Sexuality and Gender Identity in the Census

The Census in England and Wales is sent out every ten years to every household. It asks questions about issues such as demographic characteristics, the house you live in, your health and the work you do. This is important for service planning, and for identifying and addressing important social issues. As you might imagine, it’s a big piece of work and it takes a lot of planning, and so work is well underway towards the next Census in 2021.

Up until now, there have been no direct questions about sexual orientation or gender identity in the Census. In the 2011 data, the only LGBT people who are clearly identifiable are those in civil partnerships. That causes problems both for service planning and research. I’ll own up now to a bit of interest in this issue – I used to volunteer on a LGBT switchboard in a posh suburban/rural south-eastern county. It was run by a charity, on a shoestring budget, and we regularly used to go to the multiple Councils and health bodies and police teams that covered our area, and ask them for small grants to help with our running expenses. We’d also ask them to talk to us about how services could be made more inclusive. And on more than one occasion, we were greeted with genuine surprise. Were there really LGBT people needing support in this part of the world? Surely not. And then when the surprise had faded, we were asked for proof. How many LGBT people? Where? What ages? Did they have other intersectional needs? And the simple fact was, there wasn’t any reliable data we could use to make a case. Census data would have made a real difference in being able to highlight areas of need, and I believe it could have led to better services on the ground for LGBT people.

Following consultation about the 2021 census, the Office for National Statistics (ONS) has now cautiously agreed to look further into including sexual identity and gender identity questions . Here’s a summary of the key points from their report.

Sexual Identity

  • The ONS concluded that overall there was a “medium user need” for this data – however, in fact the need for this data scored 8/10 on all measures except continuity with previous surveys (since it has not been on previous surveys, it scored zero for this component)
  • The question they would be likely to use would be a variation on the one currently used by the ONS when interviewing people: “Which of the options on this card best describes how you think of yourself: Heterosexual/Straight, Gay/Lesbian, Bisexual, Other.” One of the potential difficulties is that, unlike with face-to-face interviews, the census is often completed by one household member on behalf of the others, or other household members can see the answer. The ONS has not tested this question in this situation, but there are some obvious problems with asking this question where answers are not confidential, including  the fact that it might put some people in an awkward or unsafe situation.
  • Asking this question would potentially mean a change in primary legislation. In the past, there were concerns by some politicians that asking a question on religion was too intrusive, and as a compromise it was made an optional question. Similar issues might apply here.
  • When this question has been asked in other research, there is a high proportion of don’t know/prefer not to say responses, which is often larger than the proportion of people saying they are definitely lesbian, gay or bisexual. This might affect how useful the data collected is.
  • The ONS has previously found that around one in ten people have said they do not want to provide information on their sexual orientation. There are therefore some concerns that including this question may lead to more non-responses to the Census overall.
  • The ONS suggests that the questions would only be asked for over 16s, as those below this age might not be able to answer this question. Certainly it seems inappropriate to state a sexual identity for very young children, but I’m personally not sure 16 is the right cut off. Their justification for this is that 16 is the age of consent. I think this  problematically ties sexual identity to having sex. It would also make the data less useful for youth services. If the question is likely to be optional anyway, why not allow those under 16s who wish to express a sexual identity?
  • In 2017, a large scale test of the census will be carried out. The ONS will include sexual identity questions in that, and evaluate the impact. Alongside that, the ONS will also conduct focus groups, talk to stakeholders and review existing work regarding collecting data on sexual orientation.

Gender Identity

  • This wasn’t actually included in the ONS consultation – however, they seem to have received a number of “write-in” responses saying this was important, so have decided to look into it further.
  • Overall, the ONS has concluded that there is a medium user need for this data, and that this is particularly the case because there are few other reliable sources of data on the trans and non-binary population. A number of respondents stressed that this is making it difficult to plan services and tackle inequality
  • The ONS states that “sex, as biologically determined, is one of the most frequently used and important characteristics the census collects as it is used in most multivariate analysis of data and feeds into the UK population projections”. I think this is something of a misunderstanding on the ONS’s part about the nature of the data they already collect – I filled in the census in 2011, and I did not respond to the male/female question with my “sex as biologically determined” – it genuinely didn’t occur to me that I should (and even if it had, I still wouldn’t have). Nor is it factually correct to assume that “sex as biologically determined” is a binary issue. I suspect the ONS is already collecting gender rather than sex data on this question. However, their overall point seems to be that it is  important that they be able to analyse differences between men and women, which I don’t think anyone was disagreeing with.
  • The ONS note that it is difficult to come up with a question on this which is acceptable to trans people and comprehensible by cis people. It would only take a relatively small proportion of cis people to misunderstand the question for the number of errors to outweigh the responses from trans people. (Personally, I also think there’s also likely to be an issue of “joke” responses – someone thinks it’s funny to put his housemate down as trans etc.)
  • Previous research has suggested that more than half of trans people would be unwilling to disclose trans status, however sensitively the question was asked. Additionally, census responses may be visible to other members of the household, which is likely to further increase concerns about confidentiality and lower response rates, as well as being very unsafe for some.
  • If gender identity was asked about, it would be likely to be similar to the EHRC recommended questions. This is in fact four questions: sex assigned at birth; how you think of your gender now; do you have the protected characteristic gender reassignment; and stage of gender reassignment (i.e. have you already transitioned, are you currently transitioning, do you intent to transition in the future). This adds a number of extra questions into the census,  which takes up a relatively large amount of space and may cause confusion among those for whom this isn’t relevant (e.g. some people may genuinely not understand why they are being asked about both their sex at birth and their gender now).  I am personally slightly uncomfortable with a set of questions which leads with “sex assigned at birth” – I understand why it is done, but I dislike my birth assigned sex being stated prior to my gender identity. Also, in view of the ONS’s apparent confusion above regarding the supposed importance of analysis by “sex as biologically determined”, I am somewhat concerned that they, or at least some analysts using the data, might treat my sex assigned at birth as the primary category for analysis rather than my gender identity.
  • Again, the ONS is concerned that including questions on gender identity may lead to more people not responding to the census.
  • Again, the ONS say it would need a change to primary legislation to include these questions.
  • Overall, the ONS agrees that there is a need for this data, but they don’t seem convinced that the Census is the way to do it. They therefore intend to review their previous work on collecting trans data, talk to stakeholders and talk to other countries’ statistics agencies to work out the best way forward.

Gender Identity doesn’t need science

So once again, a rash of articles and blog posts by “gender-critical feminists” asserting that we can’t offer protection for gender identity based on self-determination because there is no scientific consensus on what gender identity is. These articles then neatly segue into a discussion (usually biased) of the science, hoping no-one will notice the entirely false premise.

Scientific consensus is not required to protect a minority group. Virtually none of the “protected characteristics” in the UK Equality Act are the subject of scientific consensus. Sexuality? The scientists are still arguing whether it’s the gay gene, the size of your hippocampus, how many older brothers you have, or good old environmental factors. Race? The definition in the act talks about colour, nationality and ethnic origins, none of which is exactly amenable to ready scientific categorisation, and pretty much all scientists now consider the idea of absolute distinctions between races to be Victorian pseudo-science. Disability? The definition is an impairment with a substantial impact on someone’s day to day life – again a definition which is contingent on social factors, and hence established for equal monitoring purposes by self-definition. As for the scientific consensus on religion – well, the less said about that, the better.

We live in a pluralist society. We understand that it would be both ridiculous and hugely problematic for the government to impose central, monolithic categorisations for race, or disability, or religion, and apply them to everyone at birth, and place them on every government document. We would consider it to be a sign of an intolerant authoritarian government if you had to get formal permission to change those categorisations. And yet we can still manage to operate equality protections, despite the fact that definitions are fuzzy, despite the fact that some people may change how they describe themselves. We do this through the legislation being broadly phrased: the religion category does not come up with a list of religions and say these are the protected religions and this is how we define who belongs to which: it protects any religion or philosophical belief. Same for ethnicity and disability. That doesn’t invalidate or diminish concepts such as “Christian” or “Asian” or “blind”, nor mean that those groups aren’t protected under equality legislation, but it allows for flexibility and fluidity around these categorisations, and allows for the potential emergence of new groups who might also require protection.

As a society, we have a blind spot when it comes to the protected characteristic of “sex” (and by extension, to “sexuality” and “gender identity”, which are defined by reference to sex). Up until fairly recently, it was assumed that it was the government’s role to regulate this, and to do so in a clearly discriminatory way. It was assumed that it was necessary to be able to define who could marry who on the basis of genitals; that it was necessary to be able to treat women and men differently in tax and pay and pensions and insurance. Therefore the government needed to have a central definition, and it made sense to write M and F over every official document going. That’s out of date now. Thinking has changed, and from any feminist perspective that surely has to be a good thing. The government’s role now is to eliminate discrimination and promote equality. Therefore we should start to think about legislating for sex, gender and sexuality in the same way as we do for other equality characteristics, where multiple definitions and fuzzy categories are possible, but discrimination is still unacceptable.