You wrote: "I admit that my eyebrow couldn't help shooting up at the implication that anyone under the age of 18 should be included in a ratio of adults identifying as transgender, and the empirically dubious conflation of self-identification with an externally-defined 'larger umbrella term', but them's minor methodological quibbles." These are not "minor methodological quibbles." They're about how demographers define the subject population under study, which seems central not peripheral.
In the first case, there are many, many people who identify as transgender who haven't reached the (arbitrary) age of 18. So, we shouldn't confuse population data about minority adults with the total number of said population in the larger population. We don't do that with any other demographic category, like race, sex, gender, religion, economics, etc. Are Black children not counted when we're measuring the number of Black people in the U.S.? No we shouldn't include children in data that's about adults. My point is that transgender people aren't only adults and reports on the numbers of transgender people shouldn't be limited to only those over 18. That's inaccurate and misleading.
The "larger umbrella term" is hardly "externally-defined." In fact, the concept of the transgender umbrella emerged from within the transgender community. When demographers restrict surveys to only those who "identify" as transgender, they're inevitably missing chunks of the population that rightly should be captured by their surveys. Not every person who fits under the transgender umbrella identifies as transgender. And we'd never ask a question this way when it comes to race, age, religion, etc.
These two issues point to messy, ongoing problems with demographic research on gender and sexual minorities. The root of the problem is the external imposition of demographic terms onto a population for survey purposes, rather than starting with community understandings of themselves and using community language in any larger population surveys. When we survey populations using more open-ended questions using community terminology, we get very different data that what gets reported in national media.
Had Gallup done this and/or employed best practices for surveying SOGI populations, we'd have a better understanding of their numbers in the larger population. But, alas, they took shortcuts that produce clickbait headlines suitable for bumper stickers, rather than accurate data about the population surveyed.
