Gaydar update: Additional research on estimating small fractions of the population

Gary Gates writes the following in response to the discussion of my recent blog on the difficulty of using “gaydar” to estimate the frequencies of gays in a population:

First, here’s a better (I think, anyway) method than using AIDS deaths from the NY Times (yikes!) to estimate the % of the military that is gay or lesbian.

Gates estimates 2.2%, with, unsurprisingly, a higher rate among women than men.

He continues:

Here’s a tale of the false positive problem affecting who gets counted as same-sex couples in the Census and attached is a working paper that updates those analyses (with better methods, I think) using ACS data.

In this paper, Gates (along with Dan Black, Seth Sanders, and Lowell Taylor) finds:

Our work indicates that over 40 percent of same-sex “unmarried partner” couples in the 2000 U.S. Decennial Census are likely misclassified different-sex couples.

40% misclassification. Wow.

2 thoughts on “Gaydar update: Additional research on estimating small fractions of the population

  1. Gary Gates estimates that only 0.6% of active duty male military personnel are homosexual or bisexual, which supports my argument that low AIDS death rates among solders in the 1980s and 1990s suggest that a volunteer army is not very attractive to male homosexuals.

    My 1994 article in National Review, "Why Lesbians Aren't Gay," list several dozen tendencies on which male and female homosexuals tend to differ.

  2. Regarding the overestimation of same sex or "unmarried partner" couples by the Census, my observation was that the Census bureau was coming up with improbably high numbers of "unmarried partners" in Western states like Alaska and Wyoming. As far as I can remember, there was a higher rate of "unmarried partner" couples in some of these Wild West states than in the District of Columbia, which sounded highly improbable to me.

    My guess was that the word "partner" means something different in cowboy culture than it does to federal bureaucrats, as in "Tex and I are partners, and we're unmarried, so that must mean we're unmarried partners." I bet quite a few didn't realize the feds were intending "unmarried partner" as a euphemism for "homosexual lover."

Comments are closed.