A university vice-chancellor has warned that the “quiet disappearance” of sex as a recorded variable from academic research risks weakening policy analysis, comparing pressure to drop “politically uncomfortable” classifications with Stalin’s persecution of statisticians in Soviet-era Russia.
Speaking at an event at City St George’s, University of London to discuss last year’s Sullivan Review, the institution’s president, Sir Anthony Finkelstein, said it was vital that researchers continued to use established categories for sex in their statistical work.
“In recent years, there’s been growing unease about recording sex as a basic variable, and pressure – sometimes explicit and sometimes ambient – to replace it with or collapse it into gender identity,” Finkelstein said.
“Biological sex” was a category that was important to maintain, he argued, because it correlates strongly with major social outcomes. “It matters because sex correlates strongly with capital, health, mortality, fertility, violence, incarceration, employment, caring responsibilities,” he said. Addressing the methodological problem at the heart of the debate around population-level data, he added that researchers need distinct variables if they want to test the respective effects of sex and gender identity: hybrid “sex/gender” questions tend, in practice, to elicit sex from most respondents, making it harder to disentangle causes and draw reliable comparisons over time. “Remove it, blur it or replace it, and whole domains of policy analysis become weaker and less reliable,” he warned.
That did not mean research should not also focus on gender identity, Finkelstein added, simply that it was “a different kind of variable”. Gender identity is “subjective, fluid and not comparable across time in the way that a census requires”. It may belong in “specialist surveys and research instruments”, he suggested, “but not as a substitute for a foundational demographic category”.
Finkelstein, a software engineer who previously served as chief adviser on national security to the UK government before joining City St George’s in 2021, said he had “grave concerns” about the findings of the Sullivan Review, including “the barriers to research [and] institutional caution” related to collecting data on sex.
Authored by Professor Alice Sullivan of University College London, the review is published in two parts. The first sets out proposed principles for collecting and reporting sex and gender data in publicly funded research, and concludes that the meaning of sex is no longer stable in administrative and major survey data. Part two, meanwhile, turns to a survey of academics, recording accounts of institutional resistance to collecting data on sex. It documents, among other things, reported pressure from funders and academic publishers to adopt language “rooted in gender-identity theory”.
Multiple respondents also described research ethics committees objecting either to sex-based questions or to plain references to sex. One respondent reported that an ethics application was not approved because it opened by stating an intention to examine views on “when males should be able to compete in the female category in athletics”. This was described as “misgendering”, they said, and they were told they would need permission from their faculty dean to proceed, as well as input from an EDI team on the wording of both the survey and the ethics application.
Reflecting on the findings, Finkelstein said they pointed to a gradual retreat from recording sex “not because it lacks relevance, [but] because it’s become politically uncomfortable”.
He went on to draw parallels with the Soviet Union’s suppressed 1937 census, which produced a population figure materially below what the regime had expected to see – a discrepancy widely explained by the famine years induced by Stalin’s forced collectivisation of agriculture. Senior statistical officials who naively prioritised the dataset over ideology were subsequently arrested and, in some cases, executed, including Ivan Kraval, who headed the central statistical administration.
Researchers and statisticians, Finkelstein said, should not allow political or social pressure to reshape basic categories.
“We don’t live in a system that executes mathematicians,” he said. “But we do live in systems that reward conformity, discourage awkward questions and quietly reshape categories so that certain answers no longer appear.”
“There is a long-term danger of governing the degraded data and discovering too late that we are not describing the world we inhabit,” he warned. “A society that can’t better count itself will eventually lose itself and its ability to govern itself.”
This article forms part of the Committee for Academic Freedom’s monitoring of developments in UK higher education. To receive monthly updates and analysis on cases affecting academic freedom, click the link and subscribe to our newsletter.