A recent article about the Thai army’s decision to stop branding transsexual conscripts as mentally disturbed, reminded me that Thailand’s national draft lottery is one of the more interesting measures about the prevalence of trans-ness. It’s pretty close to a random sample of the country’s entire population of 20-year-old males. In the latest article, an Army spokesman said transsexuals make up less than 1 percent of conscript, although a different spokesman in a 2006 article said that 1% to 2% of those who show up for the draft lottery are either transvestites or transsexuals.

(Guess Eddie Izzard will be getting his 1st Battalion, Transvestite Brigade, Airborne Wing…)

It’s interesting that that 1-2% figure roughly tracks with the 2.8% of men and 0.4% of women, ages 18 to 60, who reported at least one episode of transvestic fetishism in a Swedish study that’s the only one I know of that posed the question to a random sample of the general population. (The question was actually part of a larger survey on a variety of health issues, which is why the Swedes surveyed the entire country.)

It’s seems reasonable to assume the Thai statistics represent a minimum estimate because despite Thailand’s reputation for trans tolerance, being dismissed from the Army for being trans has serious life-long consequences. (Because men are required to prove if they have completed their national service when they apply for jobs or bank loans ,and those with a “mental disorder” discharge are automatically disqualified from many jobs and mortgages.) So it’s reasonable to assume that only the most gender variant trans people are being spotted by the Army and that those who can closet themselves do so.

I’d treat the Swedish study as a rough estimate, since on the one hand, even though the survey promised anonymity, it’s not unreasonable to suspect some people probably didn’t admit to trans behavior. Especially since (according to those who’ve seen the full study—I’ve only seen the abstract myself) the question posed was “Have you ever dressed in clothes pertaining to the opposite sex and become sexually aroused by this?”. So besides people who weren’t willing to admit to this, there were probably trans people who answered “no” because they didn’t see their crossdressing as sexual motivated. On the other hand, since reportedly anyone who did so at least once was counted, it’s likely they may have counted some people who crossdressed as an experiment, but who probably wouldn’t be considered—nor see themselves—as trans.

I’ve heard rumors of a UK study that looked at what percentage of male patients brought to the emergency room were underdressed—i.e. wearing or more article of women’s clothing under their men’s clothing—but so far I’ve not found any evidence that the study exists. Since not all trans people underdress, the statistic (if it exists) would drastically under-report the actual numbers of trans people. But one might be able to combine it with some survey work looking at the prevalance of underdressing. Obviously once again you run into issues of properly sampling a closeted population. So any result number would be an extrapolation of uncertain data based on other uncertain data, and would be at best a rough estimate.

Trying to estimate the size of a closeted population is inherently imprecise. But my take on it is that it’s a bit like on-GPS navigation in sailing, in which compass/celestial sightings are inherently imprecise (due to the motion of the boat and other reasons). Nonetheless you can triangulate among enough of them to calculate your position in a useful way—albeit within what’s referred to as the “area of uncertainty.” So we’ll never have exact numbers, but if different methods end up with results in the same neighborhood, it may be possible to develop some estimates that have a greater reliability than the individual surveys.