(or, how the numbers in pittsburgh just don’t add up)
I’ve been wanting to get to this one for two weeks now, but a few other projects, and then that damned and damnable arbitrator’s decision, got in the way. Let’s not keep it waiting any longer.
Two Mondays ago, the Pittsburgh Tribune-Review published a piece about the new state law that requires colleges and universities to provide sexual assault prevention programs on campus. While there is much to discuss (like PA Rep. Scott Conklin’s claim that some people believe when you go to a university, nothing bad can happen) and to deride (like Pitt freshman Dominique Benzio’s insistence that some people don’t even understand if they have been sexually assaulted or not) in the piece, I want to focus, clearly and simply, on the numbers it presents. And the contradictions it, and almost everyone else, ignores.
You can not have an article on sexual assault in college, of course, without a solemn invocation of that infamous, oft-repeated, almost-as-oft-debunked One-in-Four statistic. The Trib does not disappoint:
The National Sexual Violence Research Center in Enola, northwest of Harrisburg, estimates 20 percent to 25 percent of women are victims of forced sex during their time in college.
But where the Trib, like everyone else who uncritically accepts this uncritical notion, does disappoint is in its failure to acknowledge that the numbers do not add up. And, in fact, that they do not even come close.
As a supplement to that same article, the Trib published a Campus Safety sidebar that provided a list of reported sexual assault offenses for eight local colleges over the past three years. All told, there were 65.
65. At 8 colleges. Among tens of thousands of female students. Over 3 years.
That’s a long way from 1-in-4. And thus a complete repudiation of the now-boilerplate statistic claimed in the article.
Those numbers, however, only represent reported assaults that occurred in student housing. Surely the numbers would be much higher, and much more alarming, once you counted sexual assaults that occurred elsewhere on campus and in the community at large. Surely that would get us much closer to the canonical 1-in-4 wisdom.
I decided to test that theory by examining the sexual assault statistics at Pittsburgh’s three largest residential universities: the University of Pittsburgh, Carnegie Mellon University, and Duquesne University.
Thanks to the Clery Act, all higher-ed institutions must publish and distribute an annual campus security report, complete with full crime statistics from the past three years. These documents are easily found online. You can check them yourself, and follow along with the numbers, here:
The University of Pittsburgh. Carnegie Mellon University. Duquesne University.
At the University of Pittsburgh, there are roughly 14,800 female students. If their chances of being sexually assaulted are 1-in-4, there should be about 3,700 sexual assaults each year. In 2009, the most recent year for which full statistics are available, Pitt students reported 4.
At Carnegie Mellon University, there are roughly 3,900 female students. If their chances of being sexually assaulted are 1-in-4, there should be about 975 sexual assaults each year. In 2009, CMU reported 6. (That figure was a three-year high.)
At Duquesne University, there are roughly 5,700 female students. If their chances of being sexually assaulted are 1-in-4, there should be about 1,425 sexual assaults each year. In 2009, Duquesne reported 3.
Just to be clear, and so those numbers stand out, here are the total number of reported sexual assaults for each of the three campuses in 2009, followed in parentheses by the numbers those universities should have suffered, according to the 1-in-4 figure:
PITT: 4 (3,700)
CMU: 6 (975)
DUQ: 3 (1,425)
Which means that, instead of 1-in-4, their chances of being sexually assaulted in 2009 were:
There is, of course, a widely reported (if virtually unverifiable) statistic that says 90 percent of sexual assaults go unreported. The Trib piece took note:
90 percent or more of those victims do not report the assault.
If we grant that claim and adjust the numbers, then here are the total number of sexual assaults that occurred in 2009, followed in parentheses by the numbers those universities should have suffered, according to the 1-in-4 figure:
PITT: 40 (3,700)
CMU: 60 (975)
DUQ: 30 (1,425)
Which means that, even if we grant the 90%-are-unreported figure, women’s chances of being sexually assaulted while attending those universities in 2009 were not 1-in-4 but:
Even after adjusting for the possibility that 90% of sexual assaults on those university women went unreported, to get to 1-in-4, sexual assaults on those campuses in 2009 would have to be increased (and unreported) to the tune of:
Is it possible that these numbers are just anomalies? That they represent a down year for sexual assaults on these local campuses, one that is not indicative of a typical year?
Not according to the crime statistics published for the past three years.
In 2009, Pitt reported 4, CMU 6, Duquesne 3. For the period 2007-2009, Pitt averaged 3.33 per year. CMU averaged 4 per year. Duquesne averaged 1.33 per year.
Which means that, in each case and on each campus, the three-year average was lower than the 2009 figure.
If we use those three-year averages, women’s chances of being sexually assaulted while attending those universities from 2007-2009 were not 1-in-4 but:
If you combine all of the crime statistics for Pitt, CMU, and Duquesne — again, the city’s three most populous urban campuses — and run an average for 2009 (which, as we’ve already seen, is above the most recent three-year average), here’s what you get:
# of Female Students: 24,400
# of Reported Sexual Assaults: 13
Chances of Being Sexually Assaulted: 1-in-1,877.
If you accept the claim that 90% of college sexual assaults are not reported, you get:
# of Female Students: 24,400
# of Reported Sexual Assaults: 130
Chances of Being Sexually Assaulted: 1-in-188.
To get to 1-in-4, you still need 47 times — not 47 more, but 47 times more — sexual assaults.
As I noted at the top of this post: the numbers in reality don’t come anywhere close to matching the numbers in the claims.
Is Pittsburgh just 47 (or, without a guess of an adjustment, 470) times safer than any other city, suburb, exurb, or small town in the country? Are Pitt, CMU, and Duquesne (plus all those other colleges on the Trib’s PDF) just terrific anomalies, just unexplained and uncharacteristic outliers in the data?
If so, and either way, there must be some city, some university, some collection of colleges somewhere for which the real, actual, federally mandated crime statistics — even bumped up to account for 90% of non-reports — work out to that 1-in-4 figure. Or at least come close. Right?
In fact, there ought to be a lot of them. Right?
If not, then the people who insist on advancing those dramatic claims while also ignoring these decidedly less dramatic realities would seem to have an awful lot of explaining to do.