Saturday, March 25, 2017

The UT Austin Campus Rape Crisis

Just some notes for later (still taking a break).

Headline in the Dallas Morning News: "15 percent of female undergraduates at UT have been raped, survey says"

Some quick math puts that at about 3000 rapes on a campus of about 40,000 undergraduates. The official crime stats for the City of Austin says there were under 500 rapes in a population of about 930,000 people.

Fortunately, the Dallas News did a follow-up story on the how the survey was done, which casts some light on things:

In the study, rape was defined as "having oral sex with someone, making someone perform oral sex, or penetrating someone's vagina or anus with penis, fingers or other objects without their consent, by use of verbal pressure, taking advantage of them when they're incapacitated, threatening to harm or using force."

An example of rape in the survey, for instance, would be if a perpetrator pressured someone to perform oral sex, after they'd said they didn't want to, by threatening to end the relationship or threatening to spread rumors about about the victim.

I just don't know what to say. I was made aware of this survey through Kate Clancy's Twitter feed, when she retweeted this. These are the people who are constructing the facts about academic harassment. Enough said.

There's video of the press conference here. And the report can also be downloaded here. It really does say, "Fifteen percent of undergraduate females experienced rape since their enrollment" (p. 18).

* * *

(Update March 26)

The Guardian headline reads "Nearly 15% of female undergraduates at UT Austin report being raped". This is misleading in an important sense that is actually emphasized in the study's methodology.* Respondents were not able to "report being raped"; they were asked to answer a series of "behaviorially specific questions" that the researchers then interpreted. As I point out above, the final report phrases the conclusion a bit more carefully: "Fifteen percent of undergraduate females experienced rape since their enrollment."

As far as I can tell, the consensus view among researchers in this area is not to let the respondents themselves interpret their experiences as a "rape" or otherwise. Rather they elicit descriptions of experiences that they then interpret on their behalf as rape. It would seem that if a woman allows her boyfriend to penetrate her anally under "threat" of his "ending the relationship" then the researchers consider this an instance of "rape".** Indeed, even if she performed oral sex under those conditions, they would consider her "raped". It's hard to know what to do with this other than simply point it out. In any case, it is not true that the CLASE study found that 15% of undergraduate women reported being raped. The survey wasn't set up in that way. Many of the women that the survey counts in its 15% would probably not describe their experience as a rape.

Jenny Miller at New York Magazine gets the lede right: "The University of Texas at Austin is reporting that 15% of its female undergrad students have been raped."

Endnote 1 of the CLASE report: "The terms employed in this study are used in the context of social science research, and not in their legal context. They are not intended to indicate that the responses of results of the survey constitute or evidence a violation of any federal, state, or local law or policy." I take it this means that we are being told that 15% of UT Austin's female undergraduates have been raped but not necessarily in way that breaks the law (or even a UT policy.) If you've been looking for an example how the terminology of the social sciences has been completely detached from reality, I give you a new species of rape: one that doesn't constitute an illegal act.

* * *

(Update March 27)

Imaginary headline: "Mayor Announces that 15% of women aged 18 to 25 in Austin have been raped."
Imagine journalists not following this up with the Chief of Police.

Actual statement in UT Austin President Gregory L. Fenves' message to the UT Community: "[The survey found that] 15 percent of undergraduate women at UT Austin reported that they had been raped."
I have not seen any comment from the chief of the UT Austin PD.

I know that the reason for this is that what the CLASE report means by "rape" has nothing to do with what the police, or anyone else for that matter, means by "rape". My point of the comparison is that we don't let a mayor define behavior that is policed in his city in ways that diverge radically from the way the chief of police defines that behavior. Nor should a university president.

In another post, I'll take up the question of how free social scientists are to define terms for behaviors independently of how they are defined by a society's administrators and enforcers.

The story has reached Newsweek. So far I haven't seen one story that approaches this in a critical way. It's like everyone is writing the same story.

* * *

(Update March 30)

Campus Safety Magazine covers the story. Not much new here, but this is worth noting: "The UT System-wide rape rate among female students was 10 percent." What I noted was that they are talking about "the" rape rate, as if it its something that this study is merely providing a recent measure of. The truth is that it is constructing an entirely new fact.

And a clipping from the Houston Chronicle:

Fifteen percent.

For any parent who has sent a daughter off to college, for any young woman who's enrolled in a university, the story behind that number is downright terrifying. That's how many undergraduate women recently reported they've been raped since they started attending the University of Texas at Austin.

A shocking study conducted at the behest of the UT System Chancellor William McRaven made headlines last week, indicating thousands of students have been sexually assaulted after enrolling at one of the largest institutions of higher education in the United States. More than half of the victims said they were attacked by fellow students.

The idea that you have a 1 in 7 chance (slightly better than a dice roll) of being raped while at college is, indeed, "downright terrifying". But the horror, I would think, subsides a little once you realize that "rape" here includes being "made" to perform oral sex on pain of losing a boyfriend, or sleeping with someone who tells you he's single only to discover that he's got a girlfriend. Now, before you tell me to shut up because I'm not a female college student, please note that the Houston Chronicle addresses itself to the fears of a "parent who has sent a daughter off to college". I'm not quite there yet, but I do have a daughter. And my concerns about letting her attend UT Austin were seriously moderated, let's say, by looking closely at the report's definition of "rape". (Of course, I'm being a bit disingenuous there; I was pretty sure about what I would find when I looked.)

In any case, notice all that inflammatory, fear-mongering language. Even though the great majority of incidents probably do not involve physical force of any kind, we're presented with "shocking" facts about "assault". And although the vast majority of cases are probably not even instances of coercion but manipulation, we are being told that the victims were "attacked". Parents reading this can be forgiven for imagining undergraduate men hiding in the bushes waiting to pounce on their helpless prey.

* * *

(Update May 1)

The Daily Texan thinks there should be more focus on the motives of male rapists in the study. Citing a number of other studies that suggest UT Austin is the hunting ground of 1600-6500 student rapists, it finds the situation "harrowing".

*(Update March 28): In fairness to the journalists, not even the CLASE report's authors seem to understand this point. On page 49, we read "Fifteen percent of female undergraduate students reported having experienced rape since enrollment at UT Austin." I may be wrong about this, but I don't think that question appeared on the questionnaire, i.e., "Have you been raped [or experienced rape] since enrolling at UT Austin?" The correct way to state the result, I think, would have been to say, "Fifteen percent of female undergraduate students reported having experiences since enrollment at UT Austin that constitute rape on the definition used in this study."
**(Update March 30): As far as I can tell, the encounter would be deemed "rape" in the study even if she found it to be a pleasurable experience and granted that he had been right to stake the relationship on pursuing it. The fact that he "made" her do it by "threatening" to leave her turns this sexual negotiation into a rape, it seems.


Anonymous said...

An alarming statistic is that the enrollment in UT's incoming class is now only half what it was just four years ago! I base my assertion on the demographics on page 30 of the report, where I see the Freshman to Senior ratio equals 14/27.

Apparently one has to correct for self-selection. My guess is that slightly more Freshman are enrolled than Seniors, due to growing enrollments and some attrition with matriculation.

From the document itself:

This document presents the web-based findings for UT Austin .... (45,000) Enrolled undergraduate and graduate students were randomly selected to participate in the study. Students anonymously and voluntarily answered
questions using a web-based platform... At UT Austin, 7,684 students participated. The response rate was 17.1%

Interesting to me, even with the self-selection, the M/F ratio of respondents (50:49) nearly equals the invited population as a whole (51:49). (The number self-reporting an additional gender identity is 1%.) The ratio of victims is 58:40:2. This means in this study, women report behaviors interpreted as those of a victim at a rate of about 50% more than that of men. The report discusses this on page 13. "The high rates of disclosure among men are surprising, although not without precedent..."

Also interesting, 45% report being in a romantic relationship. (Seems like a lot to me.) 2% of respondents were in a sorority or fraternity, (compared to the 17% in the general population (6700 of 40,000 undergraduates)). [ Apparently Greeks are 8 times less likely to respond to such a survey as their non-Greek peers.)

The executive summary leaves the definitions to the methods, which are here.

It's not nice of me, but a point can be made that it would be "Attempted rape" (in its weakest form) if a person asks another person for oral sex and then expresses displeasure when rejected:

"Even though it didn’t happen, someone tried to have oral... sex with you without your consent by ... Showing displeasure ...after you said you didn’t want to." (page 17 of above PDF).

I think your original post made a similar point about broad definitions.

Jonathan said...

How much more likely is someone to respond to a survey if they feel it is relevant to their lives? If a victim of sexual assault is much more likely to respond, then that already skews the results in an extreme way. So we are talking about 15% of 17%? (of people who could have responded).

Thomas said...

Yes, I suspect that there's a self-selection bias in this sort of survey too. It was addressed directly in the report:

"...some may assume that students with the experience of victimization are more drawn
to this type of study. Three strategies were used to minimize overestimation of victimization. First, the study was not advertised as a study about victimization. Second, weighting strategies were used so that the findings were reflective of the student population by gender, race/ethnicity, and school classification at the institution. Third, margins of error were also calculated to reflect relative confidence in the findings." (p. 16)

I haven't looked at closely yet, so I don't know what the unweighted results would have looked like. For the results we're talking about, "The margin of error is ≤ +/- 2% at 95% confidence" (p. 48).

The first strategy is probably the most important. Given that they tried not attract people who have an interest in the topic, I guess one could ask how many people were repelled by it after they started answering questions.

Kate Clancy et al.'s SAFE13 study addressed the problem somewhat differently:

"The sample was potentially biased by ethical, pre-participation disclosure that questions regarding these topics were in the survey. Some people may have been more likely to participate in the survey if they had negative experiences, some people may have been more likely to forward the survey link to individuals who had previously disclosed negative experiences in private conversation (snowball sampling), and some people may have been less inclined to participate in this survey to avoid emotional stress of sharing their experiences. Several colleagues directly informed the study authors that they would not participate because revisiting their experiences was too traumatic. Thus, it is unclear if the self-selection of this sample produces over- or under-reporting of negative field experiences."

Here the problem of self-selection bias is apparently dismissed because the effect could go either way and they don't know which.

Jonathan said...

The usual number bandied about is 25%. So the 10% or 15% would be lower than that? But of course we would expect a higher rate among seniors than Freshman, so I don't know how to adjust it for that. The 15% is more shocking if we think it is 3 and it ends up five times more, but you don't see headlines saying: "sexual assault rates lower than previously thought."

Other considerations:

What is the rate for these various kind of events among college age people who are not in college? Without that comparison we are pretty lost about the meaning of these numbers.

We also need to know what percentage is "my boyfriend wanted sex and I didn't feel like it but I did anyway." Just off the top of my head I'd say then that I am a rapist too by that definition, and that therefore this might be rather more frequent than other kinds of assault.

What I'm suggesting is that we don't have the tools for thinking about this in a critical way at all. I don't want to minimize a real problem, but just get a handle on what it is we are really talking about when we talk about sexual misconduct.

Thomas said...

The 1 in 4, or 1 in 5, figures are based on studies with a bunch of methodological flaws too, but they are normally more carefully distinguished from claims that the relevant act is rape. Here are the authors of a key study in Time:

"the 1-in-5 statistic includes victims of both rape and other forms of sexual assault, such as forced kissing or unwanted groping of sexual body parts—acts that can legally constitute sexual battery and are crimes."

I completely agree with you, Jonathan, that we lack the basic tools (at least collectively) to think about this in a serious way. You are exactly right that we need to corresponding numbers for other sectors of society. I just found this in Wikipedia:

"In a survey of 1,788 male inmates in Midwestern prisons by Prison Journal, about 21% claimed they had been coerced or pressured into sexual activity during their incarceration, and 7% claimed that they had been raped in their current facility."

The shocker here, of course, is that, if we take the use of words at face value is that female undergraduates are twice as likely to be raped on the UT Austin campus as inmates in Midwestern prisons.

If we think more carefully about this, we realize that what a UT undergraduate considers "coercion" (and what the study therefore considers "rape" when it results in sex) is probably on a different order than what prison inmates are thinking of. And even they distinguish between between being coerced into having sex and being raped.