Surveying rape

The social life of methods

What is the perspective broadly known as the social life of methods? As Savage (2013) has noted, this is (in broad strokes) the position that methods are not merely the technical or instrumental means of testing or measuring something, but are rather objects of inquiry in and of themselves. More powerfully, when we look at methods as objects of inquiry they can tell us about the ‘affordances and capacities which are mobilized in and through the methods themselves’ (2013: 5). This perspective is consistent with viewing methods as performative – as articulated by sociologist John Law:

There are two great views of method in science and social science. On the one hand it is usual to say that methods are techniques for describing reality. Alternatively it is possible to say that they are practices that do not simply describe realities but actually tend to enact these into being. (Law, 2009: 239)

The public discourses about rape and sexual assault that have been fueled by these surveys are illuminating on many levels. Historically, changing definitions of rape itself have signaled shifts in gendered and racialized conceptions of citizenship, and hence political power. As Estelle Freedman has argued in her political history of rape from the 1800s through the mid-20th century, ‘Changing the definition and prosecution of rape has challenged the very meaning of citizenship in American history’ (Freedman, 2013: 2). She demonstrates that rape’s history has been intertwined with the question of who was deemed worthy of privileges and obligations such as voting, jury duty, office-holding and access to due process. To a large extent, the politics Freedman discusses have been driven by constructions of race and gender that have reinforced white male sexual privilege and legal authority by enforcing damaging stereotypes about black men, black women and to some extent white women (for histories that deal with race, gender and rape see Feimster, 2009; McGuire, 2010; Rosen, 2009).

While these constructions have persisted well into the late 20th and 21st centuries, in this article I focus on the definitional and empirical claims made about rape and its prevalence by white feminist social scientists who elaborated the concepts of ‘hidden’, ‘unacknowledged’, ‘date’ and ‘acquaintance’ rape in the 1980s, largely in the context of – predominantly white – college campuses. This story threads alongside the more overtly racialized history of rape and sexual assault, and thus begs the question of how these statistics and surveys have served, to some extent, to efface race from contemporary public conversations about campus rape. ⁷

In typical accounts focusing on reactions to Mary Koss and her colleagues’ work in the mid-to-late 1980s, the empirical efforts of feminist social scientists to uncover the prevalence of campus sexual assault ran afoul of traditional gender ideologies and the rising tide of conservative politics and anti-feminist backlash of the late 1980s and 1990s (see, for example, Neame, 2004; Raphael, 2013; Russell, 2000; Schoenberg and Roe, 1993). Attempts by these scientists to uncover the ‘true’ scope of the problem, and to demonstrate that most rapes were committed, not by strangers, but by men known to the victims, ⁸ led to intense outcry from critics who saw this work as both methodologically flawed and strongly politically motivated, even ideological. I begin with a brief overview of the emergence of the 1970s anti-rape movement to contextualize the empirical work of these social scientists and subsequent reactions to it. ⁹ It was in the context of this movement that social scientists began to view rape and sexual violence as objects of scientific inquiry and, more specifically, as phenomena that could be surveyed and measured. ¹⁰

The 1970s anti-rape movement and feminist social science: Rape emerges as a research topic

In 1977, historian Edward Shorter reviewed journalist Susan Brownmiller’s now-classic 1975 work, Against Our Will: Men, Women, and Rape. In his review, Shorter critiqued Brownmiller’s thesis that throughout history rape has functioned as a form of male political and cultural domination. Charging Brownmiller with misunderstanding the nature of rape in times past, he argued that in the early-modern era, rape could be better explained as the result of the mass experience of ‘misère sexuelle’ on the part of men denied outlets for their sexual urges in a society that feared eroticism and zealously policed extra-marital sex (Shorter, 1977; for a sharply critical feminist reply to Shorter, see Hartmann and Ross, 1978). It was in this context that Shorter noted the absence of serious scholarship on rape, commenting in his review that rape was a topic ‘hitherto as well known to conventional scholars as the dark side of the moon’ (Shorter, 1977: 471). The ignorance of conventional scholars mirrored the lack of public discussion about the prevalence and nature of many forms of violence against women at that time. As historian Elizabeth Pleck has noted, ‘there were virtually no public discussions of wife-beating from the turn of the century until the mid-1970s’ (Pleck, 2004: 182). This silence reinforced the perception that rape, incest and domestic violence rarely happened and that when they did, they were perpetrated by a few sexual deviants (see also Breines and Gordon, 1983). This perception underwent a radical challenge in 1971 when Susan Griffin published ‘Rape: The All-American Crime’ in Ramparts Magazine, in which she highlighted the prevalence of rape and the links between male sexuality, violence and power in American culture, as well as the racism endemic to the history of rape in the United States (Griffin, 1971).

Also in 1971, the New York Radical Feminists held what has been heralded as the first public event in the United States at which women spoke out about their experiences of being raped. The speak-out was held in a small Episcopal church in midtown Manhattan. The church filled with 300 women, 40 of whom spoke about their experiences of being raped, both at the hands of the rapist and then again in their interactions with the criminal justice system. A few months later, a follow-up conference was held which included another speak-out, self-defense workshops and research presentations. ¹¹ This time, the New York Times sent a reporter. Particular concern again focused on the way the legal system was biased in favor of those accused of rape, and much ensuing feminist activism centered on challenging and changing rape laws to help protect survivors of rape from further victimization at the hands of the police and the courts.

In this newly politicized arena, both outside and within academia, feminist social scientists began to take up rape as a research topic. As feminist psychologist Nicola Gavey has observed, ‘Before feminism called public attention to the issue of rape as a social problem, the social sciences had largely been quiet about the subject’ (Gavey, 2005: 28). One of the early academic voices to break this silence was Mills College sociologist Diana Russell. South African by birth, Russell was active in the African Resistance Movement, then moved to the United States to pursue graduate work at Harvard, where she focused on the sociology of revolution. ¹² A few years after graduating and taking up her faculty position at Mills College, she published The Politics of Rape: The Victim’s Perspective, a ground-breaking and best-selling work based on interviews with over 90 rape survivors and 4 convicted rapists (Russell, 1975). In her book, Russell utilized a prominent strategy of the anti-rape movement: she let the voices of victims/survivors speak for themselves. Most of the 25 or so chapters are dominated by verbatim transcripts of 22 interviews Russell selected from her larger sample. Echoing other feminist writers of the times, she showed that rape must be understood as a manifestation of a culture of male privilege rather than as a deviant act perpetuated by a few sexual psychopaths: nothing less than traditional masculinity under patriarchy was the cause of rape.

Reactions to Russell’s book, which was one among a small handful of books on rape published around the same time, were modest and largely neutral or positive in tone. ¹³ Her strategy of presenting the victims’ own stories in detail, however, would later be referred to (and discounted) by a prominent critic as the deployment of ‘atrocity tales’ (Gilbert, 1997: 123). Russell noted that after the publication of her book, she became intrigued by the observation that while most feminists agreed that rape was a common crime against women, most non-feminists believed it was relatively rare. Echoing this observation in a 1980 article in the journal Signs, sociologist Allan Johnson prefaced his analysis of the prevalence of rape in the United States with the following assessment:

Many who have engaged in the struggle against sexual violence believe such violence is widespread; I think a more common perception in the general population is that, while the individual rape is a horrible experience for the victim, such events are statistically rare and therefore not reflective of everyday life in American society. (Johnson, 1980: 137)

The emergence of the rape survey

This need for ‘accurate’ statistics on rape next impelled Russell to prepare a grant proposal to the National Institute of Mental Health (NIMH) to undertake a large-scale survey of the prevalence of rape and other forms of sexual violence in a probability sample of 930 women who were 18 years and older in the San Francisco area. For this study, she carefully trained 33 female interviewers of various ethnicities to individually administer the surveys. Acutely aware of the stigma surrounding the reporting of sexual abuse and rape and its influence on under-reporting, Russell stipulated that interviewers receive 10 hours of education about rape and incest. She also ensured that the wording of the interview protocol minimized or eliminated victim-blaming. The survey used the legal definition in California at the time of the study to assess whether respondents had experienced completed and/or attempted rape: forced intercourse, or intercourse obtained by threat of force, or intercourse completed when the woman was drugged, unconscious, asleep, or otherwise totally helpless and hence unable to consent. At the time, although marital rape had been articulated by social scientists and others, the legal definition in California excluded rape within marriage. Following this protocol, Russell reported that 19% of her sample had experienced a completed non-marital rape, and 31% had experienced attempted rape. Combined, the rate for attempted and completed rape stood at 41% (Russell, 1983).

Russell took great care to explain the steps she had taken to overcome the problem of under-reporting and to make her methods transparent. Revisiting Johnson’s methods and findings, she and a colleague wrote a detailed article in which they used their own data, subjected them to the same life-table analysis that Johnson had used, and concluded that his 20–30% lifetime probability was in fact conservative; their lifetime probability estimates stood at 26% for completed rape and 46% for completed and attempted rape combined (Russell and Howell, 1983). Aware that readers might find these statistics implausible, Russell noted that previous estimates provided by National Crime Surveys were artificially low due to a questionnaire design that was insensitive to the phenomenon of under-reporting. In fact, the only way respondents could report an experience of sexual assault on these surveys was if they indicated ‘Yes’ to the general question assessing criminal victimization ‘Did anyone try to attack you in some other way?’ Otherwise, this line of questioning was abandoned, along with any opportunity to report completed or attempted rape.

Russell’s work, along with several other feminist analyses at that time, conveyed at least two central points: (1) that the rates of reported rape from previous federal surveys in the United States far underestimated actual rapes; and (2) that most rapes were committed by acquaintances of the victims, not by strangers. Both of these points would impel the birth of the campus sexual assault survey. In 1976, psychologist Mary Koss became intrigued by the difficulty of gathering accurate statistics about sexual assault. As she later recalled:

I had recently read journalist Susan Brownmiller’s book, Against Our Will: Men, Women and Rape. Using what few data were available at the time, she carefully documented that the scope of rape was underestimated. The challenge of finding these unmeasured rapes pulled me in…(Koss, 2011: 349)

Koss administered this initial survey to 3,862 students at Kent State University. One of her most startling findings was not, in fact, the absolute prevalence of rape on campus – other work, such as Russell’s, had already signaled this. What was more intriguing to Koss, and what became controversial in ensuing debates, was that over half of her female respondents who endorsed the behaviorally specific items reflecting the legal definition of rape, such as ‘Have you had sexual intercourse with a man when you didn’t want to because he used some degree of physical force?’, did not simultaneously endorse the item ‘Have you ever been raped?’ The likelihood of this disconnect in reporting was especially true if the rape was perpetrated by an acquaintance, as opposed to a stranger. As Koss reflected later, ‘The conclusion that women could report experiencing behaviors constituting rape but not perceive themselves as raped was a light-bulb moment revealing an alternative interpretation of what initially looked like a dismaying measurement problem’ (Koss, 2011: 350). ¹⁵ This last item, ‘Have you ever been raped?’, she reported, was initially included only as a validity check. Its inclusion made visible the disconnect between experiencing certain acts and labeling them rape. This ‘dismaying measurement problem’ in fact became crucial in realizing date/acquaintance rape, but was also invoked to try to destabilize its ontological status, and thus its political reality.

Based on these results, Koss identified what she originally called the ‘hidden rape victim’, a woman who endorses having experienced the behavioral (and legal) definition of rape but who does not report it and who may not describe herself as a victim of rape (Koss, 1985). These were the hidden rape victims, the unreported rapes, that Koss was trying to find a way to measure. In response to this light-bulb moment, Koss then refined her measurement tool to include only the behaviorally specific items, but adding a more nuanced four-item ‘perception of victimization’ scale that asked directly about the degree to which the respondent perceived that she had been a victim of a crime, including rape. This scale included the following four items: ‘I don’t feel I was victimized’; ‘I believe I was a victim of serious miscommunication’; ‘I believe I was a victim of a crime other than rape’; and ‘I believe I was a victim of rape’. She applied for an NIMH grant and designed an expanded study that would use the new measure with a nationally representative sample of colleges and universities. Ms magazine became interested in her work and offered to provide office support and help disseminate the findings.

With NIMH support and Ms magazine sponsorship, ¹⁶ Koss then administered the revised survey to over 6,000 male and female university students across 32 institutions in the United States. Subsequently dubbed the ‘Ms magazine survey’, Koss’s study quickly became the standard reference work on the prevalence of sexual assault and Koss was frequently consulted and cited as a scientific expert on the topic (see, for example, Sherman, 1985). In October 1985, Ms magazine published preliminary results from the study in an article entitled ‘Date Rape: The Story of an Epidemic and Those Who Deny It’ (Sweet, 1985). In 1987, Koss and her colleagues published a formal report of their national results in the Journal of Consulting and Clinical Psychology. Their conclusion: 27% of women reported having experienced attempted or completed rape, the vast majority perpetrated by a date or acquaintance (Koss, Gidycz and Wisniewski, 1987). This ‘1 in 4’ statistic spread like wildfire, sometimes accurately characterized as referring to attempted and completed rape, but more often distorted to ‘1 in 4 women have been raped’. Koss found her work picked up by community activists, politicians, self-proclaimed ‘equity’ feminists and numerous media outlets, resulting in what she referred to as ‘a fruitful, although at times a painful, public dialogue’ (Koss, 2011: 352). Painful in part because Koss, and to some extent Russell, became the target of attacks by the conservative media as well as other academics. Why?

In 1988, Ms magazine published a book based on findings from the study, entitled I Never Called it Rape which focused on raising awareness about the nature and prevalence of date and acquaintance rape, and what to do about it (Warshaw, 1988). The book (and its title) reinforced the idea that women were experiencing rape and attempted rape but not labeling it as such, and subsequently not reporting it. In the same year, Koss and her colleagues published another journal article in which they further explored the concept of the ‘hidden’ rape victim. Using the perception of victimization scale, they compared the responses of victims of stranger rape and victims of acquaintance rape. They found that among the sub-sample of female survey respondents that endorsed the behaviorally specific definition of rape, only 55% of the stranger rape victims acknowledged their experience as rape, and an even smaller percentage – only 23% – of acquaintance rape victims did so (Koss, Dinero et al., 1988). This was despite the fact that on the ‘perception of victimization’ scale outlined above, almost all of the rape victims perceived themselves as having been victimized to some degree, even if it was short of directly labeling their experience as rape. Critics charged that Koss and other feminist researchers were overruling women’s own perceptions and engaging in ‘politically-motivated data distortion’ (Koss, 2011: 350). More damningly, some accused them of creating a phenomenon that simply did not exist. The majority of date rapes, they argued, were artifacts of the rape survey itself as cleverly deployed by radical feminists masquerading as objective scientists.

Realizing contested realities

In 1990, Columbia University journalism student Stephanie Gutmann, writing for Playboy magazine, asked: ‘Date Rape: Does Anyone Really Know What It Is?’ In her article she accused researchers of making the definition of rape so elastic as to include incidents where women had simply regretted having sex (Gutmann, 1990). In 1991, Berkeley social welfare policy professor Neil Gilbert published ‘The Phantom Epidemic of Sexual Assault’ in The Public Interest – a prominent neo-conservative quarterly – that was subsequently picked up by the Wall Street Journal and eventually incorporated into the conservative feminist writings of Katie Roiphe and Christina Hoff Sommers (see Roiphe, 1993; Hoff Summers, 1994).

Disparagingly referring to Koss as an ‘advocacy researcher’, Gilbert argued that the apparent epidemic of sexual assault revealed by Koss and Russell was due to something more pernicious than simple definitional quarrels among academics. He charged that it was a full-scale attempt by radical feminists to change how we view intimate relations between the sexes. As he put it:

…the function of advocacy numbers is to alter consciousness more than raise it, to change social perceptions of what constitutes common experience in heterosexual relations. The difference between a sexual assault rate of 25 or 50 percent and one of .1% is more than a matter of degree…To argue for the higher rate is to try to shift our understanding of the battle of the sexes; the model suggested by Spencer Tracy and Katharine Hepburn is to be replaced by one in which Conan the Barbarian violently thrashes his cavemate. (Gilbert, 1991: 64–5)

In 1993, Katie Roiphe, then a graduate student in English at Princeton, picked up Koss’s work through Gilbert and wrote an article for the New York Times Magazine that came out in advance of her book The Morning After: Sex, Fear, and Feminism On Campus. In this article, she wrote:

Is there a rape crisis on campus? Gilbert points out that in a 1985 study undertaken by Ms magazine and financed by the National Institute of Mental Health, 73 percent of the women categorized as rape victims did not initially define their experience as rape; it was Mary Koss, the psychologist conducting the study, who did. (Roiphe, 1993: 27)

In 1994, philosophy professor Christina Hoff Sommers, branding herself a mainstream ‘equity feminist’ in contrast to the gynocentric ‘gender feminists’ she critiqued, published Who Stole Feminism: How Women Have Betrayed Women. ¹⁸ In this wide-ranging book, she took aim at the gender feminists’ portrayal of multiple issues, including a chapter devoted to rape research and the purported crisis it had created. As Gilbert and Roiphe had done previously, she linked this research to the anti-male agenda of angry feminists who preferred to rail against patriarchy rather than work towards the more moderate and desirable goal of achieving equality with men. She took extensive issue with Koss’s work and the ‘1 in 4’ statistic, both quibbling with the wording of one of Koss’s survey questions, and problematizing the finding that the majority of women who were classified as having been raped did not answer ‘yes’ to the direct question about rape. Contending that politics had trumped science, she argued that rape-hype feminists were heavily invested in high rape prevalence rates to further their claims about the sexism and misogyny of American culture, and to direct large sums of money to campuses that served young women with socio-economic standings and backgrounds much like their own. This, she said, directed money away from communities where services for rape victims were more urgently needed. She concluded: ‘Gender feminist ideologues bemuse and alarm the public with inflated statistics’ (Hoff Sommers, 1994: 225).

Thus, between the mid-1970s and the early 1990s, rape went from a topic ‘hitherto as well known to conventional scholars as the dark side of the moon’ to a phenomenon that feminist social scientists, through their surveys, were implicated in actually creating. Date and acquaintance rape, specifically, came into focus through the use of surveys that were designed to address the problem of under-reporting in the context of an emerging feminist social science that took up rape as a topic of scientific study. While Russell’s and Koss’s work was an outgrowth of the 1970s anti-rape movement, their results came to the fore as the radical social movements of this earlier era paradoxically provoked, as historian Robert Self has put it, ‘the rise of a powerful new politics on the right: breadwinner conservatism’ (Self, 2012: 5). This new politics, which took as its primary agenda the defense of the nuclear family against incursions by feminists and gay rights activists, ushered in the Reagan era of the 1980s. It is notable that Gilbert, one of Koss’s primary critics, invoked her work as a threat to the very structure of traditional norms and relations between the sexes, thus aligning himself with those conservatives who sought to defend the traditional family against attacks by ‘radical’ feminists and other enemies of traditional gender roles.

This first wave of controversy over the sexual assault ‘epidemic’ came to a peak in the early 1990s and informed the political debates leading up to the passage of the Violence Against Women Act (VAWA) in 1994, for which Koss herself gave Senate testimony. The National Center for the Prevention and Control of Rape had been established in 1975. In the 1980s it was recommended for de-funding by Reagan and eventually phased out by 1987. By the early 1990s, Joseph Biden and others were formulating the VAWA, and the specter of increasing appropriations to rape crisis centers and rape prevention and education programs no doubt unsettled neo-conservative welfare policy proponents. As Koss remarked in her 1992 response to Gilbert, ‘Defending Date Rape’:

Their [Gilbert’s and others’] assertions cannot be ignored because they support a clear-cut agenda: to reduce public support for the appropriations directed at rape crisis services and education/prevention as proposed in the Violence Against Women Act. This legislation was unanimously voted out of the Judiciary Committee in July 1991 and is scheduled to be placed before the full senate late in the year. (Koss, 1992: 122)

The contemporary social life of the sexual assault survey

Conservative critics in the early 1990s charged that, in the hands of feminist scientists, sexual assault surveys had literally created an experience, and a resulting epidemic, that did not exist. Despite their critique, survey results were highly successful in raising public awareness and supporting policy change. In the first decades of the 21st century, surveys have once again emerged as active and controversial participants in public debates about campus sexual assault.

In 2007, the United States Department of Justice, through the National Institute of Justice, commissioned the Campus Sexual Assault (CSA) Study (Krebs et al., 2007). Researchers randomly sampled 5,446 undergraduate women from two large public universities in the south and the Midwest. Participants completed a web-based survey of their experience of attempted and completed sexual assault up to and during their time in college. These results were similar to Koss’s original findings, with 28.5% of the sample reporting having experienced attempted or completed sexual assault to that time. ¹⁹ The percentage was about 19% when only the time since college was considered – hence the ‘1 in 5’ rather than ‘1 in 4’ statistic that is now more commonly reported.

Reactions to this study echoed the responses to Koss’s work 20 years earlier, and focused pointedly on the survey methods. Critics and commentators in the popular press argued that the survey had overestimated sexual assault and that the definitions of assault it used were too elastic (e.g. Contorno, 2014). This time, critics also claimed that a low response rate and the fact that the study was based on samples from only two universities decreased its credibility and representativeness. Washington Post fact-checker Glenn Kessler gave it a single Pinocchio rating, meaning the statistic should be considered unreliable pending further study (Kessler, 2014).

In the wake of another round of press about campus sexual assault, a New York Times feature asked whether colleges and universities were doing enough to prevent rape ( New York Times, 2014). Invoking the ‘1 in 5’ statistic, one invited commentator, a sociologist, highlighted the need for even more complete and unbiased statistics about sexual assault on college campuses, highlighting how current reporting requirements place too much power in the hands of university and college administrators who are motivated to minimize these numbers for public relations purposes (Dirks, 2014). Notably, much of the reader reaction to her call for more surveys and statistics expressed the deep skepticism with which respondents viewed survey methodology, along with the usual mistrust of women’s (not universities’) reporting practices: ‘the survey producing that data is not disciplined or scientific’; ‘The often misquoted “one-in-five” factoid clashes with Clery Report statistics because it is based on a survey that so broadly defined sexual assault that the majority of coeds counted as sexual assault victims disagreed they had been sexually assaulted’; ‘Surveys are a cheap but unreliable way to gather data (not information). All surveys should be validated, but they seldom are. It is a false promise to think surveys will provide transparency. If we ask women if they have ever been victimized by men, they will answer yes.’ ²⁰

In response to continued skepticism about existing surveys and their results, another survey was undertaken in 2015 by the Washington Post–Kaiser Family Foundation. It employed a random sample of 1,053 men and women who had been undergraduates at four-year colleges since 2011. This survey confirmed the ‘1 in 5’ statistic for women and reported a 5% sexual assault rate for men. ²¹ Although this confirmation appears to have assuaged some doubts (see, for example, Jaschik, 2015), online reader responses indicate ongoing preoccupation with survey practices: ‘Nothing has been validated. Look at the survey itself. First off, this was sent as an email to students with only a 19% response rate…This survey is bogus and has lots of flaws’; ‘Great, the bogus 2007 study was validated by a brand new bogus study, using the same bogus criteria like “nonconsentual [sic] touching”, where someone brushes by you on campus. It’s amazing how far these feminists will go to grossly inflate a problem to meet their agenda’; and

Simply taking the various types of offense and adding up ‘yes’ values is dishonest and inflationary. A single incident could trigger multiple ‘yes’ answers on this survey, and rather than treat it as one incident, it’s being statistically treated as separate events. That 20–25% value for all offenses combined in most schools is statistically useless as a metric, as a result. It’s guaranteed to be high. ²²

However, as I contend here, the gathering of statistics about rape is an activity that simultaneously structures the phenomenon it is measuring and is enacted within and through the society that is being ‘made up’ by these numbers. ²³ Diana Russell’s and Mary Koss’s efforts to define and estimate the scope of sexual assault both arose out of and produced versions of the social and political order that were, and are, undergoing continued contestation.

Despite this ongoing contestation, 2013 was a watershed year for political mobilization against sexual assault on college campuses. In January of 2013, five students and a former dean at the University of North Carolina Chapel Hill filed Title IX complaints against their university for not appropriately handling sexual assault charges. In February, a University of Southern California student went public with her rape experience and was joined by other rape survivors to found Student Coalition Against Rape. This was followed by the formation of a network called ‘Know Your IX’ to link rape survivors from more than 50 colleges seeking justice from the federal government. In September of 2014 a Columbia University performing art student appeared on the cover of New York magazine because of her ‘Carry that weight/Mattress performance’ in which she vowed to drag her dorm room mattress around campus until the school expeled her alleged rapist. The article described her political statement as part of ‘the most effective, organized anti-rape movement since the late ’70s’ (Grigoriadis, 2014).

In March of 2013, US President Barack Obama reauthorized the Violence Against Women Act, which included a new amendment: the Campus Sexual Violence Elimination (SaVE) Act. This amendment updates the Clery Act, in the original VAWA, that mandated every college and university in the United States receiving federal money for financial aid to gather and report crime statistics, including statistics about sexual assault. The Campus SaVE Act actually goes beyond this by requiring colleges and universities not only to collect and report better statistics on campus sexual assault, but to guarantee victims enhanced rights, provide for standards in institutional conduct proceedings, and provide campus-wide sexual assault prevention education. Both President Obama and Vice President Biden, who was instrumental in the drafting and passage of the original 1994 act, have repeatedly invoked the ‘1 in 5’ statistic as a justification for these, and other, policy changes (e.g. ‘Remarks by the President and Vice-President’, 2014).

Rape surveys as complex assemblages

Koss’s surveys, and those that have succeeded them, no matter how technically sophisticated, have not provided unmediated access to ‘the truth’ about rape; rather, they have materialized and rematerialized – via numbers and statistics – experiences that had been individual, private, unarticulated and – before the 1980s – unmeasured, and they did so within a particular social milieu that has itself shaped the ways these statistics have circulated. As feminist historian Michelle Murphy has noted in her work on stress and health surveys in the making of sick-building syndrome, surveys function not merely as information-gathering technologies, but as constitutive sites producing new ways of structuring and communicating experience (Murphy, 2006). More than simply providing a set of numbers about the prevalence of campus sexual assault, Koss’s survey methodology actually structured the experience of date and acquaintance rape. By intentionally separating out behaviors from the labeling of those behaviors, the Sexual Experiences Survey actually made legible previously unlabeled, but nonetheless identifiable, incidences of sexual assault. Quantification then made these experiences visible to bureaucrats, policy-makers and the media and subsequently leveraged the ‘1 in 5’ statistic to a prominent position in American political and cultural life.

The ongoing contestation of the statistics originally generated by feminist researchers such as Russell and Koss stands as a compelling demonstration that surveys are never neutral, apolitical ‘mirrors’ of a reality that exists independently of their creation. They are constructed, acquire meaning and are deployed in and through a multifaceted scientific and socio-political milieu. They are, to borrow Ken Alder’s Geertzian nomenclature, ‘thick things’ – complex assemblages of ethical and political prescriptions that mobilize and coordinate the diverse sets of actors who create, use and even oppose them (Alder, 2007). The fact that statistics and surveys are ‘thick things’ does not invalidate the numbers they produce, but it does invite scrutiny into how they function in the creation of contested realities.

That Russell, Koss and other feminist scientists often opted to defend their so-called ‘advocacy’ research by reiterating the scientific soundness of their methods reflects a long-standing commitment to an empiricist philosophy that has generally served them, and other scientists, well. ²⁴ Their work also arose out of and contributed to unprecedented public and political support for multi-faceted campaigns to prevent rape and other forms of violence against women, despite the ongoing debates about rape’s ‘true’ prevalence and nature. As Koss well observed, however, fighting ideological barriers solely with numbers may side-step the issue of what other strategies may be required to bring about large-scale shifts in beliefs about what constitutes and causes rape. As Law, Ruppert and Savage have pointed out, feminists (among others) have long sought to ‘simultaneously reground expertises and forms of knowledge, versions of the real, their advocates, and the larger ecologies of knowledge and expertise in ways that contest those that subsist most comfortably within the standard arrangements for knowing’ (2011: 13). This regrounding is a task requiring the simultaneous shifting of many different parts of the assemblage, and proceeds messily and haltingly at best.

So to return to the paradox I outlined earlier: sexual assault surveys and the statistics generated by them have been highly effective despite their ongoing contestation. One way to understand this, I would suggest, is to view surveys as performative – that is, as practices that help materialize experiences in new ways, with both personal and political consequences. This claim, of course, has important scientific, ethical and political implications. If surveys are performative, how can survey researchers actually claim that they have described a ‘real’ state of affairs – what is actually the case? This is especially important if their numbers and surveys are to be used to inform policy and even law.

Gilbert’s argument that Koss’s survey did more to alter consciousness than to raise it is a serious – and telling – one. I would argue that altering consciousness is exactly what viewing methods as performative entails. But rather than seeing this as a damning critique of this enterprise, as Gilbert did, I see this as offering an opening for important critical reflection – not on whether date rape is real – but on the politics of practice in the social sciences. It provides an opportunity for examining how social scientific methods – and the scientists who use them – participate in creating, changing and expanding the conditions of possibility in which certain experiences become ‘real’ and have political and social valence, while other experiences do not. What are the consequences thereof, and are these desirable? For whom, and in what ways? This is not to presume that there will be straightforward answers. To borrow loosely again from Mol (1999: 83), engaging with ontological politics entails ‘tolerating open-endedness, facing tragic dilemmas, and living-in-tension’.

Importantly, as John Law has pointed out, to view surveys as performative does not mean that they can materialize anything. As he puts it:

First, they need to be able to create knowledge (theories, data, whatever) that work, that somehow or other hold together, that are convincing and (crucial this) do whatever job is set for them. But then secondly and counterintuitively, they have to be able to generate realities that are fit for that knowledge. (Law, 2009: 240)

The sexual assault survey has performed in myriad ways, mobilizing different actors in the service of trying to enforce multiple versions of ‘the real’. In treating the rape survey not as a transparent measuring instrument, but rather as a practice that performs within a complex assemblage of implicit and explicit beliefs, attitudes, institutions, communities and politics (including, importantly, feminist politics), social scientists can be more deliberative about the social worlds they realize through their methods and, perhaps more importantly, engage more effectively in debates with the critics of these contested realities and their stakeholders. To reiterate, it equips them to engage with the ontological politics of their practices:

Ontological politics is a composite term. It talks of ontology – which in standard philosophical parlance defines what belongs to the real, the conditions of possibility we live with. If the term ‘ontology’ is combined with that of ‘politics’ then this suggests that the conditions of possibility are not given. That reality does not precede the mundane practices in which we interact with it, but is rather shaped within these practices. So the term politics works to underline this active mode, this process of shaping, and the fact that its character is both open and contested. (Mol, 1999: 74–5)