No, there aren’t 15,000 transgender troops in the US military

How Flawed Data Skews the Debate

By Alexander Raikin

The telltale sign of a bad statistic is how often it is invoked. Last week, the British newspaper The Times claimed that Donald Trump is planning an executive order on his first day back in office to remove all transgender members of the US military. More precisely, all 15,000 transgender soldiers.

Never mind that Trump’s campaign denied that any decision was made. It is rather that six times in the same 900-word article, the same statistic reappeared. “There are believed to be about 15,000 active service personnel who are transgender,” The Times stated. The question is: who believes it?

Repeating the same figure doesn’t make it true. Yet there is a good reason for why The Times declined to provide a source or a link for this statistic. It is simply outdated, unreliable, and proven to be false. In more technical language, it is nonsense.

The origin of this statistic is from a 2014 report—a decade ago, for those keeping score—by the UCLA’s Williams Institute based on 2008 polling in the National Transgender Discrimination Survey. Moreover, it is an activist survey, as if the name wasn’t enough of an indication: it is a routine study conducted by the National Gay and Lesbian Task Force and the National Center for Transgender Equality.

Yet let us set aside any potential bias concerns and focus on the main claim: despite the ban on transgender soldiers openly serving at the time, which was formally enforced by Trump in 2019 and cancelled again by Biden in 2021, this survey claimed that there are 15,500 (an extra 500) active-duty transgender soldiers and an astonishing 134,300 transgender veterans, which corresponds to 21% of the entire transgender population in the United States, over double the rate of the rest of the US population.

In other words, we apparently militarized our transgender population. If it seems unbelievable, it is because it is. Why wasn’t this 2014 report updated with more recent polling? Perhaps because the most recent complete survey, in 2021, was that only 0.3% of transgender respondents were active-duty soldiers. Yet even that is too high.

The problem with modern polling is that polling is expensive, especially when sampling a small subsample of the population. While it is difficult enough to sample enough respondents for a representative sample of the national population, it is much more difficult to sample a much smaller and varied subsample, especially one that is geographically distributed across the United States.

The solution by pollsters is to fudge the methods. The first is called a “convenience sample”, which is to use non-random sampling—the National Transgender Discrimination Survey is even less representative than most polls because it was conducted by sending the survey not through any random methods but by distributing the study through organizations and listservs where transgender people could be found.

The second solution is to manipulate a random poll. The Department of Defense issues a regular survey on “Workplace and Gender Relations Survey of Active Duty Members” to a representative sample of their population, which includes questions to service members on whether they identify as transgender, and whether their sex listed on their birth certificate is different to their current sex. They then adjust their responses by gender—only male and female—to reflect the total pool of service members.

In 2019, the Palm Center, a think tank, claimed on the basis of this survey that there are 8,980 active-duty transgender troops, since 0.7% claimed they were transgender. Yet the same 15,000 figure of service members was maintained by claiming that an additional 4,300 transgender service members were in the reserves.

Unfortunately, the Palm Center neglected to provide any error margin in the original statistic, which was plus or minus 1%. In plain English, it means that the number of transgender respondents to the survey was almost a statistical zero—or it could be double the purported number.

This error rate was high precisely because the Department of Defense’s polling sample is manipulated to reflect the total gender distribution, meaning that the small subgroup of polled transgender soldiers is likely not reflective of the real number of transgender soldiers.

If we apply the same methodology by the Palm Center to the most recent study in 2021, right after Biden revoked Trump’s transgender ban, we find that 1% of female and 0.3% of male respondents claimed that they were transgender. Yet this will mean there are 34,000 transgender active-duty soldiers, which even supporters don’t claim.

There is an easier approach to identify the number of active transgender soldiers. It won’t require any random polling. Instead, we can look at the number of active-duty service members who are diagnosed with gender dysphoria.

In 2024, according to the Pentagon, there are only 3,700 US active duty diagnosed with gender dysphoria. “This includes 1,240 Army soldiers, 1,046 Navy soldiers, 1,024 Air Force airmen and 278 Marine Corps,” reported the Daily Mail, and which accounts for only 0.3% of all military personnel, half the rate of the transgender population in the US.

While it is possible that more service members identify as transgender than those that seek out medical care, those service members won’t be affected by any “ban” on transgender soldiers. Trump’s previous ban was based solely on those diagnosed with gender dysphoria.

After all, the First Amendment protects all personal beliefs, including a belief that someone born as a biological man is a woman. It doesn’t protect the rights of soldiers to serve if they are diagnosed with a medical condition. On that basis, there are only 3,700 active duty transgender soldiers—a fifth of the purported number that The Times claims.

Alexander Raikin is a Visiting Fellow in Bioethics and American Democracy Program at the Ethics and Public Policy Center. His research focuses on the dignity of human life and end-of-life issues, especially on its impact on the field of medicine and broader ethical questions of social belonging.