For this, we’re gonna be looking at 28,000 trans people and how their sexual preferences differ per how long they report taking estrogen vs testosterone. I’m also looking at a small subsection of my Big Kink Survey questions that focus on BDSM-related questions.
Summary of my impressions - skip this to look at the graphs below first if you don’t wanna be primed
Transwomen seem to have greater shifts over time on hormones than transmen do.
Trans sexuality tends to start out quite different from cis sexuality, even before starting hormones. The size of this gap overshadows the size of shifts within trans preferences.
While there's some mild changes, sexuality doesn’t seem to significantly change over hormone length for most BDSM items measured, with the exception of sadomasochism.
After time on hormones, trans sexuality shifts to match that of cis people (of their preferred gender) for some instances, but not all.
My data doesn’t have the dexterity to sort out the social pressures confound. Maybe it feels identity validating to lean into being submissive as a transwoman, or maybe, as transwomen transition, they get more into transwomen communities, which encourages greater acceptance for alternative sexuality or whatever.
But it feels strange for this to explain everything we’re seeing. Why are we seeing shifts primarily for transwomen and not transmen? Do transmen not get into accepting communities as much that encourage their transition? Really I’ve seen a lot of people arguing for opposite things, that transmen are more susceptible to social contagion. This doesn’t seem reflected in this specific data.
And if the shifts are from identity validation or social affirming into having what seems to be cis women’s sexualities, then why are the shifts stronger for heavier things like sadomasochism and heavy bondage, and much more mild around stereotypically feminine preferences like general submission and light bondage?
This confusion also is evidence against my theory that hormones have something to do with the submissive/dominance gap; if hormones themselves are causing changes, then it’s really counterintuitive for this impact to happen more on the extremes of kink. It might in fact work that way, but it’s not what I’d naively expect.
In my survey, I asked people if they were cis or trans, and if they were trans I asked if they were on estrogen/testosterone, and for how long? The following graphs are drawn from the resulting data.
(disclaimer: I occasionally make mistakes, whether it be typos in my code, mistyped y-axis, etc. If I find any I will update this blog post with a fix. If you notice confusion, feel free to ask me to doublecheck various things, and I can also share the raw data I used in these graphs)
Dominance and Submission
This question forced people to pick - on a scale from dominant to submissive, where are you?
It seems like being on hormones is associated with pulling people slightly in the direction of their traditional gender, but only early on and only slightly, with the effect fading over time.
You might also notice that trans people, especially trans women, are already really far away from cis people’s scores. Trans women are more submissive than cis women, even before ever taking hormones.
That’s a compressed question though, on a single scale. What happens if we separate them out?
There seems to be a clear, brief bump for transmen - presumably some evidence in favor of the social pressure theory? Hard to tell, though. Regardless, it seems like both transmen and transwomen end up roughly where they started pre-transition, though the confidence interval is pretty wide.
Man, trans people and cis women really like being submissive. Like, friggin love it.
Consent
Not a huge shift, but it seems like both transmen and women become more interested in nonconsent over time on hormones.
You might be asking, “but what about age? Do you just have sexual changes when you age?”
Let’s check a narrow 22-26 aged sample:
Doesn’t seem like age is having any real impact here!
Bondage
The questions people got in the survey:
And, the results:
Trans men seem stable across bondages, showing a very slight but mostly even increase. While for transwomen, the heavier the bondage, the higher the interest over time.
It’s interesting that the key shift here seems to occur between 7months-2 years.
I find the gaps here somewhat meaningful; it’s not super common to see shifts like this in my data.
Sadomasochism
…Huh. Pretty big shift for transwomen, stabilizing at around 2-3 years, whereas for transmen there seems to be smaller movement.
Transmen and transwomen start out with almost identical gap between cis women and cis men, just shifted towards more interest - but over time they switch places, becoming slightly more like their current gender (relatively speaking.)
I would chalk up a lot of the pop in interest in sadomasochism and heavy bondage among trans women in the 7 months - 2 year period to be about coming to terms with having erectile dysfunction and then losing interest in more vanilla sex. Maybe a bit crass of an explanation but I think it fits with what I've seen.
Interesting question on the community explanation, I wouldn't be surprised if trans men tend to have communities more similar to their pre transition communities. Especially when it comes to dating it seems much more common for trans women to need to seek a new partner pool after transition than trans men since trans women are either dating straight women or gay men before transition, neither of whom tend to be particularly interested in dating women, while the butch/femme spectrum inside lesbianism can more easily accommodate trans masculinity. So maybe that additional insularity can explain some of the additional divergence among trans women.
1. Broadly saying "trans people are more ___ than cis people" is a common but very misleading way to distill these types of results. For a more obvious example, it's probably true that "men have [on average] slightly fewer fingers than women." However, while that framing suggests to the imagination a whole population of ~9.95-fingered men standing next to their 10-fingered girlfriends, the numbers actually just capture a vast majority of 10-fingered men plus a small sub-population of dramatic outliers due to violence or workplace accidents.
A more honest form of language would be "The average of values in respondents indicating transwoman identity is slightly higher than the average of values in respondents indicating ciswoman identity," or similar. Likewise, It'd be helpful to see distributions and standard deviations for many of these poll responses both on the cis and the trans side, maybe graphs that parse out trajectories for medians or for 20th vs. 80th percentile, etc.
2. These polls are great fun, but it absolutely contributes to the society-wide decline in information hygiene when you regularly:
- poll an extremely highly selected group of people (on Twitter/ Substack (aware of and following Aella (care enough about BDSM to answer a poll on it))),
- receive oceans of responses enriched with an unknown but likely large quantity of bots, trolls, culture-war-motivated lying, and deliberate multi-responses across alt accounts, and then
- headline your conclusions in ways that fully obscure all those reasons for nuance and epistemic caution, instead presenting your interpretations as broad Truths about group qualities across the general population (which, given the salacious nature of the polls, will go on to be reshared and vaguely internalized _as_ broad Truths by a huge proportion of the audience).
When journalists do this, it's understandable if gross. But as a rationalist-adjacent type, shouldn't you have higher standards for information handling?