This is a new service. Your feedback will help us to improve it.

Sex question recommendation report

Background information and evidence from our process developing the sex question and guidance for Scotland's Census 2022.

If you need this document in an accessible format, please contact us.

1. Introduction

Scotland’s next census will be held in 2021 subject to the approval of the Scottish Parliament. The census is the official count of every person and household in the country and the only questionnaire of its kind to ask everyone the same questions at the same time. Scotland has relied for more than 200 years on the information the census provides and it remains the best way to gather vital information which government, councils, the NHS and a range of users in the public, private and third sectors use to plan a range of public services, to improve the lives of those living and working in Scotland.

Census methodology and questions have evolved over time since the first census in 1801. In 2011, respondents were able to complete the questionnaire online for the first time and online guidance was available to assist respondents how to answer if required. The 2021 census is being designed to be ‘digital first’ and this will be the preferred route for the majority of respondents although paper questionnaires will still be available.

Sex is a key demographic variable and since 1801 the census has sought to collect data about the number of males and females resident in Scotland. This report provides the background information and evidence from the question development process that supports the proposed sex question and guidance for Scotland’s Census 2021.

2. Executive summary

2.1 Background

The census is the official count of every person and household in Scotland, and is the only questionnaire of its kind to ask everyone the same questions at the same time. Completion of the census is compulsory, although some questions are voluntary. In 2021 the proposal is that questions on transgender status and history; sexual orientation; and religion, religious denomination or body will be voluntary.

The proposed question on sex in Scotland’s Census 2021, is a mandatory question. It is a key demographic variable and since 1801 the census has sought to gather information on the number of males and females resident in Scotland.

Scotland’s Census is a self-completion survey which relies on respondents understanding the question and answering honestly and accurately within this context. Work to support people across Scotland to complete the census includes consideration of question design, guidance material and the functionality of other customer services support.

2.2 Question Development

All questions included in Scotland’s Census are subject to testing and evaluation which includes the following aspects:

  • Strength of user need
  • Suitability of alternative sources
  • Acceptability, clarity and data quality
  • Comparability
  • Operational consideration.

For the sex question, the long history of its inclusion in the census, topic consultation in 2015, evaluation of the success of collection in the 2011 census, and user satisfaction survey in 2015 confirmed that sex should be included as a mandatory question in the 2021 Census.

Scotland’s Census 2021 is designed to be ‘Digital First’ with an expectation that the majority of respondents will choose to complete online. The mandatory status of the sex question requires that it is answered by respondents online or via paper, however the functionality of the online form means that all respondents responding online must complete the question to progress further through the census.

2.3 Guidance

The sex question was not accompanied by guidance until 2011, when guidance was made available online. The guidance was developed by the Office for National Statistics and subsequently used in Scotland’s Census.

The 2011 online guidance for the sex question, supports a self-identification position. The self-completion nature of the survey, combined with there being no previous guidance can be seen as having enabled respondents to reach their own conclusions on how to complete the sex question i.e. self-identification.

The 2011 online guidance and proposed 2021 self-identification guidance therefore confirms and clarifies the continuation of the self-identification position.

2.4 Testing

Building on the evaluation of the 2011 Census, and the conclusion that sex should be a mandatory question in the 2021 Census, work progressed during 2017 and 2018 to further explore sex, gender identity, trans status and sexual orientation. Conclusions were presented by NRS in the Sex and Gender Identity Topic Report (September 2018).

In August 2018, question development and testing brought NRS to the conclusion that a non-binary sex question with self-identification guidance would be suitable. However, in August 2019 and following a recommendation from the Scottish Parliament’s Culture, Tourism, Europe and External Affairs (CTEEA) Committee the Cabinet Secretary for Culture, Tourism and External Affairs proposed that the Census 2021 would instead continue with a binary sex question with selfidentification guidance.

Following further discussions and correspondence between the CTEEA Committee and NRS, specifically on the nature and impact of guidance associated with the sex question, NRS commissioned ScotCen Social Research to carry out two within-subject surveys on

  • General population – adults aged 16 and over living in Scotland
  • Adults aged 16 and over and self-identifying as trans / non-binary and living in Scotland.

The key findings from ScotCen’s report are in Section 4 of this report, the Executive Summary is provided at Annex A, and a link is provided to the full ScotCen report in Annex B. Specific points from the general population study include:

  • Only 0.5% of online participants accessed the accompanying guidance before answering the question ‘what is your sex?’
  • There was no impact on non-response rates due to participants being pointed towards the legal sex or self-identification guidance.
  • For both the legal and self-identification sex guidance just over two thirds of participants reported that it was acceptable for use in the census (68% and 69% respectively). Just under 25% of participants expressed ‘neither acceptable or unacceptable’ and 5% of participants responded in both cases that the guidance was unacceptable.
  • Almost all (95%) reported that the impact on their census behaviour would be the same for both versions of the guidance.

Specific points from the trans or non-binary population study include:

  • Around 25% of trans and non-binary participants accessed guidance when answering the question ‘what is your sex?’
  • Those participants describing their trans status as non-binary were significantly more likely to access guidance.
  • In 60% of responses the participants answer to the sex question when self-identification guidance was used was different to the answer given when legal sex guidance was used.
  • Close to two thirds of participants viewed the legal guidance as less acceptable than the self-identification sex guidance for use in the census. Just over 21% felt that self-identification guidance was not acceptable compared to 77% who felt that legal sex guidance was not acceptable.
  • 69% of participants reported that they would answer the sex question if self-identification sex guidance was used, whilst only 23% reported that they would answer the sex question if legal sex guidance was used.

2.5 User needs and comparability over time and with the rest of the UK

Comparability of census data over time and across the rest of the UK are important factors that influence question design. As currently proposed, the censuses in Northern Ireland, Wales and England, and Scotland will use a binary sex question alongside self-identification guidance in 2021.

In addition the Census 2011 guidance in England and Wales, and Scotland used self-identification guidance, and it is NRS’s view that self-identification is in practice the most accurate description of how previous sex questions in the census – in the absence of guidance - were answered.

However, stakeholder views are not consistent or uniform. Some stakeholders and data users are clear that sex must be defined in terms of biology and that biological sex is a protected characteristic under the Equality Act. Others are clear that sex is not defined in the Census Act 1920, and that public bodies are not required to collect information on legal sex and can enable employees and service users to self-identify their sex.

Expressed views of some stakeholders is that comparability over time would be undermined by guidance on self-identification, whilst others take the view that self-identification is in fact the consistent position and guidance is required to provide clarity to data users.

2.6 Conclusion

The initial recommendation in 2018 by National Records of Scotland was for a nonbinary sex question with self-identification guidance. Following further consideration with a wide range of stakeholders and additional engagement with the Scottish Parliament’s CTEEA Committee, NRS brought forward the recommendation of a binary sex question with self-identification guidance.

Following consideration of representations from stakeholders across Scotland and the UK, including further independent research, NRS continue to recommend a binary sex question with self-identification guidance. The key strands underpinning this recommendation are:

  • The census is a self-completion form which has therefore relied on respondents considering the question, any guidance (including none) or other information before answering.
  • Most people – as confirmed by the ScotCen General Population study – will not consider guidance prior to answering the sex question. The basis on which they answer is therefore not explicitly defined and therefore NRS take the position that self-identification captures the reality of how people complete this census question.
  • Consistency over time and across the UK is enabled through the use of selfidentification guidance.
  • It is important to data users to have clarity through the availability of guidance on the basis for completion of the sex question.
  • The ScotCen General Population study and the recent NRS census rehearsal confirmed that most people don’t read the guidance on the sex question and that neither legal or self-identification guidance would change behaviour regarding participation with the census.
  • In the ScotCen trans or non-binary research it was reported that 3 times as many people would be willing to answer a sex question with self-identification guidance than with legal sex guidance.
  • A binary sex question with self-identification guidance therefore supports participation for all people with the census and clarifies to data providers and data users the basis of the question.

2.7 Proposed sex question and online guidance for Scotland’s Census 2021

What is your sex?

  • Male
  • Female

 

The proposed sex question is mandatory and should be answered by all respondents.

Proposed online guidance

How do I answer this question?

If you are transgender the answer you give can be different from what is on your birth certificate. You don’t need a Gender Recognition Certificate (GRC).

If you are non-binary or you are not sure how to answer, you could use the sex registered on your official documents, such as your passport or driving licence. The next question is a question about trans status and history. You can respond as nonbinary in this question.

I am answering for someone else. How do I answer?

If you’re answering for someone else, where possible you should ask them how they want to answer. If they’re away, select the answer you think they would choose. You don’t need to know or ask whether they have a Gender Recognition Certificate (GRC).

Why is this question asked?

The sex question provides vital information for organisations on national and local population statistics, and for long-term analysis, as it has been asked since 1801. This question is also used for equality monitoring.

3. Background

3.1 Aims of the census

The census is the official count of every person and household in the country and the only questionnaire of its kind to ask everyone the same questions at the same time.

Scotland has relied for more than 200 years on the information the census provides and it remains the best way to gather vital information which government, councils, the NHS and a range of users in the public, private and third sectors use to plan a range of public services, to improve the lives of those living and working in Scotland.

As a producer of official statistics, National Records of Scotland (NRS) is responsible for the planning, conduct and dissemination of outputs from Scotland’s Census 2021. The strategic aims for Scotland’s Census 2021 are to provide complete, accurate and accessible population statistics which meet user requirements; to build public confidence and encourage participation in the census; to protect confidential personal census information; and to provide value for money whilst building on the successes and lessons learned from previous censuses.

The entire census operation will be subject to assessment and monitoring by the Office for Statistics Regulation (OSR) to retain our National Statistics Accreditation, awarded to us for the last census in 2011. This accreditation process confirmed that the 2011 census adhered to the Code of Practice for Statistics by providing public value, producing high quality data and being produced by people and an organisation that were worthy of trust.

The whole of the Scotland’s Census 2021 programme will be assessed by OSR. This includes the initial planning stages that started in 2014, through to the production of the last standard outputs sometime in 2024. The assessment is split into three phases during which Scotland’s Census 2021 must demonstrate how it is being trustworthy, of high quality and of value to users.

3.2 Question development

Designing a questionnaire for a national census is a complicated process. The key aspects of the census are that it counts everyone in Scotland, it has to be credible, people have to have confidence in it and needs to be consistent to other comparators. It must ensure all citizens are represented and reflect with accuracy how they live in Scottish society. The questions should enable all people to respond and engage with the census regardless of their age or background in order to maximise participation and response rates. The census is a self-completion survey and it is important that respondents are able to answer all questions without the aid of an interviewer to provide support. Questions are tested to ensure that they can be easily answered by respondents. In addition the inclusion of questions should have no significant determent to the census as a whole particularly in the level of response so that the data gathered is of good quality to meet users’ requirements.

Topics can be asked in varying degrees of detail depending on the required data, and questions can be asked in different ways which can often lead to very different responses. Furthermore, questions asked earlier in a questionnaire may influence how people respond to later questions and people may be influenced by answering similar questions in previous censuses or other surveys. For all of these reasons, it is essential that the development of the questions, questionnaires and the associated guidance for the census follows a robust analytical process that is widely tested to ensure that it delivers high quality data outputs that meet a wide range of user needs. The evidence for the question development process is gathered from a wide range of sources. There is the knowledge gained from evaluating questions in previous censuses, surveys and the census rehearsal, learning from census offices in the rest of the UK and internationally, data user requirements, stakeholder feedback and question testing results and all of this evidence contributes to the decision on whether a question should be included, the question wording and format.

Questions and their response options are evaluated against five main themes:

  • Strength of user need: Data collected by the census must meet a user need for equality monitoring, policy development, resource allocation and/or service planning and delivery.
  • Suitability of alternative sources: Data collected by the census must meet a user need that cannot be met elsewhere.
  • Acceptability, clarity and data quality: Questions asked in the census must be acceptable to the majority of the public, clear and be designed with minimal respondent burden in order to obtain good data quality that meets user needs.
  • Comparability: Data collected by the census should be comparable over time where possible, and harmonised across the UK where reasonable.
  • Operational considerations: Census questions must be considered as part of the census as a whole, where effective digital and paper design, space and financial constraints must be considered.

3.3 Sex question

Sex is a key demographic variable and since 1801 the census has sought to collect data about the number of males and females resident in Scotland. As the data collection methodology for the census has evolved, so have the questions. For the first time in 2011, respondents were able to complete the questionnaire online and there was online guidance available for all questions to assist respondents how to answer if required. The online guidance for the sex question was developed by the Office for National Statistics (ONS) for the 2011 Census in England and Wales, and introduced to the census in Scotland for the first time in 2011.

2011 Sex question:

  1. What is your sex?

Tick one option

  • Male
  • Female

2011 Sex question guidance - online only

How to answer

Tick one box only

Why is this question asked?

This question provides important basic information about the population and is used extensively along with other data (for example, age and marital status) to produce statistics at national, regional and local level and to allocate public money to local authorities, health authorities and other community projects. By combining sex with age, the census provides a basis for measuring changes in society and calculating rates of morbidity, mortality, fertility, marriage and divorce.

Help with answering

No guidance provided

More questions?

I am transgender or transsexual. Which option should I select? If you are transgender or transsexual, please select the option for the sex that you identify yourself as. You can select either ‘male’ or ‘female’, whichever you believe is correct, irrespective of the details recorded on your birth certificate. You do not need to have a Gender Recognition Certificate.

If you are answering for someone who is transgender or transsexual then where possible you should ask them how they want to be identified. If they are away, you should select the sex you think they would wish to be identified as. You can select either ‘male’ or ‘female’, irrespective of the details recorded on their birth certificate. You do not need to know if they have a Gender Recognition Certificate.

The 2021 census is being designed to be ‘digital first’ and this will be the preferred route for the majority of respondents (80%). As in 2011, online guidance will be available to assist respondents how to answer if required. The sex question is mandatory and online respondents must answer the sex question in order to progress to later questions. It is therefore vital that the sex question is acceptable and understood by all respondents in order to maximise response rates which is a key factor for delivering a fundamental objective of the census, namely collecting good quality data.

Respondents will also be able to complete the questionnaire on paper and there will be telephone capture and call centre support available for those who require support.

3.4 Sex question development for 2021

The starting point for question development work for the sex question was the question that had been asked in Scotland’s Census 2011.

The NRS initial view for the Topic Consultation in October 2015 was that the 2021 census should continue to include the sex question and collect data for this variable. This view was informed by evaluation of the success of the 2011 topics and questions (including evidence from research and previous censuses), user requirements from the Beyond 2011 consultation carried out in 2013 and feedback from the Scotland’s Census 2011 User Satisfaction Survey in 2015.

NRS invited views on Scotland’s Census 2021 – Topic Consultation between 8 October 2015 and 15 January 2016. The public consultation was a key step towards understanding what information users would need from the census in 2021, and helped to build strong cases to justify the inclusion of topics. Following the consultation, NRS proposed to continue to collect information on sex in 2021.

Following the Topic Consultation NRS worked closely with stakeholders through follow-up events and meetings to gather more detailed information about data requirements to ensure user needs were understood. The wording of questions which were subject to change or new (including the sex question) were then subject to a programme of testing including public acceptability testing, cognitive testing, and quantitative testing as appropriate to the individual questions. The results of this testing and the question development for the sex question were published alongside the Plans for Scotland’s Census in the Sex and Gender Identity Topic Report in September 2018.

In Autumn 2018, as a result of the question development described in the Sex and Gender Identity Topic Report, NRS were proposing that the 2021 Census would ask a non-binary question sex question with similar guidance to 2011 followed by a newly developed trans status question in order to collect data about the sex of Scotland’s population in addition to the location and size of the trans population. A non-binary sex question had been suggested in the topic consultation and further consultation with some stakeholders confirmed that the 2011 question may have excluded those people who did not identify in a binary way as they were not represented by the binary ‘male/female’ classification.

The proposal for a non-binary sex and trans status combination of questions was considered by NRS and concluded that the approach:

  • could be answered by both the trans and non-binary populations and was acceptable in testing;
  • could be answered by the general population and was acceptable in testing produced outputs consistent with 2011;
  • produced outputs comparable with the rest of the UK;
  • produced outputs that met user need.

In October 2018, the Census (Amendment) (Scotland) Bill was introduced to amend the Census Act 1920 to enable particulars about transgender status or history and sexual orientation to be gathered voluntarily. During the legislative process, the proposed sex question came under scrutiny and while there was support for the proposed sex question, some concerns were also raised. In February 2019, the Culture, Tourism, Europe and External Affairs (CTEEA) Committee published their Stage 1 Report on the Bill and recommended that:

  • the mandatory sex question should remain binary in order to maximise response rates and longitudinal consistency with previous censuses;
  • there is a pro-active information campaign to support the approach adopted which must involve all interested parties who have expressed views on this issue;
  • future consultation on this topic, by NRS and the Scottish Government, reaches out to the widest possible constituency and carefully considers all the evidence gathered by this Committee.

Since February 2019, in line with the CTEEA Committee’s recommendations, NRS has considered the evidence from Stage 1 of the Census (Amendment) (Scotland) Bill and carried out further stakeholder engagement.

In August 2019, Fiona Hyslop, Cabinet Secretary for Culture, Tourism and External Affairs, wrote to Joan McAlpine MSP, the Convenor of the CTEAA Committee, to clarify the National Records of Scotland’s proposal for the sex question for the Census in 2021:

“After careful consideration, reflecting on the Committee’s recommendation, stakeholder evidence during Stage 1 of the Census (Amendment) (Scotland) Bill, and the approach being taken in censuses for the rest of the UK, National Records of Scotland now propose to continue with a binary sex question in Scotland’s 2021 Census. The current proposal would also be to continue with the self-identification basis of response. (2019: 9)”

In the second half of 2019, NRS commissioned further testing of the guidance associated with the binary sex question. The testing was carried out by ScotCen Social Research with the aim of understanding the impact of different guidance on question response, data quality, question acceptability and census completion.

The following chapters of this report describes the evidence from the sex question development process that supports NRS’s decision to propose a binary sex question with a self-identified basis of response for Scotland’s Census 2021.

4. Question testing

It is essential that the development of the questions, questionnaires and the associated guidance for the census follows a robust analytical process that is widely tested to ensure that it delivers high quality data outputs that meet a wide range of user needs.

Both binary and non-binary sex questions have been tested as part of the question set for sex, gender identity, trans status and sexual orientation. Throughout 2017 and 2018 a range of testing methodologies were used across thousands of households in Scotland to fully understand different aspects of asking questions of this nature. The results of this testing were published in the Sex and Gender Identity Topic Report which was published in September 2018.

In the latter half of 2019, NRS commissioned further testing of the guidance associated with the binary sex question. The testing was carried out by ScotCen Social Research with the aim of understanding the impact of two different versions of the sex question guidance which were developed in conjunction with a broad range of stakeholders.

The Executive Summary of the testing report is provided at Annex A, and a link is provided to the full ScotCen report in Annex B.

The aims of the testing were to:

  • understand the impact of different guidance for the sex question on question response, data quality, question acceptability and census completion;
  • understand if the impact is different for the two populations being tested i.e. the general population and the trans and non-binary population;
  • examine the interaction between the sex question and trans status and sexual orientation questions when different guidance is provided.

In order to understand the impact on different populations there were two separate surveys:

  • general population - adults aged 16 and over living in Scotland;
  • trans and non-binary population - adults aged 16 and over self-identifying as trans or non-binary and living in Scotland.

The surveys were conducted in parallel and the design and content of both were the same but other aspects of the methodology differed e.g. sampling methodology.

Key findings of testing

How do people answer the sex question when first asked?

General population: 96% of participants answered the sex question when first asked and 4% chose not to answer it. Of those that responded, 53% reported being female and 47% male.

Trans and non-binary population: When first asked the question ‘what is your sex?’, 47% of trans or nonbinary participants responded ‘female’, 45% responded ‘male’ and 8% chose not to respond. All of those who did not answer the sex question when first asked, described their trans status as ‘nonbinary’ and the binary nature of the response options was the key reason for non-response.

How people use guidance

General population: Around one in ten (11%) people reported reading the guidance before answering the sex question. Background data collected for online participants indicated, however, that only 15% of online participants that reported reading the guidance actually clicked on the ‘help’ button to access it (0.5% of all online participants). There was similar evidence of overreporting among those who took part on paper. Non-response to the sex question when first asked did not differ significantly by guidance type.

Trans and non-binary population: When asked, one in three (33%) trans or non-binary participants reported accessing guidance when answering the question. Background data collected on online participants indicated that a quarter of those who claimed to access guidance did not actually do so. Thus, overall, 25% (n=19) of trans or non-binary participants genuinely accessed guidance before answering the question ‘what is your sex?’ Those describing their trans status as ‘non-binary’ were significantly more likely than those who described it in another way to access guidance before answering the sex question (37% and 15%, respectively). Of the 25% (n=19) of trans or nonbinary people who accessed guidance when first asked to respond to the ‘what is your sex?’ question, non-response to the question did not differ significantly by guidance type.

How people answer the sex question when asked to consider different versions of the guidance

General population: Non-response to the sex question did not vary significantly according to guidance type (1% for self-identification and 2% for legal sex guidance). Those in the general population who reported they were trans or had a trans history were more likely than others to avoid answering the sex question, irrespective of which version of the guidance was used (18% said they would not answer if self-identification sex guidance was used, compared with 1% of people who did not identify as trans or nonbinary.) The equivalent figures for legal sex guidance were 59% and 1%, respectively.

Trans and non-binary population: When presented with the self-identification sex guidance, 23% (n=16) of trans or non-binary participants reported they would not answer the sex question. Around half (49%, n=35) said they would not answer the sex question if the legal sex guidance was used. For 60% (n=41) the response given to the sex question when selfi-dentification sex guidance was considered did not match the response given when legal sex guidance was considered. Twenty-six of the 32 trans or nonbinary people reporting that they would not answer the sex question if legal sex guidance were used, provided a valid answer to the question when the self-identification sex guidance was considered.

Acceptability of different versions of the guidance

General population: Two-thirds (69%) of participants reported that the self-identification sex guidance was acceptable for use in the census. This was not significantly different from the 68% that found the legal sex guidance acceptable. Nine in every ten (91%) people gave the same acceptability response to both versions of the guidance.

Trans and non-binary population: Sixteen of the 75 trans or non-binary participants (21%) described the self-identification sex guidance as either not acceptable or not at all acceptable. When asked how acceptable the legal guidance was for inclusion in the census, 58 participants (77%) viewed it as unacceptable. While some viewed the self-identification sex guidance as unacceptable, they made it clear that they deemed it more acceptable than the legal guidance.

Impact on census population

General population: For both versions of the guidance, nine in ten people (91% for self-identification and 90% of legal) said they would ‘answer the question ‘what is your sex?’. For the self-identification sex guidance, 2% said they would ‘skip the sex question’ in the census if the guidance was used and 1% said they would ‘not complete the census at all’. The likely impact of the legal sex guidance was the same. The vast majority (95%) reported that the impact on their census behaviour would be the same for both versions of the guidance.

Trans and non-binary population: Of the 75 participants, 52 (69%) reported that they would answer the sex question in the census if selfidentification sex guidance were used. Fourteen would skip the sex question and just 4 would not take part in the census at all if the guidance were used. Forty percent (n=30) indicated that legal sex guidance would have a more negative impact on census behaviour than self-identification sex guidance and 3% (n=2) said the opposite.

Discussion

From the testing it can be seen that very few people in the general population access or read the guidance before answering the sex question. This is in agreement with evidence from the 2019 census rehearsal where there were 52 recorded views of the on-page guidance for the sex question. By way of comparison, views of on-page guidance in the rehearsal ranged from 0 to 790 views for the Scots language question, with an average number of views across the question set of 96. For respondents completing the questionnaire on paper, there were only 2 recorded views of the additional guidance that was available on the census rehearsal website.

When the general population do read the guidance, there is no difference in the item non-response, acceptability and impact on census completion between the self-identified and legal sex question guidance. For the trans or non-binary population however, 25% do access the sex question guidance with the self-identified guidance was much more acceptable compared to the legal sex guidance. Half of trans or non-binary participants reported that they would not answer the ‘what is your sex?’ question if it was accompanied with guidance which asked them to answer according to their legal sex. This was significantly higher than the non-response rate if self-identification guidance were to be used. It was also found that the legal sex guidance would have a more negative impact on census completion: only 24% said they would answer the sex question, 30% said that they skip the question and 30% said they would not complete the census.

The testing results support the NRS proposal of asking a binary sex question with associated self-identified guidance as the guidance has no impact on how the majority of people in Scotland would answer the question. For the trans and nonbinary population, the self-identified guidance has a lower non-response, is more acceptable and has less impact on census completion than the legal sex version of the guidance.

Consultation and engagement

Since October 2015, NRS has engaged with a wide variety of stakeholders around the sex question through the public topic consultation, public events, meetings and correspondence. Consultation and engagement are essential to fully understand the range of data needs and how census can deliver good quality outputs that meet the range of data users need.

The census is unique as a data collection, and as such, maintaining or improving the quality of the data, and providing comparability across time, or across the wider UK is important and users have identified this as a key need.

5.1 Stakeholder engagement

The NRS initial view for the Topic Consultation in October 2015 was that the 2021 census should continue to include the sex question and collect data for this variable. This view was informed by evaluation of the success of the 2011 topics and questions (including evidence from research and previous censuses), user requirements from the Beyond 2011 consultation carried out in 2013 and feedback from the Scotland’s Census 2011 User Satisfaction Survey in 2015.

NRS invited views on Scotland’s Census 2021 – Topic Consultation between 8 October 2015 and 15 January 2016. The public consultation was a key step towards understanding what information users will need from the census in 2021, and helped to build strong cases to justify the inclusion of topics.

There were 47 responses received through the topic consultation on the subject of basic demographics. A summary of these responses can be found in the Topic Consultation Report. In this report, NRS acknowledged that there is a wellestablished user need for the key demographic variable sex. Along with age, the sex variable is a vital input to population estimates and household projections which are used by central and local government to inform resource allocation, target investment, and carry out service planning and delivery. Sex is a protected characteristic as set out in the Equality Act 2010 and the data are widely used to inform equality impact assessments and allow organisations to fulfil their Public Sector Equality Duty (PSED) under the Act. They are also essential for analysis and research conducted by a wide range of users, including public bodies and third and private sector organisations. Following the consultation NRS proposed to continue to collect information on sex in 2021.

Following the Topic Consultation NRS worked closely with stakeholders through follow-up events and meetings to gather more detailed information about data requirements to ensure user needs were understood. In November and December 2018, NRS held three public events in Edinburgh, Glasgow and Aberdeen in order to share and discuss the proposed questions for Scotland’s Census 2021 including the proposed non-binary sex question.

During Stage 1 of the Census (Amendment) (Scotland) Bill, it became apparent that there were stakeholders who had not previously engaged directly with NRS and since January 2019 NRS has endeavoured to meet with these stakeholders in Scotland to hear their views. Since January 2019, NRS has engaged with a wide variety of stakeholders around the sex question through public events, meetings and correspondence. These include equality groups, women’s groups, local authorities, Scottish Government, and ONS and NISRA.

Annex C provides a list of the meetings that NRS has held during 2019.

The next sections provide a summary of the evidence that NRS has collected during the consultation and engagement process.

5.2 Definition of sex

Some of the evidence from Stage 1 of the Census (Amendment) (Scotland) Bill and stakeholder engagement since January 2019 has raised concern over the definition of sex. Some stakeholders have the view that sex must be defined in terms of biology and that this in line with the statutory definition of sex:

"Sex should be defined straightforwardly, as it is in the Equality Act 2010 - biological and legal sex" (Anonymous A1. 2018: 1)

They express concerns that the proposed guidance for the sex question in 2021 allowing trans respondents to tick a response that is different from that on their birth certificate does not align with their interpretation of the definition of sex.

Other stakeholders however have a different view and do not believe that sex must be defined in terms of biology in a legal context:

"Since there is no single definition of ‘sex’ for all purposes, in practice, legally ‘sex’ has meant different things, depending on the context in which the term is used.

"This means that the Scottish Government is free to interpret and apply the category of ‘sex’ (or indeed ‘gender’) in any given context, in a manner that is appropriate to that context, without having to define ‘sex’, and without having to abide by any previous definition or application of the term ‘sex’." (Cowan, 2018: 4)

The term ‘sex’ is not defined in the Census Act 1920 nor is it defined in the Census Order. There is in fact no single statutory definition of sex. Although sex is a protected characteristic under the Equality Act 2010, the Act does not define sex. It simply states that a man is a male of any age and a woman is a female of any age and how male and female are to be interpreted is not stated.

5.3 Uses of outputs

The strength of user need and the use of outputs is one of the key considerations in question inclusion and design in the census. It was clear from the topic consultation and stakeholder engagement that there is a well-established user need for the key demographic variable sex and the uses of outputs is now examined in more detail.

5.3.1 The topic consultation showed that sex variable data from the census is a vital input to population estimates and household projections. These are used by central and local government to inform resource allocation, target investment, and carry out service planning and delivery. For example the City of Edinburgh Council said:

"Knowledge of the age and gender makeup of the local population is essential in service delivery planning. Such information is a basic requirement for all service areas in the council." (2015: 3)

NRS consulted with the Population and Migration Statistics (PAMS) group during 2019 who did not raise any concerns with the proposed binary sex question for 2021 and the proposed guidance.

There has been a lot of discussion around resource allocation and the planning of health services during evidence gathering for the sex question.

There were concerns that a sex question that allowed trans people to self-identify would cause issues with health screening that applied only to males and females For example, in their evidence at Stage 1 of the Census (Amendment) (Scotland) Bill, For Women Scotland said:

"Knowing how many male and female citizens there are enables public authorities to properly plan and deliver public services. For instance, there are a number of sex-specific health screening programmes offered by the NHS. Cervical and breast screening are offered to women to check for early signs of cancer. Abdominal aortic aneurysm screening is offered to men during the year they turn 65." (2018: 8)

However, in his evidence to the CTEEA Committee on the 13th December 2018 Gerry McCartney, NHS Health Scotland said:

"We would not use the census to identify people for screening or other such purposes; we would use the existing health record data sets such as the community health index data sets and the clinical records associated with those to identify individuals who needed particular services. However, the census is a key data source for resource allocation and planning at a population level. It is important to recognise the difference between those two data sets." (2018: 12)

In relation to healthcare planning, Louise Maclennan (NHS) who submitted evidence at Stage 1 of the Census (Amendment) (Scotland) Bill also said:

"Allowing trans people to record their sex in accordance with their lived identity on medical records and diversity monitoring forms in health rather than their biological or GRC sex enables better healthcare planning and improved patient experience." (2018: 11)

There was also concern that there should be consistency between census sex data and NHS health data. Professor Jackie Cassell from the Brighton and Sussex Medical School whilst providing evidence to the CTEEA Committee on the 13th December 2018 said:

"With the growing use of administrative data sets—a couple of censuses away, the census will look very different and we will draw on national health service data and all sorts of data—it is important for the credibility of the census as a data resource that gives information at a low level on small populations that there is consistency between the census data on sex and other data sets." (2018:2)

Evidence provided from Louise Maclennan (NHS) confirms that the approach taken in the census for the sex question in 2011 and also proposed for 2021 aligns with how the NHS records sex data in the NHS CHI database:

"The current census guidance states that trans people should answer the sex question in accordance with the sex that they identify themselves as (rather than the biological or legal sex) which is fully consistent with the way NHS Scotland allows trans people to answer questions asking them their sex and is in keeping with how the NHS CHI number database records people’s sex. If the census ceases to allow trans people to self-identify their sex in the census this may impact on trans people and /or on the ability to compare census data with NHS patient data." (2018:11)

NRS have also consulted NHS ISD whose main use of census data is population estimates by age and sex which are used for planning, modelling profiling and resource allocation purposes. They confirmed that they are content with the proposed sex question and guidance for the 2021 census as it ensures consistency with previous census data.

5.3.2 Equality monitoring

Sex is a protected characteristic in the Equality Act 2010. The Act defines a man as a male of any age and a woman as a female of any age. The Equality Act 2010 places a duty on all public authorities to have due regard to the need to eliminate unlawful discrimination, advance equality of opportunity and foster good relations (the Public Sector Equality Duty or PSED). A wide range of public authorities which are themselves subject to the PSED have indicated a need for data from the census which will support them in fulfilling their own duties under the PSED.

12 responses in the topic consultation, the majority of which were local authorities, said that they used sex data from the census for equality monitoring:

"This topic is fundamental to characterising area populations, of great importance for equality monitoring, and is used for cross tabulation with most, if not all, other topics." (South Lanarkshire Council, 2015: 15)

Evidence from Stage 1 of the Census (Amendment) (Scotland) Bill indicated that some of those providing evidence to the Committee thought that in order to fulfil duties under the Equality Act 2010, the sex question asked in the census should ask for a person’s biological sex and further consultation with some women’s groups in 2019 confirmed that this was their view:

"Biological sex is a protected characteristic under the Equality Act 2010." (Murray, 2018: 13)

Other Stage 1 evidence however disagreed with these claims and supported the NRS proposal that a sex question asked on a self-identified basis would meet the user need of those using the data for equality monitoring purposes:

"Although sex is a protected characteristic under the Equality Act 2010, the Act does not define sex. Section 212 (1) simply says that a man is a male of any age and a woman is a female of any age. How male and female are to be interpreted is not stated.

"The Act treats sex as inclusive of all those who are recorded as that sex on their birth certificate, ie both those who are ‘biologically’ that sex, and those who have been designated that sex under the gender recognition process. It also protects those who are perceived to be a certain ‘sex’ even if they are not." (Cowan, 2018: 4)

"Employers and service providers who include sex in equality monitoring statistics generally support trans people to respond with their lived sex, not their legal (current birth certificate) sex. A lived sex question in the census will therefore provide a more consistent baseline statistic for employers’ and service providers’ equality monitoring of sex." (Equality Network, 2018: 6)

A recent paper from the Equality and Human Rights Commission (EHRC) has clarified the position by stating:

"Many public bodies use the terms ‘sex’ and ‘gender’ interchangeably. The language used by a body does not dictate whether the PSED requirements have been met. Public bodies are not required to restrict the information they collect to legal sex and can enable employees and service users to self-identify their sex.

"The vital question that public bodies need to address is what they will be using the information they are collecting for and whether allowing respondents to self-identify their sex will affect that. The number of trans people is not big enough to skew data that is collected on sex and it is therefore deemed NRS Sex Question Recommendation Report Scotland’s Census Page 27 of 46 generally unnecessary, from a data collection view point, to merely focus on legal sex.

"Public bodies often compare their data to information collected through the census, which (currently) allows respondents to self-identify their sex. The comparison would therefore be more accurate if public bodies use the same approach.(2019: 5)"

The EHRC also provide the following information regarding human rights considerations:

"There are important human rights considerations that need to be considered when asking employees or service users to state their sex, especially in regard to people’s right to privacy under Article 8 of the ECHR and the Human Rights Act 1998. Forcing trans employees or service users to disclose their sex as assigned at birth would be a potential violation of their human rights, particularly their right to privacy and dignity under Article 8. In addition, forcing a trans person without a Gender Recognition Certificate (GRC) to disclose their legal sex would result in that person being ‘outed’ as a trans person. It is also a criminal offence under the Gender Recognition Act 2004 for a person who has acquired the information in an official capacity, to disclose information relating to another person's application for a GRC or their gender prior to grant of the GRC. In some instances, forcing people to ‘out’ themselves will also breach the Equality Act 2010." (2019:5)

NRS recognises the critical importance of the Equality Act 2010 and that it is the cornerstone of UK-wide equalities legislation. The Equality Act does not however, as a matter of law, restrict the questions which can be asked in the census to questions which are framed in terms of the protected characteristics described in that Act.

5.3.3 Analysis and research

The topic consultation provided evidence that information on the sex variable in the census was essential for analysis and research conducted by a wide range of users, including public bodies and third and private sector organisations. For example, the UK Data Service said:

"The UK Data Service represents a very diverse group of users who use the data to identify a range of different small groups, including specific age ranges, ethnic minorities and groups defined by socio-economic traits such as workers or the self-employed." (2015: 16)

It is clear that opinion on the most appropriate approach to gathering data on sex is divided. In September, and more recently, NRS has heard from Scottish and UK academics that a self-identified approach to collecting sex data would produce inaccurate data. Their view is this approach is no longer a reasonable and accurate count of sex and it is leading to data on biological sex being lost which would have an impact on resource distribution in the health service, research in areas of social NRS Sex Question Recommendation Report Scotland’s Census Page 28 of 46 inequality and sex-based socialisation processes in childhood, analysis of care for aging population and sex based biases in the criminal justice system.

NRS has also received correspondence from researchers, academics, practitioners and data users working in universities, higher education agencies and other organisations across Scotland and the UK supporting the approach that NRS took in 2011 and propose to continue in 2021. Their view is that a departure from the proposed binary sex and self-identification approach poses a greater threat to data quality, particularly in the case of longitudinal studies and that data collected by NRS harmonises with the data collected by the Office for National Statistics (who carry out the census in England and Wales) and the Northern Ireland Statistics and Research Agency.

NRS acknowledges that some data users would want the sex question to be asked on a biological or legal sex basis however the information users require from the sex question has a multitude of uses and whilst the majority of users are content with the proposed approach for 2021 NRS recognise that there are other views.

5.4 Respondent need

In order to produce good quality data which meets user needs, the census is designed to maximise response to the census overall, and to maximise response to individual questions. It is essential that questions are acceptable, understandable and can be answered by all members of the population in Scotland.

There has been a sex question in the census since 1801 and no guidance was provided for this question until 2011. As the census is self-completion, to that extent the question has always been a self-identified one and it can be assumed that without explicit guidance, people answered the question in the way that they felt best reflected their sex.

As testing has shown, for most people their self-identified sex equates to their biological and legal sex and they do not require or access guidance to answer the sex question. The guidance in 2011 was specifically to support the trans population and allowed those with or without a gender recognition certificate to ‘select either ‘male’ or ‘female’, whichever you believe is correct, irrespective of the details recorded on your birth certificate’. Engagement with LGBT equality groups and other groups such as Engender has confirmed that for a binary question, the self-identified approach taken in 2011 would be the preferred approach in 2021 and a sex question that only allowed trans people who hold a GRC to self-identify as male or female would be less acceptable. The recent testing supported this view as it demonstrated that the legal sex guidance would have a negative impact on census completion for 40% of trans or non-binary respondents compared to just 3% for the self-identified guidance.

Evidence from trans individuals at Stage 1 of the Census (Amendment) (Scotland) Bill confirmed that they would be content to answer the sex question with their ‘lived sex’ and answer the subsequent trans status question affirmatively. There were also concerns raised around response rates and that a change in approach would be a backward step in trans equality practices in Scotland if the question was no longer self-identified:

"It is a government’s responsibility to make sure all its citizens are represented under the census…the possible negativities of requiring the compulsory sex question to try to force trans people to answer in a way that does not reflect how they live in society would most likely mean that trans people just don't take part." (Lawrence, 2018: 10)

"A move to force trans people to disclose either their biological sex characteristics at birth, or their legal sex, indicates a shift back to a viewpoint that either biological sex at birth is the “more important” characteristic, or to the view that only a small number of Gender Recognition Certificate holders are entitled to dignity and respect of their lived sex." (STA, 2018: 14)

"Whilst trans men and women with GRCs have both enhanced privacy rights and legal recognition of their lived sex, due to Equality Act protections it is now incredibly unusual for public bodies, services or employers to treat GRC holders and non-GRC holders differently." (STA, 2018: 14)

Whilst evidence from testing and engagement supports the proposed approach of a binary sex question asked on a self-identified basis, NRS acknowledges that there are a small number of respondents for whom the respondent need is not met. In recent testing the binary nature of the response options was the key reason why some trans or non-binary participants chose not to answer the sex question. NRS are aware that some non-binary respondents do not feel that they are able to respond to a binary female/male question honestly and this may have an impact on census response for this group of respondents.

6. Outputs

One of the fundamental aims of the census in 2021 is to generate good quality outputs that meet the needs of the wide range of users served by the census. As the testing has demonstrated, guidance based on a legal sex approach would lead to a higher non-response rate for the sex question, and for census completion rates, which would have an impact on the data quality.

Irrespective of the guidance, NRS cannot ensure biological or legal sex data would be collected as respondents answer without verification of responses and may ignore the guidance if they consider it unacceptable. Respondents may also choose not to answer on paper and in 2011 there was a 0.8% non-response rate for the sex question.

The final dataset contains noise from a variety of sources (statistical imputation and estimation and statistical disclosure control methods) and NRS do not believe that the data quality of the sex variable will be affected by guidance of a self-identified basis.

Evidence from the topic consultation and engagement showed that the information that users require about the sex question has a multitude of uses and it will often be combined with data from other variables. From engagement with stakeholders it is clear that a single question would not address all needs and a trans status question is proposed for inclusion in the census for the first time in 2021 and will provide data on the size, age and location of the trans population in Scotland.

How NRS expect people to answer based on self-identified guidance

Person Sex Trans status
Cisgender man Male No
Cisgender woman Female No
Trans man with GRC Male Yes
Trans woman with GRC Female Yes
Trans man without GRC Male Yes
Trans woman without GRC Female Yes
Non-binary Male/Female/Non-response Yes

 

How NRS expect people to answer based on legal guidance

Person Sex Trans status
Cisgender man Male No
Cisgender woman Female No
Trans man with GRC Male Yes
Trans woman with GRC Female Yes
Trans man without GRC Female Yes
Trans woman without GRC Male Yes
Non-binary Male/Female/Non-response Yes

The table shows that the two versions of the guidance produce a difference in the responses for the trans population. Self-identified guidance for the sex question would allow trans people who have and do not have a GRC to answer consistently. The legal sex guidance, in contrast, would produce inconsistencies in how people who have and do not have a GRC respond. When processing the data NRS will not be aware of which members of the trans population have a GRC and in order to analyse the combined data for the sex question and the trans status question, it is important that the trans population answer these questions in a consistent manner and the self-identified guidance allows them to do this.

6.1 Comparability and consistency over time

The outputs must also look to provide longitudinal consistency with previous censuses and comparability where possible with data produced by other UK census offices.

As in previous years, there will be separate censuses conducted by the Office for National Statistics (ONS) in England and Wales, and the Northern Ireland Statistics and Research Agency (NISRA) in Northern Ireland. The three census offices have worked together to develop a set of questions that, wherever possible and necessary, will deliver harmonised outputs across the UK.

Continuing with the same approach for the sex question in Scotland as in 2011 would maintain the consistency of the Scottish data over time and also allow for harmonisation across the UK. Such comparability over time, and with the UK, is one of the key concerns of data users to allow population changes to be seen both in a historical context and in relation to other countries.

There is no evidence to suggest that the addition of the guidance in 2011 introduced any discontinuity in the data and ONS are currently proposing to take the same approach as 2011 for the 2021 Census in England and Wales.

6.2 EU Regulations and their interaction with Scotland’s Census

The agreed understanding by National Records of Scotland and the Office for National Statistics (ONS) - who provide the EU with data on behalf of the UK - is that member states use a variety of sources to gather the relevant statistics. This is envisaged by Article 4 of Regulation (EC) 763/2008. The EU legislation on the 2021 population and censuses explanatory notes state:

“EU legislation on population and housing censuses is ‘output-oriented’. The authority and responsibility to develop appropriate census methods and technology remain with the Member States. EU legislation aims to provide census data that are comparable between the EU Member States. To achieve this, the data have to follow a European programme of statistical data and metadata. (2019: 7)”

The ONS, in consultation with National Records of Scotland and the Northern Ireland Statistics and Research Agency, will agree how best to present the UK statistics and how any differences in methodology will be described or handled so as to help users when interpreting the outputs. There is no requirement on the individual nations to constrain definitions and methodology to exact duplication of approach. However, in NRS Sex Question Recommendation Report Scotland’s Census Page 32 of 46 presenting the statistics to the EU, ONS will set out in a ‘quality assurance’ section any points relevant to interpretation or use.

This approach allows member states across the EU to comply with EU requirements whilst respecting different statistical structures within the member states. In the case of the sex question, National Records of Scotland are currently proposing to ask the same question as the rest of the UK– binary and self-identification, and on the same basis. ONS will bring these data together to provide the UK output required by Eurostat.

7.Conclusion

National Records of Scotland acknowledges that this is a sensitive topic that has attracted, and continues to attract, strong and varying opposing views. Having fully considered the results of testing, approaches taken across the UK, and the feedback from the discussions with stakeholders and peers, the recommendation from the National Records of Scotland’s Government Statistical Services statistician group continues to be that a binary sex question asked on a self-identification basis provides the best balance in meeting the diverse range of user needs across the full census dataset.

Our aim for the census is to deliver a set of questions and associated guidance that enables all of Scotland to access, understand and complete the census.