are being trained to intervene to stop nonconsensual sexual encounters, but in those same contexts they are frequently simultaneously trying to facilitate consensual sexual encounters, all under conditions where often they are drinking and where verbal communication is limited and potentially difficult (the contexts often include loud music or many people at once).

The counterpart of students’ avoidance of men and organizations known to be rapey is that clubs, fraternities, teams, and individual men without these labels are perceived to be safe. This provides students with a potentially ill-justified sense that sexual assault is less likely in such venues or with such company. Take one woman’s account of her rape at a prestigious fraternity her freshman year, where bystanders not only did not stop her assault but also contributed to it by suddenly leaving her alone in a room with her ‘prestigious’ perpetrator:

I was embarrassed because I was so dumb to… come there in the first place, and… not realize that… I was only going there to… have sex with this guy… I thought, ‘Oh, I’m cool, I’m getting invited to [prestigious fraternity]… I’m [a] freshman, [and] I’m getting invited to [prestigious fraternity]. That’s so cool…’ [A]nd I was embarrassed because I… felt dumb for not knowing why I was invited.

Because of the fraternity’s prestige, she not only did not recognize the potential risks, but also felt embarrassed for not understanding that a brother would have ‘reasonably’ expected sex from her in return for an invitation. Over the course of the study, the preponderance of sexual assault incidents about which we learned were related to students or student organizations not marked as rapey. In avoiding reputationally rapey students and organizations, our data suggests students may overlook potential risks from unmarked groups.

Students’ gendered understandings of sexual assault and bystander intervention also contributed to the social illegibility of some forms of assault. Students assumed heterosexual men were not at risk for assault and reported not intervening to stop women making sexual advances to intoxicated men even after noticing the women’s sexual intent and the men’s inability to consent. In instances of heterosexual men being assaulted (see Hirsch et al., n.d.), other students were present immediately prior to the incident and noticed both the man’s incapacity and the woman’s sexual intent, but in no instance did anyone intervene to interrupt those assaults (although these incidents were still far less prevalent than heterosexual men assaulting women (Mellins et al., 2017)).

Finally, students assumed their own and others’ good motives – but some interventions actually created risks for the ostensible ‘beneficiaries’ of these bystander practices. Several sexual assaults occurred after men took on the role of bystander and walked intoxicated women home. These men either assessed that the woman was too drunk to be at a party, or others had assessed this and asked a man to be the responsible bystander and walk the woman home. For one woman, her ex-boyfriend and his friend were asked by her friends to escort her home after they identified her as being too drunk to stay in the bar; in her later retelling, the men took her flirtatious behavior and their previous relationship to indicate sexual interest, and subsequently raped her. No students interviewed or observed during fieldwork viewed men acting as bystanders to be possible perpetrators. Moreover, the framing of men as good bystanders and women as potential victims, as others have anticipated (McCaughey & Cermele, 2017; Messner, 2016), reinforced stereotypes about women’s relative weakness and (good) men’s chivalry, even when students’ experiences seemingly ran counter to this gendered framing. Take, for instance, this woman’s narrative of being ‘helped’ by a friend:

I think he was… a really good guy…. And… kind of saw that I was… sick and was like… ‘[L]et’s get you home’… And so I – I got home. I was fine. I was just – I didn’t really remember if anything happened. In retrospect I don’t think that we had sex…. But I was worried and… that week I got mono and my period stopped so I was super worried that we’d had sex. And I… took a pregnancy test… But no, I was just like really sick. But yeah, no, that was really scary.

In her own account, this ‘really good guy’ had noticed that she was too intoxicated to safely remain at a party and walked her home, at which point he may have sexually assaulted her. She was uncertain about what transpired next, and sufficiently worried to have sought a pregnancy test – and yet her story emphasises her own incapacitation and his laudable gendered behavior in ‘assisting’ her.

CONCLUSION

Like all single-sited research, generalizability to other settings is a limitation; our findings lay the groundwork for comparative ethnographic research in other higher education settings on how students take up and enact sexual assault prevention messages. The students we studied showed a high degree of stigma towards sexual assault perpetration; this may be a result of a unique student population at these universities or an indication of changing social trends since scientific research on this topic began thirty years ago. Finally, due to the fact that sexual assault is illegal, socially stigmatised, and largely takes place in private settings, it is nearly impossible to observe directly, requiring us to rely on students’ narratives of their own sexual experiences and sexual assaults, though we were able to observe interventions in public spaces like parties and bars (Hirsch et al., 2018; Jerolmack & Khan, 2014).

Students are stigmatizing sexual assault perpetration. However, designating men ‘rapey’ has also become a means of demonstrating social prestige (Pascoe & Hollander, 2016). In the field, we saw it utilised more to reproduce and justify social prestige hierarchies than to disrupt predatory behavior. Our findings suggest that the stigma attached to sexual assault perpetration has created social conditions that encourage men to minimise their friends’ and their own sexually aggressive behaviors while publicly shaming more socially vulnerable men with whom they have few social connections. This is not to say that sexual assault perpetration should not be socially stigmatised; rather, our point is that we must provide students with accurate facts about who perpetrates sexual assault, even—or especially—if those who commit assault are trusted, prestigious men. Research has long demonstrated that most perpetrators of sexual assault have intimate relationships with their victims (Koss, 1988; Tjaden & Thoennes, 1998). Student trainings need to explicitly confront the myth of the mysterious outsider and the campus serial-rapist (Swartout, et al., 2015), and acknowledge both the fact and the consequences of the fact that many perpetrators will be members of one’s ingroup. Practically, this means that trainings should address how to intervene among one’s friends, how to acknowledge that one’s friends’ consent practices might be suboptimal, and how to influence friends to change their behavior outside of the moment of directly intervening in an assault. Indeed, one potential explanation for our finding that students uniformly viewed sexual assault to be a morally reprehensible act that they wanted to prevent in light of the persistently high rates of sexual assault on campus (Mellins et al., 2017) would be that at least some of those who commit assault do not understand that they are engaging in non-consensual sexual behavior. This requires fostering a campus community in which such students can—hopefully before having sex—learn better consent practices; this includes assisting peers to help their friends.

Although existing bystander intervention trainings call attention to the potential for anyone – regardless of gender – to perpetrate sexual assault, one key take-away is that more needs to be done to disrupt the gendered, heterosexual frames within which students understand sexual assault and bystander intervention. As others have noted (McCaughey & Cermele, 2017; Messner, 2016), bystander interventions can reinforce gendered expectations of women’s physical vulnerability and men’s potential to both act against and for women violently. Although others have suggested training women in self-defense as a better means of dismantling these stereotypes (McCaughey & Cermele, 2017), we are skeptical in the efficacy of these measures given that alcohol use is an important part of the context in which many college students are sexually active as well as given the complex temporal and social nature by which students make sense of sexual consent (Hirsch et al., n.d.). However, prevention trainings and conversations about sexual assault with university students need to more intentionally disrupt students’ gendered notions about who is a potential victims and who should behave as a bystander. To this effect, we recommend that universities should provide all students (not just women as McCaughey and Cermele 2017 recommend) with sexual assault prevention training that emphasises both potential self-defense and bystander-intervention techniques; includes role playing of potential situations to target explicitly disrupt gendered expectations; and teaches physical and verbal techniques (e.g. traditional self-defense maneuvers but also strategies to deal with the awkwardness of extricating oneself from an intimate situation, etc.) that could be used to prevent an assault.

Bystander-intervention training must also acknowledge that bystanders, even effective ones, can also be perpetrators of sexual violence. That requires encouraging students to evaluate bystanders critically and to acknowledge that bystanders might have their own motivations for intervening and might assess a victim’s sobriety or interest in ways that can still put someone at risk.

Finally, our findings demonstrate that campus sexual assault and the necessary preventative measures to stop it occur within culturally-specific settings vis-à-vis gender, alcohol consumption, and campus social life. This underlines the need for comparative research in diverse locations within the United States and internationally as well as at a broader range of institutions (non-residential, two-year, et cetera) in order to identify context-specific prevention opportunities. In addition, the mapping of masculine prestige onto sexual assault prevention merits more research and consideration in designing studies that examine campus sexual assault particularly as relates to ‘commonsense’ knowledge about ‘risky’ populations.