Robert is exactly right that sexual harassment can be both overreported and underreported — especially in the sense he notes, where some real incidents are not reported and some non-incidents are. In fact, virtually every significant social problem, whether it’s theft, medical malpractice, child abuse, etc., is both underreported and overreported in that same manner.
The interesting question, I think, is whether the totality of the problem is overreported or underreported. I feel fairly confident — on college campuses, at least — that it is rather dramatically overreported. Several facts lead me to believe this is the case.
First, like Robert, I have never seen a significant sexual-harassment survey that was limited to the actual, legal definition of sexual harassment. Every survey I have ever seen used a wildly overbroad definition. Based on some of the definitions I’ve seen, I’m sometimes surprised that 100 percent of students haven’t reported that they’ve been harassed.
Second, colleges as a rule make a huge effort to make it extremely easy to file highly confidential sexual-harassment charges and make an equally large effort to publicize not just their (overbroad) definition of harassment but also the reporting process itself. So there’s little chance that any reasonably diligent student facing harassment issues won’t know where to turn.
Third, my own review of internal university documents obtained during litigation of speech-code cases shows an alarmingly high rate of non-incidents not merely reported but pursued and prosecuted (sometimes entirely behind the closed doors of “confidential” student proceedings).
Until I see a harassment study that reaches a large population and is based on an accurate definition of the term, I’m going to be extremely skeptical of the university conventional wisdom that the totality of the sexual harassment is “underreported.”