Except the posts above are talking about the DoJ (1 out of 5) report from 2005 & 2009 being significantly flawed and used out of context - even the lead author specifically states that the report cannot be used as representative. Alas, the 2009 analysis is blocked, so I'll have to look at the details later on it. I'm curious as to how they came up with the 1 in 5 from the "complete" analysis.
The rates he quoted were still from a different time frame (one year) than the percentage he quoted (over four or five years).
But I'm at work, so I may have misread something.
Also, we have many different studies being referenced here. Can you do me a favor in the future: Link to a pdf (or at least quote the title) when you say "the 2009 analysis" or "2005" or "2009" so I know we're talking about the same things, and quote from and link to where specific studies are described as flawed or as used out of context?
To try to recap, it was the study from 2005 - 2007 whose lead author said could not be used as nationally representative ("It comes from a 2007 study funded by the National Institute of Justice, called the Campus Sexual Assault Study, or CSA," from Yoffe's piece); the study from 2000 was the one which tried to multiply rates. In the Politifact article (which the New Republic article I linked to links to), similar criticisms are raised about the 2007 study. It also references that absurd attempt to derive cumulative numbers by multiplying rates when it referencing the 1996 Sexual Victimization of College Women survey; just to make things more difficult this study is referred to as the "2000 study" in Yoffe's article (it came out in 2000, but the data is from 1996). It also references several other studies with similar numbers: There is the nationally-representative
2007 Medical University of South Carolina study, mentioned in the New Republic article, which says that 11.5 percent of college-aged women have "ever been raped", but that is a lifetime number. There is also the CDC number, also lifetime, which says that
So, to make sure we're on the same page:
1996 | 2000 : Measures rates of rape among college students, finds a rate of 1.7% for completed and 1.1% for attempted rape over a 6.91 month period while a student; did not measure rates cumulatively, though does make an ill-advised attempt at multiplying rates and comes to an estimate of 20 to 25 percent.
2005 - 2007 | 2007 : Finds that among college women surveyed at two Midwestern colleges, 13.7 percent reported a sexual assault since they had begun their schooling.
2009: Additional data for the above study, which says that, specifically among seniors, almost 20% experienced some type of sexual assault since the start of their schooling. Study author (and other experts) advise against using number generally due to its not being nationally representative
2007 : This is the 2007 Medical University of South Carolina study, which found that 11.5 percent of college-aged women "have ever been raped"; they estimate a number of 18.0 percent for the general population. This study is nationally representative, but doesn't ask specifically about sexual assaults or rape that occurred while a student
2010 : The CDC study, which found a rate of 18.3 percent lifetime for all women for completed rape. This study may represent an undercount, as it doesn't cover vulnerable populations like individuals under the age of 18, the homeless, the hospitalized, prisoners, or people in residential treatment programs.
So, we have a studies are representative and studies that aren't; studies that focus on college students annually and college students lifetime, but don't ask about college students' experiences specifically at college; studies that ask about students' experiences specifically on campus but do so over a short time frame (6.91 months) and / or aren't nationally representative. And we have numbers for both "sexual assault" and "rape" and "sexual assault" is by necessity defined more broadly than is "rape." And we have to keep this in mind:
Koss broke it down this way: the FBIÂ’s Uniform Crime Statistics equal what gets reported to police, the Justice DepartmentÂ’s National Crime Victimization Survey equals what people disclose to surveyors, and academic research surveys are "what people tell to nongovernmental entities when survey authors have the freedom to phrase the questions as they think best as scholars."
So, it can be very difficult to have an apples-to-apples comparison of studies about the experience of college students at college because there are so many different questions being asked and time frames being used. It's better with general population numbers; the lifetime rates for forcible rape, at least, tend to fall in a pretty consistent 12 - 15%-ish (give or take a couple points for the outliers) range; and the annual rates for forcible rape tend to also be in the 0.5 - 0.7% range, for instance.