Some data matters
The problem with mashing up sex and gender identity data: UK and NZ
1. The 2021 Census of England and Wales
The 2021 Census of England and Wales attempted to collect national data on gender identity for the first time. When Dr. Michael Biggs, from the University of Oxford, saw the results, he was struck by something odd. Areas with high percentages of Muslims were found to have more per-capita transgender people than commonly recognised LGBTQI+ centres like Brighton. Further, according to the census findings, one in every 67 Muslims identified as transgender.
Biggs cross-checked the findings against other data sources relating to gender identity and found discrepancies. He concluded that poorly worded and confusing census questions were misinterpreted by respondents and had yielded false results.1 A conclusion that was bolstered by his finding that, according to the census, 0.4% of those who were in the category of having the most proficiency in English were transgender compared to 2.2% of those categorised as least proficient.
The census asked two questions about sex and gender. The first was “What is your sex?” The answers were confined to male or female. The question flagged that a question on gender identity would be asked later.
Later in the questionnaire came the gender identity question.
Is the gender you identify with the same as your sex registered at birth?
(This question is voluntary.)
Biggs pointed out that this question is confusing for those with low literacy, those without fluent English, and those unfamiliar with the concept of gender identity. It also works from the premise that everyone has a gender identity. Biggs noted that assuming everyone has a gender identity is akin to assuming everyone believes in God. Biggs also documents how the wording did not originate within the Office for National Statistics, the agency responsible for creating the census.
ONS, by contrast, selected a question that had been invented by activists in Press for Change in 2007 and then adopted by Stonewall—despite this question having been tested and found too confusing for non-transgender people. Not coincidentally, perhaps, ONS was simultaneously paying Stonewall to train its staff and to audit its policies.2
2. The Sullivan Review
It must have been hard to ignore Biggs’ findings. In February of 2024, amidst concerns that the conflation of sex and gender identity had undermined data integrity, the UK government commissioned Professor Alice Sullivan to review the UK’s handling of data, statistics and research on sex and gender. In March of this year, Professor Sullivan’s review was published.3 The review said it was useful to collect information about gender identity, but it shouldn’t come, as it had, at the expense of robust sex data.
The Government has a strong interest in promoting high-quality data on sex, both in its role as a funder of research and as a producer and user of statistics. Accurate record keeping is also vital for operational purposes, for safeguarding and, within the healthcare system, for patient safety and care. (Sullivan Review)
Reviewing 800 datasets, Sullivan found that the conflation of sex and gender identity meant sex data wasn’t stable, data was getting lost, definitions were murky, and questions didn’t make sense. Data loss creates risk for individuals, particularly in health and social care, and particularly for children.
Sullivan made 59 recommendations. Among them, and these become important later in this piece:
3. The default target of any sex question should be sex (in other words, biological sex, natal sex, sex at birth). Questions which combine sex with gender identity, including gender identity as recognised by a Gender Recognition Certificate (GRC) have a mixed target. Sex as a biological category is constant across time and across jurisdictions, whereas the concept of ’legal sex’ subject to a GRC may be subject to change in the future and varies across jurisdictions. Using natal sex future-proofs data collection against any such change, ensuring consistency. (Sullivan Review)
And:
6. The word ‘gender’ should be avoided in question wording, as it has multiple distinct meanings, including:
a synonym for sex
social structures and stereotypes associated with sex
gender identity
If a question targeting gender identity is worded as a question on gender, this is likely to mislead many respondents. Questions on sex have also often been labelled as ‘gender’. Change in the use of the term ‘gender’ means that it is important that questions on sex are labelled explicitly as such. (Sullivan Review)
After the Sullivan Review’s release, UK Health Secretary, Wes Streeting, said on X:
The Sullivan Review underlines the importance of recording biological sex, not just for research and insight, but also patient safety. Doing so does not prevent us from recording, recognising and respecting people’s gender identity where these differ. We’ll act on findings. (Wes Streeting)
Pink News reported that the groups TransActual and the Feminist Gender Equality Network were urging the government to reject Sullivan’s findings.4 Pro-child transition group Mermaids said they were deeply concerned and that:
“The recommendations of this review are an attempt by a small group of campaigners to deny the existence of trans and intersex people”. (Mermaids5)
3. The Stats NZ data standard
Sullivan’s and Biggs’ criticisms of official sex and gender data collection are equally relevant in New Zealand.
Working under the chief statistician, the agency responsible for the New Zealand census, Stats NZ, is tasked with:
“ensuring that high-quality, impartial, and objective official statistics are produced relating to New Zealand to inform the public and inform decision making”. (Stats NZ Statement of Strategic Intent, 2023-2028)6
Rather than sticking to that knitting, which is in the law that mandates its activities, Stats NZ has chosen to align itself with the policy ambitions of lobby groups.
In 2019, Stats NZ announced it was reviewing its sex and gender data standards. Statistical and Data Standards become mandatory for all New Zealand agencies collecting official data. Stats NZ appointed an advisory group to help. The advisory group consisted of representatives of transgender and intersex advocacy groups, academics and a staff member from the Human Rights Commission. All members hold a belief in the contested theory of gender identity. That is, they believe that everyone has a gender identity—inner feelings about themselves and their sex— and that this gender identity is more important than sex. They also all agree that the human rights of transgender people depend on universal validation as members of their chosen gender. The advisory group did not include representatives from women’s groups nor groups which accept long-established scientific understandings of sex.
Between July and August of 2020, Stats NZ opened the data standard up for public consultation. When it called for submissions, Stats NZ did not mention the need for “high-quality, impartial and objective data”. It talked instead of best practice in collecting data on this topic, implying exceptionalism and not mentioning who determines best practice. Instead, the call for submissions talked of the need to adequately reflect transgender and intersex people, and evolving societal and cultural understanding and language. It also discussed diversity and inclusion.7 It reflected a shift from impartiality to a form of conscious diversity activism.
“Our job at Stats NZ is to reflect society in all its diversity. We are committed to improving the way we record and present information on sex and gender identity to ensure it is inclusive and meets information needs,” Ms Milicich said.
Seventy to eighty per cent of submitters agreed with all aspects of the new Data Standard. In 2021, Stats NZ published and adopted the Data standard for gender, sex and variations of sex characteristics.8 The standard states that gender, not sex, should be the default question in relevant surveys. It claims this aligns with self-determination from a human rights perspective and promotes the respect and inclusion of all people. The extensive glossary, lifted almost entirely from trans lobby groups, provides a long-winded, circular and vague definition of gender and related terms.
Sex and gender are listed as separate concepts. Strikingly, Stats NZ has adopted the lobbyists’ belief that not just gender identity can change over a lifetime, but also sex.
4. The NZ 2023 census
In designing the 2023 census, Stats NZ followed its own gender by default principle. As a result the New Zealand census is even more confusing than the England and Wales census that produced such implausible results.
The gender question: What is your gender?
In the individual questionnaire, Question 3, the first question that mentioned sex or gender, asked about gender. The options were male, female or other.
In the online form, opening up the help box (shown open below but closed by default) revealed that gender meant gender identity. The help text reads:
“Gender refers to a person’s social and personal identity, as male, female or another gender (or genders) that may be non-binary. A person’s sex at birth may differ.”
The paper form showed some more information.
This question is misleading in several ways.
Placing the gender, aka gender identity, question at the top of the form nearest other questions about primary attributes, and before biological sex, gives it unnecessary weight.
As pointed out by Sullivan, gender used in this way is misleading. Given that gender is a long-time synonym for sex, many people will assume it means sex.
The notion of a substantial inner gender identity—one that defines you and may or may not align with your physical sex—is a proposition put forward by gender identity activists. For many, and I suspect most of us, the question doesn’t make sense or reflect our biology-informed worldview. There is, however, no option to bypass the question.
In its post survey data crunching, and before producing reports, Stats NZ did not report on the numbers that left the gender question unanswered. Instead it derived answers from how people answered the question about “sex at birth”. Thus, many who do not consider that they have a gender identity, are misrepresented in census reports.
The language used in the supporting text is based on a single ideological perspective rather than a widely accepted view of the world. This would be jarring for many respondents.
The sex question: What was your sex at birth?
The sex question followed the gender (identity) question.
The sex question is also misleading:
Having established a difference between sex and gender in the gender question, the concepts are now conflated. By asking "What was your sex at birth?" implies that the characteristic of sex is changeable over time.
Survey questions should be clear and straightforward. Sex is a very simple matter. The lengthy supporting text, adds word clutter and confusion.
Supporting text should provide the clarity respondents might need to accurately answer a question. The supporting text provided here instead reads as an attempt to educate respondents in how to think.
More points on NZ census 2023
Stats NZ did test the proposed design of the census forms with a sample of users before launching.9 Particular emphasis was given to including members of the LGBTQI+ / rainbow communities in multiple rounds of that testing. On the gender question, it noted “Some respondents, especially those who were older, had difficulty with the concept of gender.” No numbers are given. It does not appear that this finding prompted any refinements to the design of the forms. An unusual decision.
Stats NZ stated of the inclusion of gender: “This will help inform better decision-making for and about gender minority groups and will allow more people to see themselves in the data generated from the census.” There is no discussion of possible or perceived risks to questionnaire design and data robustness collected about the rest of the population.
As discussed, the question about gender identity is given unnecessary substance by removing the word identity and placing it before a question about sex. Interestingly, the question on sexuality is treated in the opposite manner. Question 27 reads “Which of the following best describes how you think of yourself?” The options provided are heterosexual/straight; gay or lesbian; bisexual; another identity and; prefer not to say. While gender has been elevated to the realm of an enduring and important characteristic, sexuality has been degraded to an identity rather than a description of attraction or orientation.
Accurate data about sex and about people who identity as transgender are worthwhile collecting. This benefits everyone in New Zealand. But the census needs to make sense to everyone. There are some simple, plain language and usable solutions that could be leveraged for the next census:
Restore the term sex rather than sex at birth. Provide additional text for those who are unsure what this means.
Put the sex question, which applies to everyone, above any questions about transgender (or other gender) identification which applies to a minority.
Ask participants if they identify as transgender (or use an appropriate alternative term). This question might sit well next to questions about religion.
Final thoughts
The flaws of the UK’s sex and gender data collection, as beautifully illustrated by the England and Wales census, have been met by Stats NZ and raised some. It’s reasonable to predict that our census might have produced similarly spurious data as that of England and Wales. Someone should look into it.
Stats NZ is mandated to produce impartial, objective and accurate statistics10. It has pivoted away from this. On the matter of sex, well-meaning attempts to promote inclusion, prioritise diversity and promote human rights have subsumed the agency’s primary purpose.
It sought and took advice from groups and individuals who believe that the abandonment of scientific understandings of sex, the rewiring of language and the re-education of the general population to adopt its philosophies are synonymous with the human rights of gender minorities.11
In our 2023 census, in its haste to heed what it considered important advice, Stats NZ abandoned the common-sense, straightforward question on sex—a question that had been asked continually since the 1851 census. Stats NZ risked the integrity of a precious data set.
It was serious overreach and it went way beyond adequately reflecting transgender and intersex people. It stopped adequately reflecting the world most of us understand ourselves to live in. It introduced needless ambiguity and confusion into the census. It repurposed our country’s most detailed and ambitious national social survey to affirm the identities and worldviews of a tiny minority. 12
My friends and I have been involved in various forms of human rights activism for decades. Stats NZ does have a role in helping us progress human rights. It can do its job properly and go back to providing us with robust, accurate and detailed information we can trust. It can help tell us what’s really happening in our country and among our people—who’s doing okay, who’s struggling and maybe, where it can, why. When we properly understand the world as it is, not as a few people want it to be, we can figure out how to make it better.
In actuality, upholding human rights has never been about forcing other people to adopt a belief system. That would be, in fact, a breach of Articles 18 and 19 of the Universal Declaration on Human Rights. If doubt persists about whether one group’s human rights can be seen to depend on another group’s loss of freedom of opinion or belief, the final article helpfully clarifies Nothing in this Declaration may be interpreted as implying for any State, group or person any right to engage in any activity or to perform any act aimed at the destruction of any of the rights and freedoms set forth herein.
Worth noting that not all transgender people nor those with differences of sex development (sometimes called intersex) hold the particular views promoted by the lobby groups who provided advice to Stats NZ.
Thank you. This is very worrying and brings all stats in the last 5 years or so into question. Particularly when it comes to health, crime and money. On the recent census, I crossed out the gender questions with a note 'I object to this question' and rewrote the sex ones.
Great, forensic article. Thank you. Stats NZ are involved in marketing their own ideas, not seeking data and information about the population. I've seen OIAs about their definition of sex being changeable which refer to it changing if a person changes their birth certificate. Idiotic and captured.