Those are the real stats, 1.2 per 1,000. Not 1 in 5.
John Shebanow

I’m afraid you’ve made the same simple mistakes as Mike in the other thread, and some more on top of that:

  • The crime victimization data you’re quoting is per year, whereas the number of 20% is per lifetime. That’s why you’re off by at least a factor of a couple dozen. And that’s why the number of 1.1% in the last 12 months that you quoted matches up fairly well with the DOJ’s statistics you quoted, after you account for underreporting. I honestly have trouble understanding how you missed this, assuming your comment is in good faith.
  • Being concerned about extrapolating from 18k to 126M is a common concern among people with no mathematical or statistical background, which I don’t blaming you for not having, but I hope you’ll adjust your opinion with the new knowledge: The population size does not matter at all when estimating proportions, only sample size matters, and 18k is plenty enough to get sub-1% accuracy. Please see my response to Mike in the other thread for references; please see e.g. this standard calculator that doesn’t even let you enter the population size.
  • You are right that there is a reporting bias, i.e. people who have been raped may be differently inclined to answer a survey like this than those who haven’t. However, you’re assuming that this bias is in one direction — you’re assuming that people who have been raped love talking about it to government agencies. This is absurd: remembering and talking about how you were violated is not fun, dude, it’s traumatic; rape is something most people want to hide and forever forget and pretend to themselves that it never happened so they can get on with their life, they just aren’t able to cause it keeps haunting them. Know how combat veterans with PTSD are often not too keen on discussing their traumatic experiences in detail? Same here; something like 30–50% of rape victims suffer from PTSD. People who have been raped often spend years before being able to talk about it even to their closest friends, much less strangers, and even less so strangers from the faceless government — so if anything, the bias is likely in the other direction.

The rest of your comment is minimizing the experience of rape under intoxication to the point of being unable to consent (which is what the survey asked about) — I wonder if you are one of those people defending Brock Turner — and minimizing the experience of someone who has escaped a rape attempt. You… think… that these…don’t count?.. (To use another analogy here, as a combat veteran, you don’t only get traumatized if you die: you also get traumatized if you narrowly escaped death).

You are right, however, that a majority of rapes are committed by people the victim knows. This is because people generally spend more time with people they know, and because people’s guard is down with people they know. It absolutely does not mean that strangers should be feared less: if even people you know and trust can rape you, that means strangers should be feared even more — it simply means that women are better at proactively avoiding dangerous situations with strangers than they are at anticipating that their friend or relative might rape them.