Sex and the Scottish Census (Summary for Policy-Makers)

I am a nonbinary person in Scotland. The Scottish Parliament is currently considering how to best count trans and nonbinary people in the 2021 Census, and so I researched and wrote a long article on the history of sex in UK law, and on the history of trans people in the census, countering some of the debates that have taken place so far. This short article summarises my findings to be accessible to policy-makers, primarily MSPs who will be making decisions on legislation. I focus here on a straightforward analysis of the options, leaving aside philosophical debates. Every claim I make here is referenced and discussed in more detail in the full version.

How do trans people affect sex-based statistics?

The sociology of trans people’s lives is at a very early stage, given the local of social visibility and acceptance of trans people until recently prevented good data from being gathered. Initial results are that trans people do not have typical results either of their lived sex or their sex assigned at birth in any major area, from health to poverty to crime. A trans woman will not have the same typical health outcomes or needs as either a woman or a man; a trans man will not have the same typical economic outcomes as either a man or a woman.

Three initial studies illustrate the complications. Schilt and Wiswall 2008 found indicative results that trans women experienced pay discrimination as women, possibly even more so than cis women, while trans men experienced less pay discrimination than all women, though they were possibly not paid as much as cis men. FORGE 2011 found that trans men and trans women both experienced sexual violence at roughly equal rates, and both at higher rates than cis women. Dhejne et al 2011 found indicative results that a pre-1989 cohort of trans men committed crimes more frequently than women, whereas the same cohort of trans women did not commit crimes less frequently than cis men — but also that both effects were not present post-1989, when acceptance and social care increased.

If you want good data on trans people, and to prevent trans people’s complicated results from distorting general sex data by enabling consistent cis/trans disaggregation, then you need to ensure that trans men are all sorted into the same sex, and that trans women are all sorted into the same sex, and that a trans status question is asked. Only defining sex for the census by lived sex enables this through a mandatory census question. It is also the only option consistent with previous censuses. This is discussed below.

However you sort nonbinary people, we too are going to complicate sex data, because we too do not lead lives statistically like those who still live as their sex assigned at birth. We will complicate the sex data whether or not we are sorted by sex as assigned at birth or as recorded on our birth certificates, but counting us in a third write-in option provides the clearest and most reliable means of recording our population and what our lives are like. We exist, and we do not live lives statistically like men or women, so if you allow us to be counted we can begin to see exactly how.

How should sex be defined in the census?

I have seen four options for defining sex for census purposes in various discussions: biological sex, sex assigned at birth, sex recorded on the birth certificate, and lived sex. Here I consider each in turn.

Biological sex

There is no biological definition of sex in UK statutory law. UK case law varies the definition of sex depending on the legal circumstances and has not specified a general definition. Corbett v Corbett 1970 argued that all medical authorities agreed that sex included chromosomal, gonadal, genital and psychological factors, but that only the first three were relevant to marriage. W v W 2000 expanded Corbett to additionally consider hormonal factors and secondary sex characteristics, and ruled on the definition of sex in a marriage based on hormonal response, genital and psychological factors.

There is no strict and exclusively binary definition of biological sex generally accepted by scientists of sex and gender: that is, there is no way to sort all humans into two and only two categories on a rigorous and consistent scientific basis. The editors of the leading scientific journal Nature are on record as saying that such a proposal has no basis in science, and that “the idea that science can make definitive conclusions about a person’s sex or gender is fundamentally flawed”.

If the Scottish Government were able to provide a satisfactory definition of biological sex for the census, the question could only be asked on a voluntary rather than mandatory basis, as the European Court of Human Rights ruled in Goodwin & I v UK Government 2002 that it is a breach of the right to privacy to require transgender people to reveal their gender history. Even if the census could get around this problem, the majority of smaller-scale data gathering (as more commonly used by service providers) could not, which would make the census data incommensurate with the majority of data gathering.

Sex assigned at birth

As above, such a question could only be asked on a voluntary basis, due to the Article 8 right to privacy.

In addition, it would require all trans men — people who present as men in all social situations, who use male services, and so on — to answer as “female”, and all trans women to answer as “male”. This is likely to produce a high rate of census noncompliance from trans people, as well as sorting people living male lives into female statistics and vice versa.

Sex as recorded on the birth certificate

Under the Gender Recognition Act 2004, trans people can apply for a Gender Recognition Certificate to have their birth certificate modified to show their lived sex. This makes trans people their sex “for all purposes”. Both the GRA Section 9 and the Equality Act 2010 Section 7 make it explicit in the language that trans people with a GRC are legally their sex as acquired as well as their gender. Other legal mechanisms, such as driving licenses and passports, already operate on the basis of trans people’s acquired sex without the measure of a GRC.

Most trans people do not have a GRC. This is because the process is long, burdensome and costs money, and a GRC is not required in order to access services by your lived sex. Having a GRC or not does not indicate that a trans man lives any more or less consistently as a man. Additionally, surgery and other medical interventions are not required to have a GRC, only a diagnosis of gender dysphoria and two years of experience living in the acquired sex: a GRC does not determine the type of transition.

If sex is defined by birth certificate for the census, the sex question could be mandatory, but this process would separate trans people into two groups for statistical purposes, making analysis very difficult. For example, trans men without a GRC and trans women with a GRC would both be recorded as both “female” and “trans”, but with no ability to tell them apart in the statistics, despite their very different lives.

Lived sex

In 2011, the census for England and Wales and the census for Scotland both advised trans people to answer the sex question as “the sex that you identify yourself as, whichever you believe is correct”. In 2001, the census for England and Wales advised trans people to answer the sex question as “whichever you believe to be correct, irrespective of the details recorded on your birth certificate”. No previous census provided advice to trans people, or a specific definition of sex: it is safe to assume that trans people have always preferred to answer by lived sex. It is worth noting here that, prior to Corbett v Corbett, trans people did regularly have birth certificates and other legal instruments changed to acknowledge their acquired sex. 1970–2004 can thus be regarded as a 34 year interregnum, and census advice reflected lived sex prior to the GRA. Answering the question by lived sex is thus at least 20 years of census practice, if not, given trans people’s preference to answer this way and disinclination to identify themselves against lived sex, the entire history of the census, and so any change would create an inconsistency and thus require considerable justification.

As a question on trans status will be asked in the 2021 census, disaggregated data on trans and cis people could be provided for the first time, improving the statistical basis for both trans and cis people for all purposes, but only if the sex question is clear and does not divide trans people by another definition. That is, even if including trans people by their lived sex created an anomaly in the sex data, it could be made clear in the disaggregation by sorting the data by trans status. Any other definition will make this either difficult or impossible.

Lived sex is consistent with previous censuses, enables a mandatory question, enables clear trans data disaggregation, and is the most inclusive option for trans people.

Should the sex question have a third write-in option?

The benefits of a third write-in option fot the sex question are:

• It would provide a general population figure for nonbinary people.
• It could provide some further statistics about nonbinary people’s lives.
• It would be inclusive of nonbinary people and thus produce the highest rate of nonbinary census compliance and the least risk to that data.

The risks of the third write-in option are:

• It would be a change from previous censuses and thus create a small inconsistency in previous sex data in proportion to the nonbinary population

The decision can be made by weighing this risk against the benefits in line with the Government’s priorities. One mitigation of the risk is that estimates of the nonbinary population are very small — a fraction of a percentage. Another mitigation is that, as sex population proportions are relatively static, nonbinary people can be sorted into binary sex data randomly for some purposes (as they would also be if the question were not asked, given uncertainty about how nonbinary people would otherwise answer).

Intersex Human Rights Australia recommend nonbinary sex options in surveys, as some (but by no means all) intersex people are also nonbinary. The Intersex Campaign for Equality in the US also campaign for nonbinary legal recognition. I have not found an intersex-led organisation that recommends otherwise, but it is crucial that a diversity of intersex-led organisations are centred in the discussion about how the census can best serve them.

Including a third write-in option for sex is not an intervention into political or philosophical debates about what nonbinary people are. The only pertinent question facing the census is whether nonbinary people should be counted, and if so which method of counting us produces the most reliable statistics. Including a third write-in option provides the best balance of risks and benefits, not just for nonbinary and trans people, but for everyone.