For campaigners and commentators opposed to the inclusion of LGBTQ people in society, what is the ideal number of LGBTQ individuals? A statistic big enough to stoke fears about the ‘explosion’ of queer identities and impressionability of young people but also small enough to discount the views of LGBTQ community leaders as an irrelevance. Trying to establish that perfect percentage – neither ‘too big’ nor ‘too small’ – is a type of Goldilocks Data and is a trap that accompanies the publication of LGBTQ population data.

On 27 June, the National Records for Scotland will publish headline data from its 2022 national census, which included new questions on sexual orientation and trans status/history. For the first time, this data will provide a national-level estimate of the numerical size of Scotland’s LGBTQ population.

Among some LGBTQ people, this historical moment – seeing your life and your community represented in a census – marks another milestone in the Scotlands’s long journey towards LGBTQ inclusion. Yet, the collection of more data about LGBTQ lives also invites new risks. Also readying themselves to crunch the numbers are anti-LGBTQ campaign groups – such as Fair Play for Women – who have repeatedly tried to change data practices to make the lives of LGBTQ people worse rather than better.

My research explores intersections where LGBTQ lives meet data practices: who benefits and who loses when we collect, analyse and use more data about gender, sex and sexuality? And how do our decisions about data – what to count, how to count and who to count – invite differential harms for LGBTQ communities?

In the run-up to Scotland’s 2022 census, I organised and participated in events and social media campaigns that tried to encourage LGBTQ communities to participate in the count. I wanted to give the NRS and the Scottish Government the benefit of the doubt and see how the data collection exercise was conducted and if the data gathered about LGBTQ communities was used meaningfully.

While it is too early to say whether my optimism was misplaced, my enthusiasm for what a national count of LGBTQ people can achieve has soured. Following the publication of census data for England and Wales, I now find myself asking if there is a place for all LGBTQ individuals in state-organised counting machines (and if these are systems we want to join). There exists a long history of contested institutions co-opting some LGBTQ individuals – for example, the police, border control, global finance and the armed forces – to improve their image and pinkwash their reputation. Might ‘inclusion in data’ mark the most recent iteration in this longer trend?

The story so far

A census is different from other types of data collection exercises as it does not extrapolate from a sample – rather, it tries to count everyone. Therefore the design of questions and response options are particularly important as they need to make sense to a wide demographic of respondents.

Between 2018 and 2020, the Scottish Parliament’s Culture, Tourism, Europe and External Affairs Committee oversaw the design of the census and took particular interest in questions related sex, sexual orientation and trans status/history. The Committee included a small cast of parliamentarians including the Convener Joan McAlpine, who was quick to share her views on what does and does not count as a sexual orientation and how adding options such as ‘demiromantic’ and ‘asexuality’ would make the census ‘more like a Cosmopolitan quiz instead of a serious data gathering exercise’. Also on the Committee was Kenneth Gibson, who had no qualms admitting – during the design process for a question to establish the number of transgender and cisgender people in Scotland – that ‘until six months ago, I hadn’t heard the term and I didn’t realise that I was apparently “cisgender”’. As a case study, the composition of the CTEEA Committee highlights major concerns as who is in the room where decisions are made about data practices that have disproportionate impacts on the lives and experiences of LGBTQ people.

After much deliberation, Scotland’s 2022 census asked respondents aged 16 and over to describe their sexual orientation:

Which of the following best describes your sexual orientation?

• Straight/Heterosexual
• Gay or Lesbian
• Bisexual
• Other sexual orientation, please write in…

The census also asked whether respondents identify as trans or as someone with a trans history. The term ‘trans history’ can refer to people who no longer identify as trans but were assigned a different sex at birth (in other words, they have transitioned):

Do you consider yourself to be trans, or have a trans history?

• No
• Yes, please describe your trans status …

The addition of sexual orientation and trans status/history questions addressed the design fault in data collection exercises that – for too long – assumed everyone to be straight and cisgender.

However, when we scratch the surface, we start to see that this national data exercise was not as inclusive as it might first appear. While the census did include new questions on sexual orientation and trans status/history, it continued to require all respondents to identify their sex as either ‘male’ or ‘female’. For non-binary people, who may not identify in this way, selecting a binary sex option was compulsory with non-completion risking a fine of up to £1,000. Although some LGBTQ people were now counted, the design of the data collection tool ensured that particular characteristics – or ways of thinking about identity – were ‘designed out’ of the process.

The continuation of a binary sex question was not enough to satisfy anti-LGBTQ campaigners who sensed an opportunity call for further tweaks to the census that would erase the lives and experiences of trans people. In February 2022, the campaign group Fair Play for Women took the NRS to court to demand a change to the guidance that accompanied the sex question. Their case built on a tactic already tested in England and Wales, where Fair Play for Women took the Office for National Statistics (the government body responsible for delivery of the English and Welsh census) to the High Court to demand a minor (but meaningful) change to the guidance that accompanied the sex question. The original guidance for the 2021 English and Welsh sex question, explained: ‘If you are considering how to answer, use the sex recorded on one of your legal documents such as a birth certificate, gender recognition certificate, or passport’. Fair Play for Women demanded that the guidance be changed to remove the words ‘such as’ and ‘or passport’ so that respondents were advised to answer according to the sex recorded on their birth certificate or Gender Recognition Certificate.

The ONS – for some unknown reason – chose not to defend their position and the guidance was revised on 9 March 2021, more than two weeks after the census had gone live. Fair Play for Women did not enjoy the same luck in Scotland as NRS defended their position in court and won, which meant that guidance that accompanied the sex question remained unchanged and advised people to complete the question according to their sex recorded on a range of legal documents, including their passport.

Goldilocks Data

We have a blueprint for what to expect following this week’s publication of data on the size of Scotland’s LGBTQ population, as census data on the size of the LGBTQ population in England and Wales was published in January 2023. In England and Wales, 3.2 per cent of the population aged 16 and over identified as lesbian, gay, bisexual or a sexual orientation other than heterosexual and 0.5 per cent of the population stated their current gender was different from their sex registered at birth.

The census has also revealed interesting insights, including the jobs with the highest and lowest proportions of LGB+ workers. The jobs with the highest proportion of LGB+ people were air travel assistants (13.7 per cent); actors, entertainers and presenters (12.3 per cent); leisure and theme park attendants (11.0 per cent) and coffee shop workers (10.8 per cent). The jobs with the lowest proportion of LGB+ people were roofers, roof tilers and slaters (0.5 per cent); and bricklayers, farmers and scaffolders (all 0.6 per cent).

Less excitingly, a huge amount of media coverage has related to whether the trans population is ‘too big’ or ‘too small’. In October 2023, Kemi Badenoch informed the UK Parliament that she had asked the ONS to explore ‘whether the census got the number right’ after concerns were raised about respondent’s comprehension of the gender identity question used in the English and Welsh 2021 census, particularly among communities where English is not the first language.

---

---

I was never a champion of the gender identity question, which asked respondents ‘Is the gender you identify with the same as your sex registered at birth?’ and provided the response options, ‘Yes’ and ‘No, write in gender identity’ with space provided. An imperfect question risks returning imperfect data, fueling debates over the numerical composition of LGBTQ communities. Regardless of the question design and the specifics of the data published, we need to avoid falling into the trap of Goldilocks Data where the ideal number of LGBTQ people is always just out of reach. For opponents of LGBTQ inclusion, the LGBTQ population is always ‘too big’ (an omnipotent lobby group that is actively recruiting young people) and ‘too small’ (a numerical insignificance not deserving of specific funding or support). A perfect figure, statistic or percentage does not exist: by trying to meet their demands, we get caught in contradictory arguments and end up in a data trap.

The future of counting LGBTQ communities

The idea of conducting a count of the population every ten years seems increasingly out of step with recent data developments. The ONS has consulted on the future of the census and the possibility of using administrative data and annual surveys to replace this major data collection activity.

I am open to both sides of the argument. A census can reach corners of society that administrative data has so far failed to capture (for example, on sexuality and trans status/history) and foregrounds the agency of people completing the census as they are asked to actively share information about themselves and their households. There is something meaningful in asking people how they wish to be recorded rather than simply relying on inferences and extrapolations, or the scraping and extracting of data from existing sources.

By saying ‘no’ to the choices presented to us – man or woman, straight or gay, trans or cis – we are not just being awkward or difficult but challenging the narrow options that are understood as possible.

Aside from the huge cost of conducting a census and how data collected is very quickly out of date, we must also ask what LGBTQ communities have gained (and stand to gain) by sharing data about themselves in an exercise like a census. Creating situations where the meaningfulness of trans lives is debated in the courts or the production of contested datasets that provide talking points for media commentators does not improve the lives of the LGBTQ people who willingly participated in these data collection exercises.

Ultimately, we need to reflect on whether our existing data practices are worth fixing. We might come to realise that our current data systems have inequality and exploitation baked into them. And that, no matter how hard we try, they cannot be fixed to build a better world for everyone.

This leaves us with another possible option: abolition. If we stop trying to fix things, this requires a radical rethink of our relationship to data systems. By saying ‘no’ to the choices presented to us – man or woman, straight or gay, trans or cis – we are not just being awkward or difficult but challenging the narrow options that are understood as possible.

When we talk about inclusive approaches to data we need to consider the good and the bad, and ask whether we should dedicate our energy to fixing data practices in need of repair or taking things apart and building something new in its place.

---

---