Yesterday, we published “When a Trans Person Commits a Crime, We’re All Put on Trial”, an article about a familiar double standard: when a transgender person commits a terrible crime, anti-trans voices often treat that individual as evidence against the entire transgender community.
The point was not that transgender people never commit crimes. In fact, we explicitly said the opposite. Trans people are human beings. Some human beings commit horrible crimes. Hold those people accountable.
The argument was that one person’s actions don’t establish something about millions of unrelated people.
Then something rather predictable happened.
In responses to the article’s social media posting, gender-critical and anti-trans commenters began sharing a graphic that supposedly proved the point wrong. According to the image, the sexual-offending rate for women is 3 per million, the rate for men is 395 per million, and the rate for transgender women is an astonishing 1,916 per million.
Checkmate, apparently. Except there’s a problem.
That isn’t actually a sexual-offending rate.
The graphic takes real numbers from British government sources, combines statistics measuring different things, performs some basic division, and gives the result a label the underlying data don’t support.
Welcome to another episode of Statistics: Please Read Past the Meme.
Where Does 1,916 Come From?
The best way to understand what’s wrong is to follow the numbers backward. In January 2022, the UK government answered a parliamentary question concerning transgender prisoners in England and Wales.
According to the Ministry of Justice, there were 146 transgender women in prison on March 31, 2021. For purposes of that particular dataset, the government defined these as prisoners who were legally male and identified as female. The figure excluded some prisoners with Gender Recognition Certificates.
The Ministry of Justice reported that 87 transgender women in the men’s prison estate had a conviction for at least one sexual offense associated with their current imprisonment. Fewer than five transgender prisoners were housed in the women’s estate, with fewer than five having a sexual-offence conviction.
Notice what we’re talking about. Prisoners.
Not transgender women randomly selected from the population. Not every transgender woman arrested during 2021. Not every transgender woman convicted during 2021. Not the number of transgender women who committed sexual offences during 2021.
People who were sitting in prison on a particular date.
Now we need the other half of the equation.
Britain’s 2021 Census included a voluntary gender-identity question for people aged 16 and older. The Office for National Statistics reported approximately 48,000 respondents identifying specifically as transgender women.
Take approximately 92 incarcerated transgender people with sexual-offence convictions, divide that by 48,000, multiply by one million, and you get roughly: 1,916 per million.
So the arithmetic isn’t some mysterious fabrication. The sleight of hand happens when the result gets labeled a “sex offending rate.”
Being in Prison Is Not the Same as Committing a Crime That Year
Suppose someone committed a sexual offense in 2015 and received a 12-year prison sentence.
That person would potentially appear in a prison snapshot taken in 2017, 2018, 2019, 2020 and 2021.
Did that person commit another sexual offense during each of those years?
Obviously not.
That’s one of the problems with using the number of people incarcerated at a particular moment to calculate something presented as a population offending rate.
Prison populations accumulate people convicted across many different years. Longer sentences also increase the likelihood that someone will appear in a snapshot.
Meanwhile, someone convicted of an offense but given a noncustodial sentence wouldn’t appear. Someone released before the snapshot wouldn’t appear. Someone whose crime wasn’t reported wouldn’t appear. Someone who wasn’t prosecuted wouldn’t appear.
The Ministry of Justice itself was remarkably clear about what its numbers represented. It said it could provide figures only for offenses that led to an individual’s current imprisonment. Calling it an “offending rate” doesn’t magically transform it into one.
There’s Another Problem With the Famous 48,000
The denominator deserves scrutiny too. The 48,000 estimate really does come from Britain’s 2021 Census. But there’s some important context that wasn’t available when the viral graphic appeared.
In 2024, the UK’s Office for Statistics Regulation removed the Census 2021 gender-identity estimates’ status as accredited official statistics. They are now classified as official statistics in development.
Then, in March 2025, the Office for National Statistics published additional guidance about uncertainty in those estimates.
ONS found patterns suggesting some respondents may not have interpreted the gender-identity question as intended, particularly among people with lower English-language proficiency. It says the overall census estimate can still provide a broad indication of the transgender population’s size, but specifically urges caution when using individual gender-identity subcategories.
That’s relevant because “48,000 trans women” is precisely one of those subcategories.
None of this means the true number is necessarily higher or lower by some particular amount. We don’t know that.
It means pretending 48,000 is a perfectly precise denominator from which we can calculate a precise criminal-offending rate to the nearest person per million is difficult to defend when the agency responsible for the census is itself warning researchers about uncertainty.
The Graphic Doesn’t Even Reproduce All the Prison Data Cleanly
There’s another clue that we’re looking at advocacy rather than a serious criminological analysis.
The viral graphic lists 103 incarcerated female sexual offenders, while the Ministry of Justice’s June 2021 prison-population dataset is the source being cited for the male and female prison numbers. The underlying government release contains detailed prison population tables by sex and offense group.
Independent examinations of the graphic have noted that the June 2021 dataset reports 119 females serving immediate custodial sentences for sexual offenses, rather than the 103 shown in the meme.
That discrepancy isn’t even the central problem. Changing 103 to 119 wouldn’t rescue the methodology. It simply reinforces why a social media infographic shouldn’t be mistaken for a population-level criminology study.
This Is Exactly What Yesterday’s Article Was Talking About
And this is where things get almost painfully ironic.
Yesterday’s TransVitae article argued that transgender criminals should be treated as individuals rather than representatives of transgender people generally.
The article explicitly said: “Transgender people commit crimes. There. That wasn’t difficult.”
It went further. If a transgender person murders someone, prosecute the murderer. If a transgender person sexually assaults someone, prosecute the offender. Trans equality doesn’t require pretending transgender people are incapable of doing awful things.
The problem begins when individual wrongdoing becomes collective guilt.
The article also made another point that suddenly seems particularly relevant: if someone claims transgender people represent an unusual criminal threat, the appropriate response isn’t to ask whether examples of transgender criminals exist.
Of course they exist.
The question is whether reliable comparative evidence demonstrates the claimed population-level pattern.
In response, commenters produced a graphic that takes fewer than 100 incarcerated transgender people and attempts to extrapolate from them to transgender women generally.
You couldn’t have scripted the demonstration much better.
The Answer Isn’t Pretending Trans Criminals Don’t Exist
There is a tempting but unnecessary response to statistics like these: trying to prove transgender people never commit sexual crimes or insisting every unfavorable statistic involving transgender people must somehow be false.
Don’t. We don’t need that argument.
There were transgender prisoners convicted of sexual offenses in Britain. That is a fact. Their crimes shouldn’t be minimized because they’re transgender.
But their existence doesn’t establish that transgender women have a sexual-offending rate of 1,916 per million, much less that transgender women as a class are inherently dangerous.
Those are separate claims requiring separate evidence.
If researchers want to determine whether transgender women commit sexual offenses at higher, lower, or similar rates to other populations, that’s a legitimate research question.
But answering it requires an actual research design.
Researchers would need comparable populations, consistent definitions, matching time periods, and comparable measures of offending. They would need to decide whether they’re measuring arrests, convictions, self-reported offending, incarceration, or something else. Ideally, they’d account for variables such as age and other characteristics associated with differences in criminal offending.
What they wouldn’t do is take a snapshot of roughly 90 people already selected into the prison system, divide it by a census population estimate, and announce that they’ve discovered the sexual-offending rate of transgender women.
That’s not how this works.
The Bottom Line
You don’t need a dissertation to respond when this thing inevitably appears underneath another Facebook post.
Ask one question: “What does this number actually measure?”
Because once you follow the sources, the answer isn’t “the rate at which transgender women commit sexual offenses.”
It is approximately the number of transgender prisoners with certain sexual-offence convictions who were incarcerated at a particular point in time, divided by an estimated transgender female population taken from a different government dataset.
That’s considerably less meme-worthy. And it brings us straight back to yesterday’s point.
Transgender people don’t need to be protected from accountability. A transgender person who sexually assaults someone should be held responsible for that crime. Victims deserve justice regardless of the perpetrator’s gender identity.
What transgender people shouldn’t have to accept is being treated as a collective defendant whenever someone finds a transgender criminal.
Cisgender people receive that individuality so routinely that nobody even notices it.
We’re asking for the same thing.
And if someone wants to make the much larger claim that transgender women are disproportionately likely to commit sexual offenses, the evidentiary standard isn’t particularly complicated: Bring population-level research that actually measures population-level offending.
Until then, putting “1,916 per million” in giant letters doesn’t turn a prison snapshot into something it isn’t. It just makes the font bigger.
