Since January 2025, about 360 federal data collection surveys in the United States have removed at least one question about sexual orientation or gender identity. This followed an executive order instructing federal agencies to recognize only two sexes and to strip what it calls “gender ideology” from public forms. The order’s effects have manifested across surveys, health systems, and websites, and what has disappeared represents more than a decade of technical work and measurement, now undone through administrative changes.
None of this is accidental or confined to one country. Anti-gender movements both oppose particular rights and contest which realities are legitimate enough for the state to name. When gender diversity is presented as ideology, certain lives become politically unspeakable—and populations framed as such do not need measurement. This erasure can take place without legislation and can be enacted easily and without fanfare: a variable dropped from the next census, a health system redesigned without a gender identity field, a ministry quietly dropping a disaggregation requirement.
However, forms of statistical inclusion now at risk, from the gender identity question in Argentina’s 2022 census to systems documenting more than 5,000 reported killings of trans people worldwide since 2009, reflect years of human rights advocacy. Statistical inclusion was never a gift; it was won slowly and against resistance. In most countries, conversations around statistical inclusion have barely begun, and incomplete gains are vanishing before they become stable practice.
Invisible discrimination, real losses
Deleting a question does not look like repression. There is no arrest, no office shutdown, no victim who can testify at a hearing. The damage is slow and dispersed; it will only become visible years from now, when, for instance, someone asks how many trans people died in the 2020s and finds no information. That is what makes this method of discrimination so useful: it costs little, it is hard to litigate, and it can always be excused as routine administrative work. The losses it produces, however, are entirely concrete.
The first loss is evidence. The machinery we use to hold states accountable runs on data produced by states themselves. Treaty body reviews, the Universal Periodic Review, and the disaggregation rules of the United Nation’s Sustainable Development Goals all depend, at least in part, on governments reporting on their own populations. In 2019, the UN Independent Expert on sexual orientation and gender identity warned that this information was incomplete in many countries and that policymakers were making decisions in the dark. A state that has stopped counting can now report, with a clear conscience, that it holds no information about discrimination against a specific population. This is institutional silence: the state simply stops producing the evidence needed to protect them.
The second loss is money. Public budgets follow evidence. A need that does not appear in official statistics is not costed, not planned, and not funded, even when governmental will exists. A ministry cannot direct resources toward a population of unknown size and conditions. Statistical absence not only weakens advocacy from outside the state, it also disables policy from within.
The third loss is prevention. Statistics are how a society notices a pattern before it becomes a disaster. Population-wide averages can hide a rise in violence against one group or higher mortality concentrated in one population. Looking at data separately for different population groups can reveal otherwise-hidden patterns and make earlier action possible. Without population-specific data, every case stays an isolated tragedy.
The fourth loss is memory. Data systems are how a country maintains knowledge about inequality across governmental change. If one administration can add a category to data collection and the next can delete it without explanation, each government starts from zero, inequality is left without a history, and no one twenty years in the future will be able to establish what is really happening now.
But the deepest loss is not about usefulness. Being named accurately by the state is a right in and of itself, not just a means of obtaining other rights. Death certificates that erase how people lived damage mortality statistics, but they also affect mourning, family recognition and collective memory. Registries are places where a society decides who existed, and to be unnameable there is an exclusion that no later policy can fully repair.
Statistical justice: a rights-based approach to data collection
None of this means the answer is simply more data. In 2011, Nepal became the first country to include a third gender category in its census, but activists considered the resulting data unrepresentative and demanded changes to later census design. Where criminalization or weak data protections exist, a registry can also expose people to surveillance and violence. Trans and LGBTIQ+ organizations have long understood this tension, and their insistence on consent, self-identification, and control over data use is what separates a rights-based approach to data collection from a merely administrative one.
Together, the losses inflicted when the state undermines statistical inclusion fundamentally violate the right to be recognized in government systems and for official data to reflect the true diversity of societies. My colleagues and I call this statistical justice. This is not a trans or gender issue alone. Migrants, Indigenous peoples, Afro-descendant communities, and people with disabilities have all been counted late, partially, or through categories imposed on them. Methods of erasure move easily among populations.
Naming the right to be recognized is the easy part. Human rights advocates know how to challenge a law, document a case, and bring a government before an international body. But we have much less practice defending the apparatus that decides what will count as evidence in the first place. These data systems are rights infrastructure, too. They require the same defense because they determine what a state can later be asked to answer for. The demand is not technical but democratic: a way of ensuring that reality does not depend only on those with the power to name it.