What a New Survey Can — and Can’t — Tell Us About Campus Sexual Assault

By
New students at San Diego State University watch a video on sexual consent during an orientation meeting Friday, Aug. 1, 2014.
New students at San Diego State University watch a video on sexual consent during an orientation meeting on Friday, August 1, 2014.Photo: Gregory Bull/AP/Corbis

Sexual assault is an inherently difficult thing to address, and it’s only made trickier by the extent to which it eludes any sort of straightforward social-scientific analysis: It turns out it’s really, really hard to measure the frequency of sexual assault in any given setting. All the usual problems familiar to survey researchers and pollsters — among them the ease with which different people can interpret the same question differently, and people’s natural inclination to hide things they feel ashamed of or traumatized by — are compounded when the subject is rape.

That can partially explain why there’s such huge variation in estimates of how often rape occurs. As Dana Goldstein explained in a very helpful Marshall Project primer last year, two different approaches to estimating the rate of rape on campus, both conducted by professionally trained researchers, led to results — 0.6 percent versus the famous “one in five” figure often cited in these discussions — separated by a chasm. “The two studies are based on surveys that asked different questions of different populations under different circumstances,” explained Goldstein. In other words, every detail of how a survey is conducted matters — a lot — and it’s important not to get hung up on any one number based on a limited sample.

All of which is worth keeping in mind in light of a new survey published in the Journal of Adolescent Health. The authors, led by Dr. Kate Carey of Brown University and Sarah Durney (who worked on the survey during her time as an undergraduate at Brown), attempted to estimate the prevalence of sexual assault, both completed and attempted, at an upstate New York university. It’s important to note that it surveyed respondents repeatedly over the course of a year.

The respondents, recruited through “a mailing sent to all matriculating first-year women, campus flyers, word of mouth, and a research participant pool,” were surveyed both before they started school “and at the end of the fall, spring, and summer” of their freshman years, earning $20 for the first survey and $10 for each subsequent one. To help reduce the potential for bias in the respondent pool, the researchers presented the survey as being about health in general, not sexual assault. Within a series of more general health questions, they included a 20-item version of the Sexual Experiences Survey, an established tool for measuring sexual assault prevalence that asks respondents how often in a given span of time they’ve experienced various forms of assault and other unwanted sexual advances, either completed or attempted.

To help address some of the methodological difficulties they faced — particularly the aforementioned issue with question interpretation — the researchers broke out their results into multiple categories of rape and were careful to separate out victims who were “only” groped from those who were raped, or who escaped an attempted rape. “We decided, because the literature has been so varied in terms of how it’s characterized … to focus in on two types of rape: forcible and incapacitated,” Carey told Science of Us. As the names suggest, incapacitated rape, or IR, involves instances in which “alcohol or other drugs [were] used,” while forcible rape, or FR, deals with rapes in which the victim was physically overpowered.

Twenty-eight percent of the women in the sample had already experienced either a rape or an attempted rape (either type) by the time they started school. Over the course of their freshman year, one in six reported an attempted or completed rape, so that by the start of the women’s sophomore year, the lifetime incidence rate had jumped all the way up to 37.1 percent (given the questions’ phrasing, assaults occurring away from school, while home on break, for example, were likely counted).

Incapacitated rape was more common than than forcible rape, which the researchers say is “[c]onsistent with previous research” on college students. Women who had reported “a precollege history of sexual assault, particularly [attempted or completed] IR,” were more likely to reported being raped in their first year. (The table running down the results at each point is too wide to embed but can be viewed here.)

Suffice it to say these are depressingly high figures. But Christopher Krebs, a researcher at RTI International who wasn’t involved in Carey’s study, said in an email that it’s important to put them in context — not because the numbers aren’t alarming but because they reflect the results of only one school polled in one instance using one methodology.

He said that in the wake of the release of this sort of data, there tends to be a fair bit of overextrapolation — people want the numbers, already shocking, to mean even more than they do, and want them to signify something at the national level. Krebs knows this firsthand, having led the famous “one in five” surveys that were never intended to serve as national-level statistics — he and his colleagues only surveyed students on two campuses — and which nonetheless led to President Obama and Vice-President Biden stating that one in five women nationally will be sexually assaulted during their time on campus.

This might sound like nitpicking — obviously no survey method is perfect, and obviously each contributes a little bit of information that, in the aggregate, can add to our understanding of the issue. But Krebs said that the focus on national-level statistics misses the mark in important ways. “I don’t see much value in estimates that might represent some sort of average across schools, because although they might be good for headlines or speeches, they don’t help administrators at a given school understand or solve the problem,” he said. “It is sort of like having a national crime rate. If I am a citizen or a police chief in Mayberry, what does a national crime rate do for me? Not much.” Krebs thinks that there are probably significant differences in the sexual assault rates at different campuses that are worth exploring — data he and colleagues are currently putting together and plan on releasing by the end of the year.

As for studies like this one, he said that as long as the public understands their limitations, they can play a very important role. “To me, what are needed are good school-specific data, and results that can be used to understand the magnitude and nature of a problem at a school so people can use the information to do a better job of prevention victimization, providing services to survivors, investigating incidents, and holding perpetrators accountable,” he said.

Carey, too, emphasized the importance of digging down into the grim details of a given campus’s or community’s rape problem. She said that emerging research suggests that certain characteristics — being a freshman, for example, or being a member of a sorority — are linked with a higher chance of being sexually assaulted, and that better understanding why these links exist could help researchers develop targeted sexual-assault-prevention interventions. “Prevention of something as culturally and behaviorally complex as sexually assault really needs to be multi-pronged,” she said. “There’s not one Band-Aid that will fix it.”

As for her survey’s estimates, Carey said that they’re in line with other numbers that have been produced by similar efforts — she seemed less concerned about the idea of thinking in terms of wider averages. “The estimates that have been generated over recent years have been remarkably consistent,” she said, “if not in the precise numbers, then in the general proportion of the population.” Her numbers — again, one out of every six of the students experienced an assault or an attempted assault — certainly aren’t out of line with the idea that (approximately) one in five women will be assaulted during her time on campus.

Even if generating reliable numbers remains a difficult task, and even if any one attempt brings with it inherent limitations, Carey said that she hoped her survey and others like it would contribute to a “chipping away at the social norms of acceptance” of sexual assault. “I think that we do need to change the notion that, Oh, this just happens,” she said.