Applied Epistemology

When I inform students about the use and abuse of unreliable statistics, one very common reaction is to distrust all such claims -- except for the ones that happen to support the views the student already believes. This is particularly clear when I pose a question about the effect of college degrees on income. The question is no different than any number of weak correlation-cause arguments that I've specifically and repeatedly warned them about. But whereas the students are extremely doubtful of true claims that conflict with received opinion, they almost all buy the "college makes you earn more" pseudo-evidence.

This article annoyed me almost as much as the "leading scientists believes we're all characters in an as-yet-unreleased PeterMolyneaux game" crap they published around the same time.

The imprecision of language like this:

“Surveys and studies to the contrary notwithstanding, the conclusion that men have substantially more sex partners than women is not and cannot be true for purely logical reasons,” Dr. Gale said.

irks me. Because, in context, there is no "purely logical reason" that can't be true. What's more, the article switches back and forth between discussing averages and medians without doing an adequate job of flagging either, leaving the discussion a muddled mess.

Only towards the end, after a tedious exposition of the mathematical impossibility of the survey results, does the article mention that, uh, actually, it maybe isn't so much logically or mathematically impossible that men and women may have widely variant numbers of lifetime sexual partners, it's just seems unlikely. The mathematicians don't have any, you know, data to show that the "prostitution effect" is, as they claim, "negligible," but they're willing to assume it without discussion (presumably because a side trip to Reno or Thailand either isn't in the realm of the mathematickal experience or just isn't polite conversation within the tower).

It comes across as a bunch of Frinks complaining to a reporter about how stupid the sociologists are because they don't understand math. Next week, maybe the Post will run the sociologists' side of the story... When It Comes to Sex, Mathematicians Reject Empiricism... (I'm sure with just a little work a clever writer could make the entire article one long double entendre).

On an entirely unrelated note: WTF?

Earlier, I emailed Prof. Gale, asking for an explanation of his dismissal of the world's sex workers as negligible. No reply yet.

The mean-median switching was bad. However, the underlying point -- that surveys of any information that people are likely to lie about are inherently unreliable -- is right.

I thought about linking to the NYT crap on Nick Bostrom's Matrix fantasy. But I couldn't bear to. Reading the original 2003 paper was enough for me.

By the way, the Wizards' link was an old April Fool's prank. Though 4e is probably in the pipe. I sadly note that Dungeon and Dragon magazines have been closed down by WotC, the better to be made into consumer-shakedown online ephemera.