Census 2021: Qualitative research on the guidance for the question “What is your sex?”

The Office for National Statistics (ONS) has conducted a comprehensive programme of research and testing to finalise Census 2021 questions and guidance. The report Census 2021: Final guidance for the question "What is your sex?” is available on our website.

This methodology article provides further detail about the qualitative test, referenced as 2020:6¹, conducted as part of the review of evidence that fed into the decision-making for both the target concept and the guidance wording. As such, it was conducted at the same time as finalising the concept of sex for the 2021 Census and a range of target concepts were included.

Notes for: Overview

1. References to tests take the form (year:test number). “Year” refers to the calendar year the test was undertaken in and the test number is the position of the test within all testing that took place in that year. For example, the fifth test conducted in 2017 would be referenced as (2017:5). A full list of tests is provided in the summary of testing for Census 2021.

The aim of the interviews was to understand how different groups responded to different wording and target concepts of the sex question guidance, and the impact it may have on willingness to complete the Census 2021 as a whole and the sex question in particular.

In relation to the target concept, the interviews explored the acceptability of the different options. This included exploration of the potential for a negative impact on overall census response or item level response to the sex question. Within this we explored the potential for this to disproportionately affect specific subgroups of the population, for example, women or trans people¹.

In relation to the guidance wording, the interviews explored the acceptability and clarity of terms used within the draft guidance.

Notes for: Research aims

1. The trans population is defined as the population whose gender identity differs from their sex registered at birth. This includes those with non-binary identities.

Earlier research on the sex and gender identity questions is detailed in Sex and gender identity question development for Census 2021 and included research into the guidance for use in the 2019 Rehearsal. This had previously identified that the sex question, and the associated guidance, could negatively affect response rates among the trans population.

Following the 2019 Rehearsal, we identified that there was also a risk to response rates in other populations. The sampling criteria therefore considered:

• trans status: trans and cisgendered people
• organisation membership: representatives from trans ally groups and women’s groups
• age group: people aged 16 to 29 years old, 30 to 59 years old, and 60 or more years old

In total, 52 participants were recruited from the general population representing a mix of these criteria. 16 participants were trans, 7 participants were members of trans ally groups and 16 participants were members of women’s groups. 16 cisgender participants were not members of either a women’s group or a trans ally group and included males and females. These groups are not mutually exclusive so one participant may be in more than one of these groups. A range of geographical locations were represented in the testing.

Participants were recruited through existing databases containing individuals who had “opted-in” to take part in future research and via a professional agency. Screening questions were used to determine eligibility and the research topic was not known to participants prior to the interview. Membership of any women’s or trans ally group was sufficient to meet the sampling criteria.

One-to-one interviews were conducted remotely via video conferencing software and in line with the Code of Practice for Statistics and best practice for Government Social Research.

Interviews explored participants’ interpretations of the sex and gender identity questions before cognitively testing four versions of the guidance, with varying target concepts, and guidance wording. This wording had been further developed from the version used in the 2019 Rehearsal, to take into account feedback received during and after that event. For example:

• information on “Why we ask this question” has been moved to the top of the page as research found this was important to encourage response (2020:5); this has been applied to guidance for all questions

• all references to specific population groups have been removed; this ensures that the guidance remains relevant to all potential readers and removes terms that are not universally acceptable to the communities being described

The versions of the guidance were shown to participants in different orders in a between-participants design to reduce order effects.

The interviews also included an in-depth discussion of methods participants have used in the past to provide feedback to businesses or public organisations, such as the Office for National Statistics (ONS), on their goods or services. This was then used to gauge strength of feeling and likelihood of action when discussing the actions they could take when presented with each of the guidance options being tested.

Full analysis of the interview content was conducted independently by two researchers before comparing and agreeing findings. This was then quality-assured by a third team member. The analysis considered the potential for the order the guidance was presented in to have affected findings.

Interpretation of the sex question

Before being presented with any version of the online help guidance for the sex question, participants were asked what they considered the question to be asking them. All participants were confident of their ability to answer the question without further guidance.

Across all age groups there were participants who interpreted sex as referring to “what you were born as”, that is sex as recorded on your birth certificate. They distinguished this as different from gender identity.

The proportion that made this distinction varied by age. Younger participants rarely referenced sex and gender as being the same, whereas this was more common in older age groups.

Across sample groups there were participants who objected to the binary nature of the sex question. For example, there were older cisgender participants who questioned why data on sex were needed. Strength of feeling on this varied between participants.

Participants, including some trans participants, would typically answer with the sex registered at birth. Trans participants who would answer in this way stated that this was because it was what was “legally” written down on their legal/official documents; and because they would want to provide accurate information as they believe the census generates important data. This included trans participants with a non-binary gender identity.

In contrast, other trans participants with a female or male gender identity would answer with that identity.

Acceptability of target concepts

Wording tested, aligning with different target concepts¹, was:

A version where the target concept was not explicitly suggested was also tested.

The strongest and most shared objections was against the target concept sex registered at birth. There were participants who stated that they would refuse to answer the question and participants who stated that they would refuse to complete the census. The objection was strongest among those with a Gender Recognition Certificate.

After seeing this version some individuals stated that they would potentially join a campaign and would encourage others not to complete the census or the sex question.

The target concept “Sex recorded on birth certificate” was accepted by a larger proportion of participants across all groups than the concept of “Sex registered at birth”. The main group negatively affected by this version of the guidance was trans participants with a male or female identity who had not obtained a Gender Recognition Certificate and there were instances where participants stated that they would refuse to answer either the census as a whole or the sex question having seen this guidance. However, there was no direct evidence of a potential for a campaign response.

While cisgender participants, including those who were members of women’s groups, expressed views, none of the guidance options tested would have an impact on their own response or elicit further action, such as campaigning against the question.

Other guidance options were more positively received, although there were still objections. For example, not stating a target concept does not acknowledge the barrier this question poses to trans individuals, and not all trans people have official documents aligned with their gender identity.

Notes for: Results related to the target concept

1. Because of the range of concepts of sex being considered the wording around self-identification and sex recorded on legal/official documents was tested within a single guidance version. The concept of sex as living or presenting was not tested.

Why we ask this question

All versions of the guidance started with information on “Why we ask this question” in a guidance accordion. When the heading is selected it expands to show the full text. This read:

This information helped some participants feel more positive about answering the question. However, the explanation was criticised for not providing an adequate explanation for why the data are collected in the way it is, for example:

• non-binary participants did not feel it sufficiently justifies why there are only two binary response options

• simply saying “equality monitoring” does not explain why the Office for National Statistics must ask for sex data

• there were concerns from trans participants that the data will be used to allocate services based on sex assigned at birth, which may result in trans people being excluded from services they need; trans participants who expressed this view said that the information within the link “upset” them more than the question itself

• the sentence about the sex question being asked since 1801 was considered superfluous and irrelevant among participants across the sample; members of trans ally groups pointed out that society has changed since 1801 and asking the question in the same way for so long is “terrible”.

The recommendation was to expand this guidance to provide more detail and to repeat information on the importance of the question within the main body of the text for further emphasis.

Select either “Female” or “Male”

The information on why we ask the question was followed, in all versions, by the statement:

Participants from all groups noted that this is just “reiterating the question”.

When read in isolation, in the version of the guidance that did not indicate a target concept, trans participants felt this statement was only aimed at cisgender people and did not help them to decide which option to choose.

Introducing the target concept

The target concept was introduced next in all versions of the guidance tested.

The version collecting self-identified sex used the statement:

While considered more inclusive and empathetic by some participants it was also seen as contradicting the preceding statement. The statement was interpreted as an “acknowledgement” that the question is not suitable for everybody, but it was noted that this assumes that everyone wants to answer as either female or male.

It was also criticised as having potential to lead to inaccurate or mischievous responses because of phrasing being “ambiguous”. Participants discussed the need to carefully consider the wording to ensure that factual responses were provided whatever target concept was being collected.

In other guidance versions the target concept was introduced using the sentence stem:

This wording was considered unacceptable and offensive in each version tested as it was interpreted as meaning that the participant was confused as to whether they were male or female, rather than unsure what target concept the question was asking for.

The recommendation was to replace this question stem with one that makes it clear that the clarity being given to respondents is about what target concept of sex is being asked for.

The wording for the end of this sentence varied depending on the target concept and is shown under the Acceptability of target concepts subheading in Section 5: Results related to the target concept.

Introducing the gender identity question

The target concept text was always followed by a paragraph introducing the gender identity question:

There were lots of questions about and opposition to the gender question only being asked of those aged 16 years or over with participants thinking it should be asked of everybody. Trans participants said it implies that younger people do not “know their own gender” and “disregards” the fact that people can define themselves at a young age¹.

However, the guidance performed as intended in that participants said they might answer the sex question themselves, or urge others to do so, despite their objections to the question because there would be a later question about gender identity where they can provide more detail.

Answering on behalf of someone else

The final section of the guidance is a headed subsection:

The first part of this guidance was appreciated among trans participants and their allies because it told people they should ask others before writing down an answer for them, and not to assume how individuals would want to answer.

However, there was some objection to the second part as participants felt that even if someone is away, other people should not answer questions about them in the census.

Notes for: Results related to the guidance wording

1. The main justification for asking the question only of those aged 16 years and over relates to data quality. This is discussed in more detail in the report Sex and gender identity question development for Census 2021.

These results formed part of the detailed evaluation of the target concept of sex for Census 2021 reported in the Methodology for decision making on the 2021 Census sex question concept and associated guidance.

They also formed part of the evaluation of the final guidance wording. The final wording is reported in Census 2021: Final guidance for the question "What is your sex?”.