In the comments on my last post, Rhymes with Scrabble makes a reference to the Rita Mae Brown quote: "The statistics on sanity are that one out of every four Americans are suffering from some form of mental illness. Think of your three best friends. If they're okay, then it's you."
RWS suggests that the assortative forces of friendship make this reasoning misleading; as she says, "a lot of the crazy people hanging out with each other." I would go even farther than this, to conjecture that probably one of the better indicators that one is sane is if one's three best friends are sane. It's also, I suppose, a pretty good indicator that one is boring, and it's definitely a strong indicator that one is not selecting one's best friends from the groves of academe.
Of course, even if there was not this assortative selection, one wouldn't think that The Brown Method was actually a good way of figuring out whether or not they are crazy. If there was no assortative befriending on sanity, having three sane friends would have no implications for one's own probability of being sane (still 1/4). What is interesting would be to imagine the world where the Brown Method did work: one where sane people liked just enough insanity in their lives to want 1 of their three best friends (no more, no less) to have mental problems, while those with mental problems only allowed sane people into their inner circle.
Meanwhile, a person I know who with a long history with antidepressants told me that a similar logic applied to the statistics of side effects. The side effect literature says this drug has a 10% chance of this side effect, another drug has a 10% chance of that side effect, and so on. He said it was mostly the same 10% of people who end up getting side effects no matter what they take and they end up wandering for years from drug to drug searching for one that was going to be tolerated well by their bodies and not be totally inert for the problem for which they were taking the drug in the first place. I have no idea if he was correct about this. I do know that he certainly had the experiences that would seem to qualify him as a charter member of that unlucky One In Ten club.
* By complete coincidence, I had a roommate in college who was not related to me but had my same last name--except his family pronounced it so that it rhymed with crazy. This is less of a transformation of the original German than my preferred rhymes-with-niece, but, as long as you aren't getting the originating pronunciation right anyway, why bother with a pronunciation that no one, anywhere, is going to get right without being instructed? He even has "(Rhymes with Crazy)" between his first and last name in the signature file of his e-mail for his workplace account. Yo, I just googled him, and it's there on his workplace webpage as well. (Note also that his e-mail address on said webpage is to "E-Man." I was the one who started calling him that in college, with a whole By The Power of Grayskull subtext, and, lo, he took it to heart. It's nice to know that I can have that kind of power on username choice even more than a decade later.)
Update, Monday: Mr. Rhymes With Crazy himself responds in the comments.