The "Fact" That Never Changes

The invaluable Ashe Schow has a new post up at the Washington Examiner discussing how the bogus, unempirical, yet remarkably durable "1-in-5" factoid retains currency among Title IX rape inquisitionists. It occurred to me that there's a significant point to be made here: the figure has been around nearly thirty years. As Heather MacDonald pointed out in a City Journal piece I link to almost reflexively on this subject, the figure originates with a 1987 study by Mary Koss (PDF), which has remained unchanged ever since. Sometimes one even hears figures as high as 1-in-4, but seriously, this is what they expect us to believe:

To even come close to showing the actual reported and theoretical underreporting (68% of all rapes are not reported, per RAINN), the graph above required logarithmic scaling. (Data for actual rapes came from the FBI's Uniform Crime Reporting database.) Over that entire time, the overall numbers rose and fell, but the party line on rape is that it hasn't changed in nearly three decades. Of course, I'm being somewhat unfair in this comparison — the 1-in-5 figure represents a lifetime risk, though over which lifetime we're never sure. It's frequently sold as time on campus, which would mean at a four-year institution, a typical woman stands a 1-in-20 risk of being raped in any given year. 5% rape rates would make for the kind of panic that would shut down college campuses everywhere, and yet this does not happen, and indeed young women still eagerly attend college.

I wanted to remove the invariant 1-in-5 figure from the graph above to highlight something actually quite good:

The overall rape rate has gone down considerably and very nearly continuously since a peak in 1991. Even accounting for the RAINN-derived underreporting, we should expect to see similar changes in any real-world data collection that actually causes some variance — that is, if the individuals in question were engaged in something like actual science.