Designing policy is a tricky thing, and the difficulty is radically increased when there is a lack of good data. This is one of the reasons I dislike Stephen Harper so much; he actively resists using data and has made it a policy to defund, muzzle, or discredit science so that he can decide things without any pesky facts mucking up his ideology. Reading my Facebook feed today really brought home just how hard it is to get good data to make decisions even when we are actively looking for it. In particular people were talking about how many trans folk there are in the population.
The estimates in the thread ranged from 1 in 6,000 to 1 in 1,000 but it was pretty clear that nobody really knew. A cursory search of the internet reveals that there is nothing like an accepted number but surgery rates suggest that 1 in 10,000 is way too low - other than that we just don't know. Of course this is drastically more complex because before we ask this question we have to be very careful about what we are asking. Some people identify as genderqueer, genderfluid, nongendered, etc. and while those are different than being trans they illustrate that gender is a slippery thing to nail down even if you can rely on people to answer questions about it openly and honestly. Given how marginalized trans people are and how violently people react to gender absolutism being challenged we know the answers we get to questions aren't going to be reliably true.
Given how hard it is to figure that out I was amused at the thought of trying to figure out what percentage of people are gay or polyamorous. At least for trans folk you have things like medical procedures and government change of gender forms that might let you count things usefully (you will miss a lot but there is *some* hard data to provide boundaries) but how would one even go about trying to count gay folks? I know an awful lot of people that don't cleanly identify along bisexual / straight / gay lines, much less along the gay / straight binary. How do you answer a gay / bi / straight poll question when you have had all straight experiences but would be open to a same sex relationship if the right person came along?
Poly people are even harder. Do you count people who have cheated on the assumption that they want to be poly but can't be accepted as such? (Hint: This would not be a good assumption.) It is similar to the other questions in that there are issues with being outed so the answers you would get would trend towards dishonesty. Plus you have the added twist of people not even knowing they are naturally poly because they don't know what the word means or have never seen it in action which seems like it would be a lot more common than the analogous situation with the gay / straight or cis / trans questions.
Looking at the published data on such things tells me that the people doing the research have a major disconnect with the experience on the ground. For example, if you look at Wikipedia you see a lot of big studies that conclude that bisexuality is something like 1% of the population. However, when I look at data published by dating sites and experiences people have in that regard I notice that poly people identify as bisexual closer to 50% of the time. The same sort of correlation also exists with BDSM. It is possible that the three things are related directly but I suspect it is just that once a person comes out as being part of one sexual minority they have little reason to hide the other. I can't say with any particular authority that trans folks are more likely than the average to be poly / into BDSM / not straight but I would lay a pretty big bet on yes, and by a lot. I don't know that if our culture were suddenly completely free of discrimination against bisexuality that we would actually see a 50% rate across the board but it sure as *hell* is higher than 1%.
I have this ravening curiosity about sex, gender, and desire in the human population. I want to know how all the people live, how they see themselves, what they want, and how they want it. The trouble is that even a simple question such as "what % of the population is gay?" currently has no decent answer so my musings suffer from a shocking lack of reliable data.