Domino effects: the implications of the framing the sex question in the 2021 census

The UK is losing robust data on sex and the framing of the sex question in the 2021 census is set to make the situation worse.  

The sex question in the 2021 census will ask whether a person is male or female (there will be a different, separate question where people can provide information on how they identify). However, the three UK census authorities (ONS, NRS and NISRA) also intend to accompany the sex question with guidance that advises respondents that they can answer based on their self-declared gender identity. This follows on from similar guidance introduced in the 2011 census, although it is important to note that the 2011 guidance was not subjected to any parliamentary or other form of external scrutiny, and the change escaped the attention of most if not all elected representatives.

In practice, the muddling of sex and gender identity means that for an unknown (and unknowable) number of people, the census will collect a different type of data to sex and is unlikely to provide reliable data at the subpopulation level.

The reframing of ‘sex’ in the UK census is already attracting the attention of other organisations, as shown in the following exchanges between directors of the European Social Survey and Understanding Society survey: 


A recent YouGov survey asked respondents to identify themselves ‘on the gender spectrum’ and, as we understand it, entirely omitted a question that asked for respondents’ sex:


In response to a complaint about the question format, YouGov subsequently explained that it had implemented the change to reflect the forthcoming UK census, and that other market research companies were planning to do the same. There’s an apparent misunderstanding here, that the census will not ask a male/female binary sex question at all, but the status of the census as setting the standard for others is clear.


A muddled landscape
As well as influencing future data collection exercises, the muddling of sex and gender identity in the 2021 UK census also looks set to a cement a shift that is already well underway in UK public bodies.  

Police forces now record incidents and crimes based on a person’s self-defined gender identity, including rape in some forces, rendering criminal proceeding statistics for offences that are rarely committed by women unreliable.

Higher education institutions (HEIs) are legally required  to monitor and publish data as part of the public sector equality duty (PSED) of the Equality Act 2010, under which sex is a protected characteristic. However the sex question used by the Higher Education Statistics Agency (HESA) incorporates an ‘other’ option which it states is ‘included for staff whose sex aligns with terms such as intersex, androgyne, intergender, ambigender, gender fluid, polygender and gender queer’.

Screenshot 2020-06-20 at 23.20.48

The ACAS/Government Equalities Office joint guidance on gender pay gap reporting states that employers should collect data based on their employees’ gender identity, and that ‘in cases where the employee does not self-identify as either gender, an employer may omit the individual from the calculations’.

Screenshot 2020-06-20 at 23.22.56

In 2018 the Scottish Government changed the binary sex question in the Scottish Household Survey, to a question that asked respondents to describe their gender identity, stating that this was ‘in line with similar changes being made to other major surveys in Scotland’.


Despite clearly reported sex-based differences in the effects of COVID-19, the widely downloaded KCL COVID-19 Tracker App developed to better understand the symptoms and spread of COVID-19, quickly changed its original male/female sex at birth question, to ask people about their sex ‘assigned at birth’ and whether they are male, female or ‘intersex’:


KCL COVID-19 Tracker App, 24 March 2020

KCL COVID-19 Tracker App, 20 April 2020

Similarly a survey by the London School of Hygiene and Tropical Medicine considering how people were coping with the pandemic and how information about the disease was being shared in communities listed five potential responses to the question ‘What is your gender?’.

Why has this happened?
This shift in data collection practices has come about principally as a result of lobbying by some LGBT groups which advocate collecting data on self-declared gender identity instead of data on sex. For example the Stonewall/Scottish Trans Alliance ‘Getting Equalities Monitoring Right’ guide explicitly advises organisations against collecting data on sex:

Public authorities are obliged to monitor the gender of their staff to report back on the pay gap. It’s absolutely vital to ensure that this isn’t restricted to male and female, but is inclusive of non-binary identities. There are lots of different non-binary identities – including a space for individuals to describe their own gender identity is important, and will allow you to capture much more information. Remember that most trans people will identify as men or women – this third option is usually used by trans people who identify as non-binary non-binary (or in other non-binary terms, such as gender-fluid or genderqueer).

UK law treats sex and gender as the same thing and in general conversation people often use these terms interchangeably. However, when it comes to monitoring questions, it is better to use the term gender rather than the term sex. This will help make it clear to people that you are asking them to tell you about their self-perception of their gender rather than about their biological sex. Asking questions specifically about biological sex at birth will be ineffective – most trans people find this question offensive and will provide their gender identity, making the data collected inaccurate. However, gender data can still be used to help plan services relating to biological sex, as most people will have a biological sex that corresponds with their gender. The technical term for someone’s self perception of their gender is their gender identity – it’s the term we use most often in this guide. But it’s absolutely fine to use the term gender instead on forms, particularly for staff or service users, as this is more widely understood. It’s important to ask monitoring questions that are clear and easy to understand, so that the data you collect is accurate‘.

The main failing, however, lies with those organisations, including the census authorities, who have neglected to consult more widely on their policies, particularly with data users and statisticians working with population level data. Instead, policy development has been influenced by a specific set of interests, without due regard for other affected groups, data users, the wider population or, where monitoring under Equality Act is relevant, the law.   

Such dynamics are for example, apparent in the 2016 ONS workshop to discuss gender identity in data collection, including the census. Attended by representatives from 12 LGBT organisations, workshop participants also suggested that data collection exercises ask about sex only on a self-identification basis:

participants recognised that it was important for the respondent to feel that they have the opportunity to self-identify in the way they want to and feel comfortable in so doing, for any potential question on gender identity, including questions asking about sex’.

Yet despite the significant implications of this recommendation for legal and policy interpretations of sex, it appears that no groups representing women’s interests attended the workshop.

We see this as an example of policy capture. Policy capture has serious social and political consequences. In relation to data collection, there has been a clear failure to recognise the full range of interests affected by the profound redefinition of sex as a matter of subjective feeling.  

In framing the sex question in the 2021 census in terms of self-declared gender identity, the UK census authorities are clearly signalling that they no longer view sex as a key determinant from birth onwards of a vast range of outcomes relevant to the formulation of public policy.

Widely considered to be the ‘gold standard’ of population surveys, the UK census’ flagship status also means that the census authorities are giving implicit licence to other data collection surveys to follow suit, and that the elision of sex and gender identity is even more likely to be replicated elsewhere.

As a result, the UK is now at a serious risk of losing the capacity to gather data that, for decades, has provided the building blocks for policy-makers and researchers to monitor and tackle discrimination based on sex.