Took me a bit and I am a little drunk but here are the totals. We had a good showing.
Note: some comments got cut out because the server doesn't like displaying more than 300 from what I can tell. I did my best with what it would show, I just sorted them by new. This should be fairly representative of the currently active userbase. As an aside, I believe we did a similar survey a looong time ago, and we had only 33% of the active userbase as trans.
Random thoughts:
A lot of people were very very cute and confused and essentially asked me to decide their gender for them. Eggs? Probably. This isn't Harry Potter and I'm not a hat so I just went with what they were sounding more convinced of.
A lot of people are even cuter and don't understand how to follow instructions, though some of this is my fault. Made this a little harder to organize.
Some people were not cis and did not identify with the words 'transgender' or 'gender diverse'. If I ever do a trans/adjacent survey again, I think I will ask 'Are you cis?'
I may do a survey for queer people overall eventually, and the question will be 'Are you cishet?'
I would love to do more scientific, inclusive polling and have better and more questions and options, but we need some good secure polling tech for that, which we don't have. So I just have to ask simple questions and get a handful of answers.
Next time I will look into how feasible it is to post a couple of comments and get responders to upvote certain ones. This might fix the issue of the display of comments being limited.
Since some people have two sets of pronouns, both of their pronouns are included separately.
Since the poll was public, some marginalized groups probably shied away from answering. If we ever get a secure way to poll people, we would get more realistic estimates of the trans and cis women userbase.
#Tallies
Yes: 121
No: 137
Maybe: 37
Total: 295
If you think something is fucked, you can do it yourself, the thread is public. Hope you like pie! 
We have a very good sample for active hexbear users/commentors, basically. If you wanted it to be more scientific you need your own polling service, ask many different questions that are basically the same thing, and poll over a longer period of time.
With this limited setup though we can get a fairly good view of the general trends of the active userbase that posts and replies Wednesday through Saturday morning, as we surveyed about 1/3rd of them.
If we made multiple surveys with the same methodology over a long period of time we could speak for the general hexbear population (re people that have gone inactive or semiactive) much better.
If you want to get an idea of how much self-selection is skewing results, one option is to compare the proportions of various pronouns tags of those who answered, versus active users who did not. Not a great measure, but it'll give you a basic idea. That's assuming that pronoun tags are kept.
Suggesting pronouns tags because that's a relatively fixed and measurable quantity that mostly preserves anonymity of answers. another potential option is measuring activity quantiles for like comment and vote counts between answers and active population