Well, if the survey data wasn't cooked, to everyone's surprise, including the transbians, r/AL's userbase as of last year was majority women. Even so, I think it's safe to assume very few of these women are actually lesbians:
Ver archivo adjunto 2832815
Ver archivo adjunto 2832816
archived 2 Dec 2020 16:56:25 UTC
archive.md
archived 27 Dec 2021 20:05:02 UTC
archive.md
It's definitely cooked. First of all, studies show that women are more likely to take voluntary surveys than men are, so the results are going to be skewed towards women from the start. But second of all, as with all studies of this nature, the men involved have a vested interest in portraying the sub as majority female.
After all, imagine the reaction people would have if it was demonstrated that the big "lesbian" sub on reddit was actually majority heterosexual male pervert and had hardly any real women, let alone real lesbians, at all! They can't have that. That would make them look bad and probably drive away the real women and especially the real lesbians that still go to that sub. It would create massive issues for them.
So they cook the books - probably not directly screwing with the numbers, but probably by taking the survey over and over and over again pretending to be "cisgender lesbian women" every time, so the numbers come out the way they want them to. Since none of these surveys ever even so much as require an email address to prevent people from submitting as much as they want, they can do that to their hearts' content.
They've done this kind of thing on other subs as well - the buggered to fuck AITA survey that claimed the users were "majority female" despite the same flaws and the same issues of vested interests (male users needed the survey to show the sub was majority female so they could claim there was no way the sub trended towards being misogynistic) as this one comes to mind. Anonymous surveys that can be taken as many times as anyone wants to take them, administered by and for people who have a vested interest in getting a particular result will
never be reliable gauges of actual community demographics.