Scotland's census 2021: equality impact assessment

3. Key findings - protected characteristics

Age

Demographic overview

Scotland's population is ageing. In 2018, just under one in five people (19 per cent) in Scotland were aged 65 and over, compared with 16 per cent in 2008. The population aged 16 to 64 has decreased from 66 per cent to 64 per cent over the last 10 years. The higher number of females at older ages reflects their longer life expectancy^[4].

Internet use by age

There is a clear relationship between age and use of internet, with lower rates of internet use among older adults. In 2018, 100 per cent of adults aged 16 to 24 reported using the internet compared to 38 per cent of those aged 75 and over. Thirteen per cent of all adults stated that they did not use the internet at all. Although older adults were less likely to use the internet, the gap in internet use between adults aged 16-24 and adults aged 60 and above has fallen over time from 57 percentage points in 2007 to 35 percentage points in 2018^[5]. This result has mainly been driven by an increase in internet use amongst adults aged 60+ (from 29 per cent to 65 per cent).

Younger internet users were more likely to access the internet using a smartphone than older users, with 96 per cent of 16-24 year olds using smartphones compared to 29 per cent of adults aged 75 and above. This age divide can also be seen in the use of digital, cable or satellite television to access the internet and games consoles. Older internet users were more likely than younger users to use a tablet to access the internet. Younger people were less likely to have changed their use of the internet as a result of security concerns. For example, only nine per cent of those aged 16-24 and six per cent of 25-34 year-olds said that security concerns made them less likely to bank online, compared to 20 per cent of those aged between 60-74, and 23 per cent of those aged 75 and above.

What does this mean for different age groups of respondent?

Our experience of previous censuses has shown that older people need a range of support mechanisms to raise their awareness of the need to complete a census questionnaire and to help them to do so. However in light of the evidence above, the emphasis on online completion in 2021 will present fresh challenges in mitigating the risk of non-participation amongst older people. We will therefore ensure that paper questionnaires will be available

Information from the 2011 Census shows that 53 per cent of those aged 65 and over are limited in their ability to carry out day to day activities due to a health condition or a disability. This compares to 20 per cent for the population overall. Therefore we are working to ensure that our support services take account of additional needs. Public assistance will feature a number of impairment solutions including Telephone Data Capture, Relay and Braille. We have engaged with charitable bodies and support organisations such as RNIB to inform designs to assist their users and will continue to do so. We recognise that older people may lack confidence in using our dedicated contact centre, including automated systems such as Interactive Voice Response for paper questionnaire requests.

All of these factors are influencing design decisions and strategies around public assistance, community engagement, enumeration and fieldwork.

Contact Centre and field force staff will be able to provide assistance to aid completion. In addition, the census questionnaire can be completed on behalf of the householder by a family member/friend/carer. In the case of communal establishments (e.g. care/retirement homes) pre-census contact will seek to establish whether there are likely to be any difficulties around residents' completion of the questionnaire and if so we will make sure that there is appropriate help available. Our public assistance team will offer assisted digital support by phone and through local support hubs for those who lack digital skills. Language used in contact materials will be carefully considered to avoid discouraging those with low confidence. Clear guidance will be provided on how to contact public assistance to use the various support channels for those without the skills/confidence to complete on their own.

Older single person households may also be wary of doorstep contact by field staff. We are actively looking to recruit staff who have previous experience of working with vulnerable groups, subject to the necessary checks and clearances.

Question on age

Data on age, or date of birth, was first collected in the Census in 1821. There is a strong and well-established user need for the key demographic variable of age. By combining sex and age information, the census provides a basis for calculating rates of morbidity, mortality, fertility, marriage and divorce. These are vital inputs 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.

Social surveys generally collect information about age, either by asking for age or age bands, date of birth, or a combination. We propose to continue to collect information on age (via a question on date of birth) in 2021. No question development has been considered necessary for the date of birth question as the data collected in previous censuses was of good quality which met user needs.

As we are recommending the inclusion of a question to gather age data in 2021, data on this variable will be available in census outputs. Outputs relating to age are an important tool for creating a complete demographic picture of Scotland. Stakeholder feedback has indicated that these outputs are vital for resource allocation, targeted investment and service planning/provision across the country.

How are questions and question acceptability affected by age?

Question development work seeks to ensure that the language used in questions and guidance is inclusive, acceptable and aids people of different ages in answering questions confidently and correctly. This includes guidance for those who are completing returns on behalf of others who may be at school, retired, or young children. All potential questions are reviewed to ensure that they are asked of the appropriate age groups - specifically, questions on sexual orientation, trans status and language. Testing of all potential new questions, and any questions which are considering changes has been undertaken with communities of interest and the general population to understand the implications of posing these questions to all individuals, or targeting to relevant age groups, to understand how different age groups may respond to questions, and whether the questions are publicly acceptable to all age groups.

Our proposals seek to limit the asking of questions to certain age groups only

In 2021, we need to make sure that the census only collects the information that we need from the people of Scotland and we want to make sure that the questionnaire is easy to complete and takes as little time as possible. Testing has also shown that some questions are less acceptable when they are asked about people under the age of 16 years.^[6] With this in mind we are proposing to have some age routing in the questionnaire. This means for example that the person completing the form will not be asked how well a two year old can read English or where a 6 month old baby lived a year ago.

For the online questionnaire we propose that the response used to the date of birth question will be used to calculate a person's age and which questions will be relevant to them, so they will not need to respond to questions that are not relevant to them. On the paper questionnaire we will provide guidance about which questions should be answered by people of particular ages.

The following age-limitations are proposed:

Questions asked of people aged 16 or over:

• The new question on trans status or history
• The new question on sexual orientation
• The new question on Ex Service status
• The question on legal marital and civil partnership status
• The qualifications held question
• The questions on employment and activity last week

Questions asked of people aged 4 and over:

• The question on whether you are a school child or full-time student
• The questions on travel to place of work or study

Questions asked of people aged 3 and over:

• All of the language questions
• The question on provision of unpaid care

Questions asked of people aged 1 and over:

• The question on address 1 year ago

All other individual questions will be asked for everyone.

Communication considerations by age

Certain age groups may be unaware of the census or unsure of some of the language used. Some older people may experience difficulties in understanding advice and guidance.

Ensuring maximum participation from all age groups

Stakeholder and community engagement will seek to connect with key support and care organisations both nationally and at local level to build better understanding of the barriers to participation which may exist for older people and disabled people and in the longer term to identify the most effective routes for raising awareness and educating around the benefits of taking part and the accessibility options we will offer. This will further inform design and implementation of public assistance services and solutions as well as publicity and marketing. Reassurance around security of data will also form a key objective of our communications activities.

Young adults have also been identified as a group at risk of not participating in Census, and will therefore be the subject of targeted community engagement. Digital exclusion, whilst more prevalent in older people, also exists as a barrier to the participation of young adults to an extent. However, attitudinal and motivational factors such as lack of knowledge/awareness, disengagement with officialdom, and fears around security of data are key amongst this group, some of whom may be experiencing their first Census as a respondent, and there will therefore be an emphasis on raising their awareness of the obligation to complete a questionnaire, and highlighting the benefits of doing so.

Evidence^[7] suggests that using the internet and social media are a good way of targeting young people but other ways of reaching and engaging older people may be more effective.

Publicity and marketing campaigns will seek to maximise the potential for social media and other channels to target all age groups with tailored messaging. Messaging will also seek to reassure respondents by highlighting our commitment to keeping data secure.

Access to outputs

Evidence^[8] shows that older people are less likely to use the internet, so may not have as much access to data outputs if they are only available online. Census outputs will be made available in a variety of accessible formats both online and in hard copies on request. This will ensure that all data users, regardless of their internet access or proficiency will have access to census outputs.

Sex

Demographic overview

Mid-year estimates for Scotland for 2018 produced by National Records of Scotland show that 51 per cent of Scotland's population were female and 49 per cent were male.^[9]

What does the sex question measure?

The Census has collected information on the numbers of males and females since 1801. For the first four censuses, information was not recorded on an individual basis, only on the total number of males and females per address. Since 1841 the Census has asked individual households to record details of the person who are resident their including their 'sex'.

The Census Act 1920 is the legal framework for the 2021 census and does not define sex . Paragraph 1 of the schedule to the Act lists "Sex" as a matter in respect of which particulars can be required.

Since 1920, a variety of questions have been asked/instructions given and the question asked in different looking formats:

• 'Please tick appropriate box'
• 'Write in your sex' or
• 'What is your sex?'

The response options have always remained constant and it is therefore assumed – in the absence of any evidence to the contrary – that the data are consistent over time.

Prior to 2011 in Scotland, there was no additional guidance provided on how to answer the sex question. In 2011, in response to user requests, additional information was provided online to help trans people understand how they should answer this question. That guidance advised people who were trans that they did not need to answer the question with the sex recorded on their birth certificate.

The Census Topic Consultation in 2015 identified a continued need for data on sex to be collected in the Census, and in addition identified a need to consider how the question is inclusive of everyone in Scotland. Some stakeholders raised that the 2011 binary sex question could not be answered by those who were non-binary; this was evidenced in 2011 by some respondents ticking both boxes, writing in 'non-binary' over the response or refusing to tick either box. Investigating how to address this was considered important as the sex question is not a voluntary question and therefore completion rates should be 100 per cent.

A programme of research and development was therefore undertaken, the findings from which were published in September 2018. The topic report^[10] reported the findings that a non-binary sex question was publically acceptable and produced less item non-response than a binary sex question. National Records of Scotland reported in the Plans for Scotland's Census^[11] in September 2018 that it was continuing to investigate whether a non-binary sex question would lead to improvements in data quality for 2021.

During the consideration of the Census (Scotland) (Amendment) Bill, there was considerable discussion around the sex question to be asked in 2021, in particular, whether a non-binary question should be considered further. Whilst some stakeholders felt that such a question could lead to better result, many stakeholders raised concerns over the concept of a non-binary question. In particular, there was a widely expressed view amongst these stakeholders that sex is biological and can only be male or female and therefore any question which asks a respondent about their sex must be binary. In addition to the need for the question to remain binary, these stakeholders are clear that if sex is based on biology then any guidance must be clear that the basis of the question is around sex as recorded on an individual's birth certificate or Gender Recognition Certificate and not on the basis of how they self-identify.

In its report on Stage 1 report on the Census (Amendment) (Scotland) Bill^[12], the Culture, Tourism, Europe and External Affairs Committee of the Scottish Parliament recognised that there are different views on the issue of binary or non-binary sex question but recommended that the question remains a binary one in 2021 in order to ensure consistency over time. Having discussed with stakeholders and considered the matter further, it has been agreed that a binary sex question will be proposed for 2021.

The issue of whether or not the question remains a matter of self-identification in 2021, as it was in 2011, remains very much alive. The current proposal is to continue with the same question as in 2011 'What is your sex?' and to update the guidance used in 2011 but still retaining the concept of self-identified sex. The guidance would remain online and separate to the census questionnaire. We recognise that there are very strong opposing views on this issue. National Records of Scotland recognises that there are very strong opposing views on this issue. NRS commissioned ScotCen Social Research to test the nature and impact of guidance associated with the sex question. ScotCen carried out two within subject surveys on: the general population aged over 16 and living in Scotland; and adults aged over 16 who self-identify as trans / non-binary and living in Scotland. The full ScotCen report can be found here. Following careful consideration of all of the available evidence NRS continue to recommend a binary sex question with self-identification guidance for the 2021 Census. The NRS Recommendation Report can be found here.

We continue to work with our data users to understand what data they need from the census to do their work. A version of the equality questions and guidance can be seen in Annex A.

As part of their report on the Census (Scotland) (Amendment) Bill, the Committee also noted that whatever the outcome of the considerations around guidance, the final guidance should be clear, clearly communicated to all and accessible.

Sexual Orientation

Demographic Overview

There is currently limited data and evidence collected on the experiences of lesbian, gay and bisexual (LGB) people in Scotland.

A question on sexual orientation is now asked in the Scottish Household Survey, the Scottish Health Survey and the Scottish Crime and Justice Survey as one of the Scottish Government's "core" questions. Developed by the Office for National Statistics, the question was designed to provide accurate statistics to underpin the equality monitoring responsibilities of public sector organisations and to assess the disadvantage or relative discrimination experienced by the lesbian, gay and bisexual population. It should be noted that estimates on self-identified sexual orientation from this question are likely to under-represent the lesbian, gay and bisexual population. According to the Scottish Household Survey around two per cent of all adults self-identified as lesbian, gay or bisexual in 2018 compared to one per cent in 2011. In 2011 and 2018, 0.4% of adults preferred not to provide a response to the question.^[13]

According to a recent UK study^[14] by Office for National Statistics, in 2017,1.9 per cent of Scotland's population identified as LGB.

Question Development and Statistical Outputs

The Census has not previously asked about sexual orientation but it is proposed that a question be included in 2021.

Scotland's equality evidence strategy 2017-2021 states that data on sexual orientation had improved in recent years. However, gaps persist, and official sources are likely to undercount the proportion of the population who are lesbian, gay or bisexual.

Considerable user demand has been identified from the 2015 Topic Consultation for the collection of information on sexual orientation in the 2021 Census. The main requirement identified is in relation to the monitoring and reporting duties for public bodies, and service planning and provision. The information is also required to inform equality impact assessments, which in turn inform policies and practices. The Equality and Human Rights Commission require this information to use in a statutory review of equality and human rights, which is carried out every five years.

Research and development carried out across UK census offices found a sexual orientation question (asked of those aged 16 or over) was generally acceptable to the public and the majority of respondents would provide a valid response. However, acceptability decreased with age: while only 4 per cent of those aged 25-34 years and 11 per cent of 35-44 years indicated the inclusion of a sexual orientation question in census was unacceptable, 27 per cent of those aged 65-74 and 30 per cent of those aged 75 or over indicated that a question of this nature was unacceptable. ^[15]

Overall 14 per cent of the public said they would not answer a sexual orientation question if it was included in the 2021 Census. The majority of these (13 per cent) said they would skip the question and continue completing the rest of the form. Only a very small proportion of participants said that they would request an individual form (less than 1 per cent) or stop completing the census altogether.

Clarification on why information regarding sexual orientation is required and additional reassurances of information security should go some way to addressing reluctance to answer the question.

In the context of completing the census on behalf of another household member, the proportion who found the question not acceptable increased to 20 per cent. Similarly, just over one in five people (21 per cent) indicated that they were not comfortable with providing this information on behalf of others.

Testing of the questions showed almost all participants provided a valid response to the question on sexual orientation. Of those who did not provide a response to the sexual orientation question, the majority were aged 65 or over (59 per cent). Less than 1 per cent of participants provided an invalid response to the question. The question was voluntary and, as such, 9 per cent of participants chose not to provide an answer.

On this basis, it is proposed that a question in the 2021 Census be asked on a voluntary basis and the Census (Amendment) (Scotland) Act 2019, now allows this. The current proposal is that this question would only be asked of those aged 16 years and older.

Some concerns have been expressed during the question development process about confidentiality in relation to sensitive questions, particularly for young people who may feel unable to answer the question if they still live at home with their families. Missing this group would be an issue as young people have specific service needs. To provide complete privacy and confidentiality for any person responding to the 2021 Census, the facility to request and receive an individual questionnaire for completion in confidence will be available to all people over the age of 16 who are capable of completing a return. Any such individual will be able to complete an individual form without other members of the household being aware.

We continue to engage with Scottish Government teams and other stakeholders to ensure the outputs from the census and other surveys are comparable. This will enable data users to use a full range of outputs on sexual orientation to promote equality and identify inequalities.

Data Collection and Enumeration

It is recognised that there may be privacy issues around response to sensitive questions, albeit they will be voluntary, for residents of communal establishments who are unable to complete a questionnaire themselves and who will rely on others to record the necessary information. National Records of Scotland will work with communities of interest as well as communal establishment managers to identify the best approach to enumeration in these cases.

Communications, Stakeholder Engagement, Publicity and Marketing

National Records of Scotland is actively engaging with groups and organisations that represent the interests of LGB people.

We recognise that privacy issues may be key to respondents around sensitive topics such as sexual orientation. We are actively exploring relevant concerns through our discussions and engagement with communities of interest and they will form a key focus of our assessment activities going forward.

In the event that questions are asked in the census around sensitive topics such as sexual orientation, due care will be taken to ensure that guidance and publicity materials, including those which support the questionnaire will provide clarity and education to the general public around terminologies, definitions, concepts and language, and individual privacy both in responding to the census, and in how the data is output for use.

Gender Reassignment and Trans

Demographic Overview

A Gender Identity Research and Education Society (GIRES) publication in 2011, funded by the Home Office, estimates that prevalence of people who had sought medical care for gender variance is increasing. Its original report had estimated prevalence in the UK in 2007 as 20 per 100,000 people having sought medical help for gender variance. This represented 10,000 people of whom 6,000 had undergone gender transition. Of these, 80% were birth-registered males choosing to become females, although this percentage is noted to be dropping.^[16]

Trans and the Equality Act 2010

Part 2 Chapter 1 Section 7 of the Equality Act 2010 defines the protected characteristic of gender reassignment for the purposes of the Act as where a person has proposed, started or completed a process to change his or her sex. A transsexual person has the protected characteristic of gender reassignment.

Section 7 also explains that a reference to people who have or share the common characteristic of gender reassignment is a reference to all transsexual people. A woman making the transition to being a man and a man making the transition to being a woman both share the characteristic of gender reassignment, as does a person who has only just started out on the process of changing his or her sex and a person who has completed the process

The Equality and Human Rights Commission website notes that :

'In 2016 a Women and Equalities Committee report made over 30 recommendations calling for government action to ensure full equality for trans people.

One of the report's recommendations was that the use of the terms 'gender reassignment' and 'transsexual' in the Equality Act 2010 are outdated and misleading, something that we fully agree with. The preferred umbrella term is trans'.

Transgender (or "trans") is a term for people who are proposing to undergo, are undergoing or have undergone a process to transition to a new gender identity. We use the term trans in the sections that follow.

Question Development and Statistical Outputs

The Census has never previously asked questions around either gender reassignment or trans status. The 2015 Topic Consultation highlighted a need for information on 'gender identity'. Further consultation with data users refined this data need to being about the size and geography of the trans population – both those who currently identify as trans as well as those who might have a trans history. Given the lack of alterative data sources, and the small populations, this makes Census the only statistical collection likely to gather robust data on the trans population. A key reason for requiring census data on trans status is to be able to fulfil the public sector equality duty (see section above in relation to the Equality Act). Census data would, for example, assist public authorities in carrying out equality impact assessments when they assess and review policies and practices.

Following stakeholder engagement to fully understand data needs, National Records of Scotland tested a trans status or history question, alongside the sex question to replicate responses as they would be perceived in the census itself. Testing found the trans status or history question was acceptable to members of the trans community and to the general population, and produced good quality data.

Respondents were able to answer the question on trans status or history with ease on behalf of themselves - around 94 per cent of respondents provided a valid response to the question on trans status or history. Respondents indicated they were comfortable answering on behalf of another member of their household if they had their permission to do this. Stakeholders had a range of views on potential age limits for asking the trans status or history question. Whilst some stakeholders suggested age limits of 16 or below 16 years (such as 12 or 13 years), others suggested it should be asked of all regardless of age. Respondents in cognitive testing indicated the question should be asked of all regardless of age.

Public acceptability testing for the gender identity topic^[17] was carried out in January – March 2017, for the three UK census offices. In Scotland, 5,000 households were invited to take part. Public acceptability testing is designed to explore the views of the public on the acceptability of including sensitive questions in the census, thereby identifying particular sensitivities and potential barriers to public confidence and exploring mechanisms for overcoming concerns. This showed, in the context of providing an answer on behalf of another household member aged 15 or under, the proportion who found the question acceptable decreased from 74% to 58%, and the proportion who found the question unacceptable was 16% (compared with 9% when asked of those aged 16 or over). The proportion who were undecided increased from 16% to 26%. The acceptability testing also highlighted the proportion of the public who reported they could not answer accurately for any members of their household increased from 4% when asked of those aged 16 and over, to 9% when asked of those aged 15 and under. The results of public acceptability testing showed that while the general public found the inclusion of a question was acceptable, acceptability decreased if asked of those aged under 16 years. Therefore, whilst some need for data on trans status of under 16 year olds was identified, asking the question of those aged 15 and under was less acceptable.

Testing has shown that additional guidance in the question stem of the trans status or history question enables members of the general population to have a better understanding of the terminology and answer the question.

Some concerns have been expressed during the question development process about confidentiality in relation to sensitive questions, particularly for young people who may feel unable to answer the question if they still live at home with their families. Missing this group would be an issue as young trans people have specific service needs. To provide complete privacy and confidentiality for any person responding to Census 2021, the facility to request and receive an individual questionnaire is available for anyone aged 16 years and above, who is capable of completing. Any such individual will therefore be able to complete an individual form without other members of the household being aware.

As with the sexual orientation question, it is recognised that this is a sensitive question and no-one should be compelled to answer it. The Census (Amendment)(Scotland) Act 2019, allows a trans status or history question to be asked in the Census on a voluntary basis so long as they are specifically prescribed for that purpose in a Census Order. The draft Census (Scotland) Order includes provision prescribing the trans status or history particular for that purpose.

We will engage with Scottish Government teams and other stakeholders to ensure the outputs from the census and other surveys, where a question on trans status or history are asked, are comparable. This will enable data users to use a wider range of outputs on trans status or history to promote equality.

Communications, Stakeholder Engagement, Publicity and Marketing

National Records of Scotland is actively engaging with groups and organisations that represent the interests of trans people.

We recognise that privacy issues may be key to respondents around sensitive topics such as trans status or history and are working to ensure these are addressed.

Assuming that a question on trans status or history is asked in the 2021 Census (as is provided for in the draft Census (Scotland) Order), due care will be taken to ensure that guidance and publicity materials including those which support the questionnaire will provide clarity and education to the general public around terminologies, definitions, concepts and language, and individual privacy both in responding to the census, and in how the data is output for use.

Disability

Demographics Overview

According to the 2011 Census, the proportion of people in Scotland with a long-term activity-limiting health problem or disability was 20 per cent, the same proportion as reported in 2001. The disability rate has stayed the same despite Scotland's ageing population.

A higher proportion of people in Scotland reported a long-term activity-limiting health problem or disability than the UK as a whole (18 per cent). However Scotland recorded a lower proportion than Wales (23 per cent) and Northern Ireland (21 per cent).

Internet usage by disability

Data from ONS indicates that disabled people are less likely to be frequent users of the internet, particularly those over the age of 75; 39 per cent of disabled adults in this age group were recent internet users, compared with 49 per cent of non-disabled adults.

What does that mean for how we support disabled respondents?

Disabled people are recognised as a group who are at risk of non-participation in Census. Scottish Household Survey data from 2018 show that 27 per cent of adults who have some form of long-standing physical or mental health condition or illness reported not using the internet, compared with eight per cent of those who do not have any such condition.^[18] We are therefore developing and designing a wide range of support services and solutions to help mitigate the risk of non-participation.

Extensive stakeholder engagement has indicated those with certain disabilities or impairments may have specific needs and/or may have difficulty completing a census questionnaire. They may also have challenges in accessing or understanding contact materials and guidance.

Individuals with certain impairments may be more likely to have poorer digital skills and/or confidence.

Public assistance services will offer a wide range of support products including British Sign Language (BSL) translation, Braille questionnaire guidance and large print. These can be requested by calling our Contact Centre, web chat, eForm and Textphone. Live interpretation for BSL users will also be available. Design of contact materials will give consideration to impairments to ensure that it can be read and understood.

BSL translations of the questions will be available online, on DVD or on a USB stick. They can be used to help fill in the questionnaire online. BSL clips for each question will be accessible online on YouTube. A free video relay service will be available to request any of these products or simply to ask a question.

When enumerating some types of communal establishments, we will make prior contact with the manager of the establishment to identify the specific needs of any residents in advance so that we can ensure the correct support is in place.

Question Development and Statistical Outputs

Data about general health and long term health problems or disability is used by central government, local government and public bodies to identify health and social care service needs and to inform resource allocation at national and local level. It is also widely used to inform service planning and develop, monitor and assess policies on population health and health inequalities. There has also been extensive use in multivariate analysis undertaken by a range of users, including academics and research institutes. A suggestion made to use an impairment based model of health as a framework for data collection was not supported by respondents to the Scotland's Census Topic Consultation in 2015. Engagement with stakeholders highlighted the 2011 questions met data needs, but concerns around data quality were noted. Question development for 2021 has focused on improving the data quality, as comparable data over time was highlighted as a priority for users.

The Census outputs website will be designed to conform with accessibility guidelines to ensure ease of use for all users. National Records of Scotland is exploring the possibility of making appropriate outputs available in BSL.

Communications, Stakeholder Engagement, Publicity and Marketing

Extensive engagement with stakeholder organisations who represent disabled people has been and will continue to be undertaken to better undertand their needs and circumstances and the barriers to participation they may experience. Learnings will influence field force and contact centre staff training and the design of help and guidance.

We aim to make our website as accessible and usable as possible for all, including older audiences and those with visual, hearing, cognitive or motor impairments. This includes the use of various assistive technologies by allowing navigation using a keyboard only, by using easier-to-read colours, larger fonts and plain English.

The overall objective for content, such as question help, web and printed content, website navigation and user help and assistance is to support the user journey and user experience of Scotland's Census 2021 to be as seamless, efficient and quick as possible. It will align with the Scottish Government's Digital First Service Standards. All such written content will have the target reading level of an average nine-year-old's reading ability. This will enable people with a wider spectrum of literacy skills to make use of the service. We will strive to ensure that the tone of our content is accessible, authoritative, friendly and helpful.

When planning stakeholder engagement events venues are selected to ensure that they are accessible for disabled people. During the events, further steps are taken to ensure that disabled people are not precluded from participation, for example through the provision of a hearing aid loop for attendees with hearing difficulties.

Race

Demographic overview

The 2011 Census showed Scotland to be a more ethnically diverse nation than in 2001. Despite its increased diversity, Scotland was still a much less ethnically diverse country than England in 2011. Visible minority ethnic groups doubled, from 2 to 4 per cent of Scotland's population compared with 15 per cent in England. The non-British white group also increased, from three to four per cent of the population (127,000 to 222,000 people).

Between 2001 and 2011, Scotland's 'Asian' population doubled (an increase of 69,000 people), and the 'African, Caribbean or Black' population increased more than fourfold (by 28,000 people). 'Mixed or multiple' and 'Other ethnic group' non-white groups also showed an increase. People who identified as one of the 'Asian' categories represented 2.7 per cent of the Scottish population in 2011 compared to 1.4 per cent in 2001. Each minority ethnic group made up a larger proportion of the population in 2011 than in 2001. The vast majority (82 per cent) of those who recorded within the 'African, Caribbean or Black' group identified as 'African' in 2011. 9 per cent of the group identified as 'Caribbean' and 7 per cent as 'Black, Black Scottish or Black British'. In 2011, people who identified as 'Pakistani' made up the largest Asian group in Scotland, followed by those of 'Chinese' and 'Indian' ethnicity. 'Bangladeshi' remained the smallest Asian group of the categories listed on the 2011 Census questionnaire.

A separate 'White: Gypsy/Traveller' response category was added to the Census questionnaire in 2011. 4,200 people recorded their ethnic group within this category (0.1 per cent of all people in Scotland).

The last official count of Gypsy/Travellers living on sites in Scotland was carried out in 2009 and included council, Registered Social Landlord and private sites along with unauthorised encampments. This count estimated that around 2,000 Gypsy/Traveller people lived on sites in Scotland. This is around half the number who recorded as 'White: Gypsy/Traveller' in the 2011 Census, indicating that a significant proportion of Gypsy/Travellers lived in settled housing.

Question Development and Statistical Outputs

A question on ethnic group has been asked since 1991. As well as meeting previous legislative obligations, the data is used for resource allocation by central and local government.

Scotland is a culturally diverse nation and ethnic group is one of the most widely used census variables. Information on national identity and religion complements that on ethnic group. It has contributed to developing the Scottish Government's Race Equality Framework and Action Plan^[19] and has proved useful in helping to understand the links between national identity and ethnic identity, according to respondents to the Topic Consultation. Information on national identity and religion when used in conjunction with ethnic group data, has also been helpful in developing a fuller understanding of cultural identity.

In addition to ethnic group, information on country of birth is extensively used for a range of purposes, including area profiling, equality monitoring and to identify local areas which have experienced in-migration. It has supported ongoing work to understand the impact of migration and to assess at local level the extent and speed of integration. Given the recent shifts in migration patterns, understanding the origin country of migrants has become increasingly important, in order to better anticipate needs.

Information on ethnic group is used for resource allocation, to inform policy development and make service planning decisions. It also helps organisations meet and monitor their statutory obligations arising from Equality Act 2010 - over half of all respondents to the 2015 Topic Consultation told us that they use ethnic group data for equality monitoring purposes. Collecting this information in the census is particularly important because many minority ethnic groups in Scotland are too small to be effectively captured by sample surveys, and the census gives the only robust information on size of groups at small area level.

Data users have identified that the need to collect data that is consistent over time is important in order to monitor change over time. We continue to actively working with stakeholders to better understand how the question could be improved to meet need, whilst allowing consistency across time and recognising that public acceptability around the use of language in relation to ethnic group has changed substantially over time.

A large amount of work was done to review the ethnic group categories for the 2011 Census. In planning for 2021, National Records of Scotland undertook further question research and development in light of requests made for response options and terminologies to be reviewed and/or additional information to be collected. This includes engagement with a wide range of ethnic groups and populations using focus groups, surveys and stakeholder events, both independently and in collaboration with our colleagues in the Office for National Statistics, who are responsible for the census in England and Wales. This engagement increases the understanding of what user and respondent needs are now, language and concepts that are acceptable, and how data quality from an ethnic group question could be improved. Investigation was also undertaken around how other information collected in the Census, specifically religion and country of birth, can be used to improve the evidence base on ethnicity.

One of the major concerns which was raised in 2021 was around the concept of what is being measured under the 'ethnic group' question. This question has been asked in a broadly similar way since 1991; the response options used in the 1991, 2001 and 2011 Census questions combined concepts of colour, nationality, and ethnic or national origins. This measurement in the UK is influenced by the legal framework (The Equality Act 2010), which specifically refers to 'racial grounds' – namely colour, nationality, and ethnic or national origins.

One of the main criticisms of the ethnic group classification is its inconsistent use of colour and geography. In the review of ethnicity classification prior to the 2011 Census, many of these issues were explored.

Ethnicity is a complicated and sensitive concept. The format and content of this question has changed with each census to reflect this and best meet user needs, while retaining an element of comparability over time and with the rest of the UK.

Through consultation, research and question testing for the 2021 Census, National Records of Scotland continue to hear opposing views on the acceptability of descriptors in the ethnic group question. The consultation for 2021 highlighted a need for continuity with 2011 and/or earlier censuses, particularly to enable monitoring of equality related policy and service delivery. Changes to the existing categories would prevent this. For the 2021 Census, National Records of Scotland proposed to consider question testing and research requirements in light of a limited number of requests which were made for the response options to be reviewed and/or additional information to be collected within the existing question format to retain comparability over time.

In light of all of the consultation and discussion, a limited number of changes are proposed for recommendation to the Scottish Parliament. These changes, and the evidence supporting the conclusions are described below.

Information on Sikhs

A strong user need has been expressed for collecting information about the Sikh population in Scotland which has historically been captured through data about religion. A question on current religious belonging has been included in Scotland's Census since 2001, and this information is also captured through the three largest surveys in Scotland: the Scottish Household Survey (SHS); the Scottish Health Survey (SHeS); and the Scottish Crime and Justice Survey (SCJS). Stakeholders expressed concerns that using religion as proxy for the Sikh population risks undercounting the population, and raised concerns that some of the population do not identify with the options included in the 2011 Census ethnic group question.

The evidence for a potential undercount of the Sikh population through the religion question in the 2001 Census and 2011 Census is less strong. In 2001, 6,572 people identified as Sikh through the current religion question, and 6,821 people identified their religion of upbringing as Sikh, a difference of 3.7 per cent. Further analysis of the of write ins for "Sikh" in the ethnic group question in 2011 by ethnic group category by the respondents' response to the 2011 religion question showed that 9,055 people identified as Sikh through the question on current religious belonging. There were 873 write in responses for "Sikh" in the ethnic group question. Of these respondents, 26 (3 per cent) did not identify as Sikh in the religion question.

A tick box for "Sikh, Sikh Scottish or Sikh British" was then tested in focus groups with Sikh participants. Within each group, participants were asked to respond to the 2011 Census questions on religion, national identity and ethnic group on paper. Following this, participants were asked to feedback on the acceptability, quality and clarity of two different designs of the ethnic group question and to compare these designs with each other and the 2011 Census question.

A "Sikh, Sikh Scottish or Sikh British" tick box was included under "Other ethnic group" in one version of the ethnic group question and under "Asian, Asian Scottish or Asian British" in another version. In addition, a tick box for "Sikh, Sikh Scottish or Sikh British" was included for qualitative survey testing in Scotland under the "Other ethnic group" category following the tick box for "Arab, Arab Scottish and Arab British". Key results from cognitive focus groups and the 2019 qualitative survey test are described below. More detail is available in the Topic report published on the website. ^[20]

• The inclusion of a Sikh tick box was found to be acceptable to many focus group participants. However, there were some strong acceptability issues. Some participants found the inclusion of religion under ethnic group unacceptable, inappropriate or confusing. This made them question what they were being asked for. Some felt that religion and ethnic group should be separated or that this was repetitive since they had already selected Sikh under religion. Participants who found the tick box acceptable, and those that didn't, questioned why Sikh was the only religion included. Some felt singled out or that inclusion would only be acceptable if other religions were included or the question wording was changed. In comparison, there were very few acceptability issues with the 2011 Census question.
• The inclusion of Sikh as was largely acceptable in the qualitative online survey, but the placement of Sikh under "Other ethnic group" continued to cause difficulty for some respondents.
• Focus group participants who naturally identified as "Scottish Sikh" or "British Sikh" typically found a question including a Sikh tick box easier to answer because they could locate a response more easily. However, other participants were clear they would not select a Sikh option in ethnic group, found the inclusion of the tick box confusing and found it more difficult. Some participants multi-ticked or initially missed the Sikh tick box and later changed their answer. These results suggest that a Sikh tick box would undercount the population and that there may be other impacts on data quality.
• In the qualitative online survey, a larger proportion of the Sikh population could be identified as Sikh through the alternative question. However, the religion question provided the best information on the Sikh population.
• There is a risk that if data from ethnic group was used alone, this figure would undercount the Sikh population. National Records of Scotland will consider how equality related outputs are presented to meet user need.
• Acceptable comparability at category level would be achieved with a Sikh tick box under the Asian category. Comparable category level data would otherwise be achievable by aggregating a Sikh tick box under "Other ethnic group" with the Asian category. Acceptable comparability at category level, which might be achieved by aggregating different groups, would allow for harmonisation with census statistics across the UK.

On the basis of these findings, we do not propose to include a Sikh tick box under the ethnic group question in 2021.

Information on the Jewish population

A strong user need has been expressed for collecting information about the Jewish population in Scotland which has historically been captured through data about religion. A question on current religious belonging has been included in Scotland's Census since 2001, and this information is also captured through the three largest surveys in Scotland: the Scottish Household Survey (SHS); the Scottish Health Survey (SHeS); and the Scottish Crime and Justice Survey (SCJS). Stakeholders expressed concerns that using religion as proxy for the Jewish population risks undercounting the population. They highlighted the difference in the numbers identifying as Jewish in the current religion question included in the 2001 census (6,448) and in the religion of upbringing question included in the 2001 census (7,446), and their continued application of this difference (13 per cent) to provide an estimate for the size of the Jewish population in Scotland.

The religion of upbringing question was not included in the 2011 Census. Analysis of the number of write ins for "Jewish" in the ethnic group question in 2011 by ethnic group category by the respondents' response to the 2011 religion question showed that 5,887 people identified as Jewish through the question on current religious belonging. There were 812 write in responses for "Jewish" in the ethnic group question. Of these respondents, 219 did not identify as Jewish in the religion question.

A tick box for "Jewish, Jewish Scottish or Jewish British" was included for qualitative survey testing in Scotland under the "Other ethnic group" category following tick boxes for "Arab, Arab Scottish and Arab British" and "Sikh, Sikh Scottish or Sikh British" and the key results from the 2019 qualitative survey test are described below. More detail is available in the Topic report published on the website. ^[21]

• The inclusion of Jewish was largely acceptable. The placement of a Jewish option under "Other ethnic group" was mostly acceptable, but caused difficulty for some respondents.
• Results indicate that the best estimation of the Jewish population is likely to be the estimate obtained from the religion question (around 90 per cent) plus those who identify ethnically only by responses across ethnic group categories (around 10 per cent). We will consider how outputs can be created to best meet this user need.
• Including a tick box improved data quality on the Jewish population gathered by ethnic group but did not improve estimation of the total population size in the test. There was no evidence that including a tick box increased the capture of people who identified as ethnically Jewish but not religiously Jewish in this test. These people mostly chose to express their Jewish identity by writing this in to the ethnic group question with 2011 Census response options.
• Including a prompt to write in to indicate to respondents that Jewish is an acceptable response to the ethnic group question would likely result in the same overall improvements in data quality and may reduce issues with acceptability of the placement of a tick box
• Acceptably comparable category level data would be achievable by aggregating a Jewish tick box with the "White" category. Acceptable comparability at category level would allow for harmonisation with census statistics across the UK.

As a result of this work, we intend to include a prompt to write in "Jewish" in the "Other ethnic group" category

Information on the Roma population

A strong user need has been expressed for collecting information about Roma in Scotland. There are no alternative sources for this information. Research in England and Wales by ONS showed "Roma" was considered the most appropriate term to use for the Roma community and that the placement of a "Roma" tick box under the "White" high-level category was acceptable, with participants feeling that they should be placed in close proximity to the "Gypsy or Irish Traveller" tick box. This placement made it easier for respondents to locate the tick box. A tick box for "Roma" was included for qualitative survey testing in Scotland under the "White" category following the tick box for "Gypsy / Traveller" The key results from the 2019 qualitative survey test are described below. More detail is available in the Topic report published on the website. ^[22]

• The inclusion and placement of a "Roma" tick box was acceptable and preferred by most respondents. However, there were a small number of Roma respondents to the 2019 qualitative survey.
• The alternative question improves data quality to meet user need. All respondents identified as Roma in the alternative question by combining responses to the Roma tick box with write in responses compared to half in the question with 2011 Census response options.
• Although the alternative question improved data quality to meet user need, not all respondents chose to select the "Roma" tick box. While the location of the tick box has been shown to improve data quality, be generally acceptable, and easy to locate, National Records of Scotland will consider how outputs can be created across ethnic group categories to best meet user need.
• Acceptable comparability over time is expected at category level. Comparability at category level will also allow for harmonisation with census statistics across the UK. ONS have also recommended that for 2021, the ethnic group question in England and Wales should include a tick-box for "Roma" within the "White" category under "Gypsy or Irish Traveller", providing for the collection of comparable data on Roma.

As a result of this work we are proposing including a Roma tick box under the White category.

Information on Showpeople

A strong user need has been expressed for collecting information about Scottish Showpeople in Scotland. There are no alternative sources for this information. A tick box for "Showpeople" was included for qualitative survey testing in Scotland under the "White" category following tick boxes for "Gypsy / Traveller" and "Roma". The key results from the 2019 qualitative survey test are described below and more detail is available in the Topic report published on the website. ^[23]

• The inclusion of a tick box for "Showpeople" was highly acceptable, with the majority of respondents selecting this option in the alternative ethnic group question.
• The alternative question improved data quality to meet user need. However, many write in's across the ethnic group questions in this test were for "Showman" or "Showperson" rather than "Showpeople". A tick box for "Showperson" rather than "Showpeople" may improve both acceptability and data quality
• While the location of the tick box has been shown to improve data quality, be highly acceptable, and generally easy to locate, National Records of Scotland will consider how outputs can be created across ethnic group categories to best meet user need.
• The alternative question design provides acceptable comparability over time at category level. Comparability at category level will also allow for harmonisation with census statistics across the UK.

As a result we are recommending the addition of Showman/Showwoman tick box in the White category.

In addition to these changes, we are also proposing a design change to the "African" category to improve data quality, and a design change to the "Caribbean or Black" category to improve data quality. More details on this can be found in the Topic report.^[24]

Data Collection and Enumeration

People from some ethnic groups may require information to be available in a range of community languages; the guidance sheet will be available in 24 languages. Our public assistance channels will also ensure that translation and interpreting advice is available and respondents will be able to request a call in their own language to help answer any questions they might have.

We are taking forward a programme of stakeholder engagement with a range of communities including, for example, Gypsy/Travellers, Roma, Showpeople, Sikh, Muslim, Jewish, to encourage and support increased participation in Scotland's Census 2021 and to inform enumeration strategies and processes, which will be tailored to specific identified needs and circumstances where necessary. In particular, we wish to ensure that the approaches and definitions used make sense to all groups – such as the concept of a household – in order to ensure that we can collect the necessary information and ensure that all people are able to be counted.

Communications, Stakeholder Engagement, Publicity and Marketing

Language, lack of awareness and understanding will be significant barriers to participation in the census for some ethnic groups and national identities. Support materials, messaging and guidance will be available in languages other than English. The programme will explore and test solutions such as the use of specialist media and translation services. Field staff training, enumeration processes, publicity, marketing and all communications will take account of the need for appropriate language and cultural sensitivities. The following types of language support will be available:

• A live interpretation service via the Contact Centre for up to 200 languages.
• A language support sheet signposting to help and support in 24 languages.
• Translated guidance in 15 community languages and Gaelic to aid completion of the paper questionnaire.

Religion and Belief

Demographic overview

Scotland became a less religious country in the decade to 2011 according to the 2011 Census. Almost two fifths of the population (2 million people) stated they had no religion in 2011, an increase of over half a million people from 2001.

Those reporting a 'Christian' denomination represented the majority of the Scottish population (54 per cent). The next largest religion was 'Muslim' which represented over 1 per cent and the other religions combined (including 'Hindu', 'Buddhist', 'Sikh' and 'Jewish') represented a further 1 per cent. 7 per cent did not state their religion.

The number recording their religion as 'Muslim' increased by 80 per cent and those reporting as 'Hindu' almost trebled.

Question Development and Statistical Outputs

The question on religion was introduced in the 2001 census, and its inclusion was allowed on the basis that answering it was voluntary. Consultation with users has shown that public bodies use the census information on religion to assist with monitoring discrimination, linked to the introduction of the public sector equality duty. The data has also been used to inform service provision for health, social care and education.

Data on religion gathered in the census is used by a range of users, including central government, local government, public bodies and religious organisations, to plan and deliver services. It is widely used by local government in equality monitoring, area profiling and to identify demand for denominational schools. The data are also used for planning a range of services and for research and analysis.

There are a number of ways in which a question on religion can be asked and these give rise to different results. Respondents to the Topic Consultation in 2015 and stakeholder engagement noted that continuing to ask a question in 2021 which was conceptually similar to that asked in previous censuses was important in order to be able to track changes over time and for monitoring purposes, and continuity with 2011 and/or earlier censuses was thought important by the majority of users. The Plans for Scotland's Census 2021 proposed to continue using a question which asks about 'belonging to' a religion, religious denomination or body, on a voluntary basis. Development and testing have been undertaken to ensure, where there have been requests for additional response options, these produce good quality data which meets user needs. As national identity, religion, and ethnic group are different but related facets of how people think of themselves, any testing is conducted across all the relevant questions to understand how people understand and respond to the questions.

Two changes are being proposed to the religion question in the 2021 Census:

Further information on the Muslim population

User need has been expressed for more detailed information on branches or schools of Islam to be captured through the census religion question, in a similar way to how data on Christian denominations is captured. There are no alternative sources of this information. A prompt for Muslim respondents to write in, leaving how to identify open to respondents, was included for qualitative survey testing in Scotland and the key results from the 2019 qualitative survey test are described below. More details on this can be found in the Topic report.^[25]

• The alternative religion question was found to be slightly more acceptable than the 2011 religion question.
• The majority of respondents wrote in and further specified their religious group in the alternative question, gathering data to meet user need.
• Comparability over time is expected at category level ("Muslim") and would allow for harmonisation with census statistics across the UK.
• Some respondents indicated that they would prefer a tick box for their denomination. User need for this information is not stronger than the user need for other census questions or religion response options, and this test has shown that user need is met by having a write in prompt. Two additional tick boxes would be required in the question to meet this data need.

A write in box to collect additional information on denomination is being proposed for 2021.

Information on Pagans

User need has been expressed for a Pagan tick box to be added to the religion question for the 2021 Census. A tick box for "Pagan" is included in the religion question in the Scottish Surveys Core Questions set. The core question is asked in all Scottish Government cross-sectional surveys, including the three largest surveys in Scotland: the Scottish Household Survey (SHS); the Scottish Health Survey (SHeS); and the Scottish Crime and Justice Survey (SCJS). These questions are recommended for use in other surveys to improve comparability, reduce costs of designing surveys, to ensure participants understand what is being asked by using thoroughly tested questions, ensure outputs can be grouped in ways that are useful for analysis and reduce risk of offence when asking about sensitive subjects because the questions have been widely consulted on.

The SHS provide estimates at council area level for the population by religion of belonging. However, due to the small population size and therefore sample size of this group along with many other religious groups, data is unreliable and may not be available for analysis against other variables. It is common for data on Buddhist, Sikh, Jewish, Hindu, Pagan and other religious groups to be combined for analysis. As a result, census is the best source of information for the Pagan population.

A tick box for "Pagan" has previously been tested for the census in Scotland. A tick box was included in the 2006 Census Test and 2009 Census Rehearsal. There were no issues with acceptability, clarity, data quality or comparability concerns with the inclusion of a tick box for "Pagan" highlighted through these tests. A response option was tested for "Pagan" because this was the most frequently written in religion response under the Other category in Scotland in 2001. A tick box was not included in the 2011 Census. As in 2001, Pagans used the write in box to specify their religion.

A tick box for "Pagan" was included in the alternative religion question included in the 2019 qualitative survey. There are no acceptability, data quality or comparability concerns about the inclusion of this tick box for the Pagan population. As such, this testing did not include a quota of Pagan participants. Testing showed that there were no acceptability, data quality or comparability concerns with the inclusion of a tick box for "Pagan" for other respondents.

Data Collection and Enumeration

National Records of Scotland are developing a specific approach to the enumeration of religious establishments which will take account of necessary protocols, stakeholder feedback and the needs of residents, managers and staff.

Field Operations and Recruitment

Cultural sensitivities and protocols will be taken into account in field and contact centre staff training.

Communications, Stakeholder Engagement, Publicity and Marketing

Terminology and language used in questions and materials, and its acceptability to different cultures, faiths and beliefs, carries a risk of being a barrier to participation or response to specific questions. Such considerations will therefore influence design decisions.

Engagement is helping us to better understand any barriers to participation in the census and informs our designs around messaging, publicity and marketing, community engagement and support materials.

Pregnancy and Maternity

Question Development and Statistical Outputs

We have never asked a specific question around pregnancy or maternity in the census and have no plans to do so in 2021.

Data Collection and Enumeration

Our experience of previous censuses has shown that there can be sensitivities stemming from confusion or misunderstanding about whether to include babies on questionnaires if recently born or pre-natal. Cases of still-born or infant deaths also require careful and sensitive consideration. We will take full account of issues around registration, parental preferences and legal considerations in the design and provision of advice and guidance to respondents and the training of field and contact centre staff.

Question development work includes a consideration around whether to alter the term 'maternity leave' to 'maternity or paternity leave' in the labour market questions to provide more information to our users to understand the impact of policies.

Field Operations and Recruitment

Employment practices will adhere to all relevant employment legislation and applicable statutory requirements around pay and leave provisions in the case of pregnancy and maternity.

Communications, Stakeholder Engagement, Publicity and Marketing

Parents of young children are recognised as a group who can experience 'disconnection' with some aspects of everyday life due to the chaotic pressures and time constraints which parenthood often brings. For those people the effectiveness of 'mainstream' broadcast publicity and marketing campaigns can be limited and so the programme will work with stakeholders and contractors to identify the best routes to raise awareness and encourage participation, including tailored publicity and the use of social media and other channels.

Marriage and Civil Partnership

Question Development and Statistical Outputs

Marital or civil partnership status information is used by a wide range of users, including central government and other public authorities in equality impact assessments, which in turn inform policies and practices.

As a result of the Marriage and Civil Partnership (Scotland) Act 2014, the question has been reviewed. Consultation work identified a low user need for information on same sex marriage, and alternative data sources are available from official marriage statistics. The Census marital and civil partnership status question collects data on legal marital and civil partnership status. As there is no difference in the legal status of same sex and opposite sex marriage, it is proposed the question will not be changed to separately identify same sex marriage.