Do trans women commit sex crimes at higher rates than women ... or men?

In my dissection of a couple of critical reviews of Jerry Coyne’s “Biology is not Bigotry” essay, I noted that there was one comment by the critics that is deserving of further discussion – the question regarding rates of sex crime commission by trans women.

This is what Coyne wrote:

But even here Grant misleads the reader. They [Grant uses the pronouns they, them] argue, for example, that “Transgender people are no more likely to be sexual predators than other individuals.” Yet the facts support the opposite of this claim, at least for transgender women. A cross-comparison of statistics from the U.K. Ministry of Justice and the U.K. Census shows that while almost 20 percent of male prisoners and a maximum of 3 percent of female prisoners have committed sex offenses, at least 41 percent of trans-identifying prisoners were convicted of these crimes. Transgender, then, appear to be twice as likely as natal males and at least 14 times as likely as natal females to be sex offenders. While these data are imperfect because they’re based only on those who are caught, or on some who declare their female gender only after conviction, they suggest that transgender women are far more sexually predatory than biological women and somewhat more predatory than biological men. There are suggestions of similar trends in Scotland, New Zealand, and Australia.

In his critique of Coyne’s essay, “Friendly Atheist” Hermant Mehta wrote about this claim by Coyne:

Coyne says trans women are more likely to be sexual predators… based on a stat published by a UK-based hate group focusing only on prisoners. And even that stat is wildly out of context. He writes:

A cross-comparison of statistics from the U.K. Ministry of Justice and the U.K. Census shows that while almost 20 percent of male prisoners and a maximum of 3 percent of female prisoners have committed sex offenses, at least 41 percent of trans-identifying prisoners were convicted of these crimes. Transgender, then, appear to be twice as likely as natal males and at least 14 times as likely as natal females to be sex offenders. While these data are imperfect because they’re based only on those who are caught, or on some who declare their female gender only after conviction, they suggest that transgender women are far more sexually predatory than biological women and somewhat more predatory than biological men.

Among the problems with those statistics? There are only a small number of trans women in prisons. (It’s a very tiny, very select sample size.) There are actually fewer trans women in prison relative to their existence in the general population, an argument that suggests trans women as a whole are more law-abiding than cisgender people. It’s also possible some of the “sex offenses” include getting arrested for consensual sex work.

Aaron Rabinowitz similarly wrote (On the “Friendly Atheist” blog) that:

Coyne’s misuse of data is so egregious one could easily get sidetracked with extensive debunking of just his empirical claims. To take the most harmful example, Coyne supports his bigoted claim that transgender individuals are far more likely to be sexual predators by citing what he calls “a cross-comparison of statistics from the U.K. Ministry and the U.K. Census,” which makes the data sound thorough and reliable. However, if you click the link he provides, it goes to an anti-trans activist group called Fair Play for Women, who in turn cite what they call a “freedom of information” request, but link to a BBC article that highlights several flaws with the data and concludes that “these figures are not yet a reliable reflection of the true numbers.” Crucially, the BBC notes that the study has a small sample size—only 125 transgender prisoners—and that it likely excludes a significant number of transgender prisoners for various reasons. It also does not clarify what behaviors classify a prisoner as a “sex offender.”

Coyne clearly knows this, since he briefly acknowledges that the numbers he’s citing are “imperfect” for some of the reasons mentioned, but that does not stop him from claiming “they suggest that transgender women are far more sexually predatory than biological women and somewhat more predatory than biological men”.

Are these criticisms of Coyne’s reasoning valid? Actually – to some degree – yes. The problem, though, is not so much the small sample size; even if the sample size were much larger, the percentage of trans inmates who have been convicted of a sex crime does not translate in any direct way into estimates of the percentage of trans individuals within the population who have committed sex crimes. In fact, even if 100% of trans inmates have been convicted of sex crimes, it could be the case that trans individuals are simply otherwise very law abiding, and therefore are never convicted of any OTHER kind of crime – even while having a lower rate of sex crime conviction than (non-trans) women or men.

In order to draw the conclusion that Coyne reached, what is needed is to have some reasonable estimate of the total number of trans women convicted of a sex crime (the numerator) and a reasonable estimate of the total population size of trans women (the denominator). That percentage would then be compared with the parallel measure for non-trans women and for non-trans men.

Interestingly, even though Coyne did not discuss those data, his article does actually include a link to a study the DOES provide the best data of this kind that we have. When Coyne wrote that “There are suggestions of similar trends in Scotland, New Zealand, and Australia.”, he included a link to an excellent article that analyzed data from a 2021 census in the UK. I will quote at some length from that article:

One of the few places we’ve been counting transwomen, transmen etc all along has been the prison population, for the fairly obvious reason that we (officially) have single sex prisons in the UK, and trans offenders need to go somewhere just like everyone else does.

You would have needed to suffer from a fairly extreme form of sensory deprivation not to have at least heard of or read the “scare stories” like Karen White, a trans-identifying male sex offender who was put in the women’s estate, and proceeded to sexually assault female inmates who could not escape White’s predations because, y’know, prison.

Trans activists have targeted prisons as a place where self-ID can be rolled out in practice before it was made law. A series of FoI requests of the Ministry of Justice had found an ever increasing number of trans prisoners, and there was a significant kerfuffle when Fair Play For Women (FPFW) – a feminist organisation that has done a lot of work on the clash between women’s rights and trans rights – reported that 48% of trans prisoners in the England and Wales system were incarcerated for sex crimes, while the percentage for women was miniscule.

BBC Reality Check infamously followed this up in 2018. I say “infamously” because when their initial enquiries proved that FPFW were broadly correct – 60 of 125 trans prisoners at the time were sex offenders – they provided a series of excuses and reasons why readers shouldn’t believe those figures.

Some of those reasons were not wrong. There very well may be trans prisoners in the system who are “closeted” for the stay and don’t show up in those stats. The hard number of 60, however, was reliable. It was the floor, the number we were certain of at that time. There were at least that many trans-identifying offenders in the prison system for sexual offences.

At the time, we simply didn’t know how many transwomen were in the country, so we relied on estimates by groups with skin in the game. In 2018 the Government Equalities Office, using some work done by trans advocacy group GIRES, produced a tentative estimate of between 200,000 and 500,000 trans people in the UK, which would suggest (assuming a roughly even distribution by sex) between 100,000 and 250,000 transwomen [a figure much higher than the later 2021 census data suggested].

The Prison Service also helpfully provided statistics for the number of prisoners by sex and offence type for the period in question.

At the moment in question there were 125 women incarcerated for sexual offences (mostly as accomplice to male offenders and mostly against children). The same spreadsheet shows 13,332 males were in prison for sexual offences

2017 was bang between censuses, so we only have an official estimate of the number of women in England and Wales at the time: 30 million (and 29.5 million men). Still, we now can do some basic mathematics to work out proportions of incarcerated sex offenders to wider population:

13,332 men out of 29.5 million = 1 in 2,212 men

125 women out of 30 million = 1 in 240,000 women

Given the tentative GIRES figures, there was a little ambiguity over the possible trans figures, so we’ll calculate both a min and a max.

60 out of 250,000 (max) transwomen = 1 in 4,166 transwomen

or

60 out of 100,000 (min) transwomen = 1 in 1,666 transwomen

So, it seemed, transwomen were somewhere between a bit more likely than men to be imprisoned as sex offenders and a bit less likely. In any case they were, many, MANY times more likely to be sex offenders than actual women.

But then a funny thing happened.

Partially due to COVID, the number of incarcerated sex offenders of both sexes went down. As of 2021, men dropped as of 2021 from 13,000 to around 11,000 and women dropped from 125 to around 110.

So now we have hard counts for sex offenders in prison as of March 2021: 11,660 men and 103 women, plus 92 transwomen.

11,660 men serving time for sex offences out of 29.5m = 1 in 2530 men

103 women serving the same time out of 30.4 million = 1 in 295,000 women

92 transwomen serving the same time out of 48,000 = 1 in 522 transwomen

That suggests transwomen are five times more likely than other men, and 566 times more likely than women, to commit sexual offences. That’s not a sample, not a poll, not a survey, not a representative focus group. It’s a hard count.

Given these data, it is certainly challenging to claim that trans women are not MUCH more likely than biological women to be convicted of, and incarcerated for, a sex crime. But I’ll give it a try anyway.

Perhaps – when trans women do commit a sex crime, they are MUCH more likely than biological women who commit a sex crime to be charged with the sex crime violation. I’m not sure why that would be the case, but it is not totally beyond the realm of possibility. Or perhaps trans women are more likely than biological women to receive a prison sentence when they are convicted of a sex crime. Again, I have no reason to think that IS the case, but if it were true, then it COULD be the case that the data presented above are accurate AND that trans women are not more likely than biological women to actually commit a sex crime.

It’s possible that there are data that could be brought to bear on these questions – but I’m certainly not aware that any such data exist. And so – while acknowledging that these two possibilities cannot be ruled out with ABSOLUTE certainty, it does seem quite farfetched to claim that either of these possibilities could be of such a magnitude that the effect could explain the data described above. When the estimates of the rate of incarceration of trans women for sex crimes relative to the rate for biological women ranges from 50 times more likely (at the low end) to 550 times as likely (at the high end), the explanatory challenge for anyone who wants to claim that there is really no difference in the rates at which members of these groups commit sex crimes is a particularly daunting one.

Having said that, it is at the same time important to note that these are essentially correlational data, and as such do not indicate what it is about members of the trans woman population that makes them so much more likely than biological women to commit a sex crime. The much higher rates at which biological males commit sex crimes relative to biological females might provide an important clue about the reason; given that trans women are individuals who were born male and, in most case, developed as males at least through their teen years, it may be that trans women may retain whatever it is biologically that is responsible for biological males transgressing at so much higher rates than biological women. Note that this hypothesis does not claim that ALL, or anything close to all, trans women are likely to transgress; just as most biological males do not commit sex crimes, one would not expect more than a very small percentage of trans women to commit a sex crime even if this explanation is correct.

Alternatively, there may be a variety of other, non-biological, factors associated with being a trans woman that may explain why trans women commit sex crimes at much higher rates than do biological women. There may, for example, be lifestyle factors, or the phenomenon may reflect the effects of discrimination, or the nature of the sexual and/or romantic relationships that trans women form.

Regardless, it is also important to note that the focus on rates of sex crimes committed by trans women is a bit of a distraction from the central issues regarding women-only and girl-only spaces. Fear of being assaulted is not the sole reason why most women prefer to be able to use a women’s bathroom in public, prefer to be able to use a women-only locker room, prefer as a rape counselor someone who was born a female, and of course, that is not a factor at all in the reason why we have women’s sports. Even if it could be conclusively shown that trans women were no more likely to assault anyone (sexually or otherwise) than biological men, girls and women would still want, and would still deserve to have, girl-only and women-only spaces and activities. Even if it could be conclusively shown that trans women were no more likely to assault anyone (sexually or otherwise) than biological women, girls and women would still want, and would still deserve to have, girl-only and women-only spaces and activities.