In the last few weeks, as a result of incidents such as a sexual assault investigation leading to the suspension of the University of Ottawa men’s hockey team and its coaches  and a University of Ottawa student politician alleging online sexual harassment, there has been a great deal of heated discussion about whether a “rape culture” exists on Canadian university campuses.

Columnist Barbara Kay at the National Post newspaper waded into the fray with this column, in which she states “[rape culture] does not exist” and presents statistics which she claims prove that statement. She also asserts that “[i]f these statistics do not convince you, then I suggest you are in the grip of a serious ideological virus. There is a remedy for it, called critical thinking.”

Okay, then. Let’s look critically at the statistics in Kay’s column.

The source for the statistics and their analysis – the single source that Kay cites in support of her assertion that there is no “rape culture” –  is a blog post on the Pittsburgh Post-Gazette newspaper’s website. (The link to this blog post in the online version of Kay’s column doesn’t work; here’s the direct link.) Kay’s column describes Chad Hermann, the author of the post,  as “a management communication professor at Carnegie Mellon University’s Tepper School of Business” – a position which, according to his LinkedIn profile, he hasn’t held since 2007. Kay’s column also fails to note that Hermann’s blog post was written in 2011 – not, as the wording of the column seems to suggest, in response to the most recent accusations of “rape culture”.

The analysis in Hermann’s blog post uses publicly reported data from three Pittsburgh-area universities on on-campus sexual assaults, and compares the numbers of reported assaults to the numbers of female students enrolled at each university. There are two major problems with Hermann’s methodology.

First, Hermann doesn’t explain how “student” was defined for the student numbers he used in his analysis. Students can be counted in several different ways: per person, per course, per semester, per academic year, or per calendar year. For example, one student taking five courses in one semester could be counted as one student, because it’s one person, or could be counted as five students, because the person is enrolled in five different courses. The definition of “student” is important for Hermann’s analysis, because different definitions of “student” could increase or decrease the student numbers he uses in his estimates. But since Hermann doesn’t explain the definitions used to generate the student numbers he uses, his calculations of rates of sexual assault are meaningless.

Second, Hermann misses some significant information in his source data. His data are drawn from institutional reports made under the Clery Act  – which, as a very quick Google search will show, requires American post- secondary institutions to report on-campus crimes in a number of categories. Those categories include whether the crime occurred on “on-campus property” or  in “residential facilities” (i.e. dorms or student residences operated by the university). Here are the 2009 data from Duquesne University’s main campus which are cited in Hermann’s analysis.

Hermann accurately reports that there were seven “forcible sexual assaults” reported at Duquesne that year. However, he neglects to mention that all of them took place in on-campus student residences. This suggests that a meaningful assessment of female students’ risk of sexual assault on campus should compare numbers of sexual assaults not only to the total number of female students, but also to the much smaller number of female students enrolled at the university and living in student residences. Hermann only uses the total number of female students, which results in a mischaracterization of the risk of sexual assault.

Kay mentions neither of these problems in her column.

Another quick Google search on “Clery Act statistics” brings up a very useful US government website: the Campus Safety and Security Data Analysis Cutting Tool, which aggregates all the Clery Act reports into a searchable database. Kay’s column never really gets around to explaining how 2009 Clery Act statistics from three universities in a single US city prove that “[rape culture] does not exist” in 2014 at all Canadian universities. But since Kay apparently didn’t look at any of this website’s data when researching her column, I decided to explore it myself.

First, here are the most recent data (2010 to 2012) from all US post-secondary institutions covered by the Clery Act. This includes public and private institutions (both private for-profit and private not-for-profit); two-year and four-year institutions; and institutions with programs under two years in length.  I restricted my search to only include institutions with on-campus student residences, because the Duquesne University data indicate a possibly higher incidence of sexual assaults in student residences. And for comparison, in addition to the data on forcible sexual assaults, I’m presenting the data on the other two most frequently reported crimes in each campus area.

Here are the most recent US-wide data for the most common crimes on campus as a whole:

	2010	2011	2012
Forcible sex offenses	2,984	3,442	3,944
Burglary	22,028	20,214	18,554
Motor vehicle theft	3,615	3,621	3,244

And here is the subset of that data for crimes in on-campus student residences:

	2010	2011	2012
Forcible sex offenses	2,092	2,425	2,802
Burglary	11,919	10,985	10,318
Aggravated assault	780	825	817

Clearly, students living in student residences are at much more risk of sexual assault than students not living in residence. And the numbers of forcible sex offenses in residences, and on campus in general, have increased in each of these three years, while the numbers of burglaries have decreased.

I then looked for Clery Act data that might be comparable to post-secondary education in Canada. Since the majority of post-secondary institutions in Canada are public institutions, I looked at the data from two- and four-year public institutions. These institutions could be considered as similar to colleges and universities in the Canadian post-secondary system.

Here are the most recent campus-wide crime data for US four-year public institutions with student residences:

	2010	2011	2012
Forcible sex offenses	1,468	1,655	1,883
Burglary	10,358	9,485	8,760
Motor vehicle theft	1,641	1,635	1,443

And here are the data for crimes in those institutions’ student residences:

	2010	2011	2012
Forcible sex offenses	1,051	1,168	1,319
Burglary	5,719	5,227	4,874
Aggravated assault	405	435	443

The overall crime data for two-year institutions with on-campus student residences:

	2010	2011	2012
Forcible sex offenses	214	262	261
Burglary	2,438	2,351	2,031
Motor vehicle theft	1,034	943	793

And the crime data for the residences at those institutions:

	2010	2011	2012
Forcible sex offenses	62	72	73
Burglary	741	660	630
Aggravated assault	78	79	82

These data indicate that students at two-year institutions seem to be at more risk of sexual assault on campus in general, since a lower percentage of the total number of sexual assaults took place in student residences at these institutions than at four-year institutions. However, the numbers of reported sexual assaults are not decreasing at both the two- and the four-year institutions, unlike the number of reported burglaries.

I don’t claim for a minute that these data prove that a campus “rape culture” exists. Kay’s column also neglects to mention that “rape culture” isn’t just about rape. It’s a culture in which sexual assault and aggression is tolerated or even normalized – so there are many more characteristics of “rape culture” than how many rapes occur or are reported. Simply counting the numbers of sexual assaults on a campus will not determine whether a “rape culture” exists there.

But that doesn’t make Kay’s denial of  “rape culture” any less inflammatory, illogical, and poorly researched. Her statistics do not prove anything that she claims, and that’s not because of “a serious ideological virus” or “delusion” – it’s because there are major methodological flaws in her single source of evidence. There’s a potentially interesting and provocative discussion in the much more extensive and current data that are so easily available, but Kay’s column isn’t that discussion. If Kay wants to argue that there’s no “rape culture”, that’s fine – but to dismiss the notion of “rape culture” without any reliable supporting evidence is bad journalism and irresponsible writing.