Policy paper

Quality Improvement Plan: government ethnicity data

Published 17 April 2020

Introduction

The government announced the Race Disparity Audit (RDA) in August 2016. The aim was to show how people of different ethnicities are treated across public services by publishing data held by the government. Following the announcement, the Race Disparity Unit (RDU) was set up to lead the audit. RDU published an overview of the main findings in October 2017.

RDU publishes data from across government on Ethnicity facts and figures, and produces transparent statistical outputs for different ethnic groups and topics. The content represents a unique way of describing the experiences and outcomes of people of different ethnic groups in society.

The Women and Equalities Select Committee’s report on the RDA recommended that RDU should produce an action plan “to improve the consistency and robustness of the data it collects on the basis of ethnicity”.

RDU has prepared this plan in response to that recommendation. It provides a detailed overview of RDU’s plans to address issues related to the quality of ethnicity data.

This is intended to be a 2-year work plan starting from publication in April 2020.

Summary of RDU’s quality improvement objectives

RDU will:

• work with other government departments to improve the consistency of the way ethnicity data is recorded and reported

• improve the robustness of ethnicity data, particularly for smaller ethnic groups and in relation to particular datasets and topics

• improve the granularity of data we publish, reducing the number of measures that use the binary (White/Other) classification - this is with a view to moving to a new Census 2021 classification but moving to the Census 2011 5+1 classification as a minimum (we understand that making this change may be difficult in the short to medium term)

• work with government departments to provide data to address some important gaps in the evidence base

• improve the availability of sub-national data, especially for local authorities

• provide leadership and coordination of cross-government work on the quality of ethnicity data, and provide support and help to analysts in other government departments who are collecting, analysing and disseminating ethnicity data

We intend to meet these objectives over the 2 years from April 2020.

RDU’s approach to quality

Our objectives broadly derive from the definition of quality in the Code of Practice for Statistics:

• “Quality means that statistics fit their intended uses, are based on appropriate data and methods, and are not materially misleading.”

• “Quality requires skilled professional judgement about collecting, preparing, analysing and publishing statistics and data in ways that meet the needs of people who want to use the statistics.”

They are also consistent with the emphasis on value in the Code of Practice:

• “Value means that the statistics and data are useful, easy to access, remain relevant, and support understanding of important issues.”

• “Value includes improving existing statistics and creating new ones through discussion and collaboration with stakeholders, and being responsible and efficient in the collection, sharing and use of statistical information.”

The objectives will continue the direction for the trustworthiness of ethnicity data as defined in the Code of Practice:

• “Trustworthiness is a product of the people, systems and processes within organisations that enable and support the production of statistics and data.”

• “Trustworthiness comes from the organisation that produces statistics and data being well led, well managed and open, and the people who work there being impartial and skilled in what they do.”

The objectives also link to the European Statistical System’s Dimensions of Quality.

The Code of Practice for Official Statistics requires official statistics producers to inform users about the quality of statistical outputs using these dimensions of quality. These dimensions aim to give a broad understanding of quality, as represented by how fit for purpose a statistical output is (summary RDU objectives in brackets):

Relevance

The degree to which the statistical product meets user needs in both coverage and content (granularity, sub-national data, data gaps).

Accuracy and reliability

Accuracy is the proximity between an estimate and the unknown true value. Reliability is the closeness of early estimates to subsequent estimated values (robustness).

Comparability and coherence

Comparability is the degree to which data can be compared over time and domain. Coherence is the degree to which data that is derived from different sources or methods, but refers to the same topic, is similar (consistency).

Accessibility and clarity

Accessibility is the ease with which users are able to access the data, also reflecting the format in which the data are available and the availability of supporting information. Clarity refers to the quality and sufficiency of the metadata, illustrations and accompanying advice.

Timeliness and punctuality

Timeliness refers to the time gap between publication and the reference period. Punctuality refers to the gap between planned and actual publication dates.

While not as relevant for this plan, accessibility and clarity are important as we make more data available for different geographies (and also for gender and disability) by ethnicity. Timeliness is an important quality measure for the RDU’s work on publishing ethnicity data from other departments on Ethnicity facts and figures.

1. Consistency

Objective: Work with other government departments to improve the consistency of how ethnicity data is recorded and reported.

RDU will work with Government Statistical Service (GSS) Harmonisation Champions to create a harmonised set of ethnicity classifications. This will involve:

• cross-GSS work to develop new harmonised standards for household and individual surveys to reflect the 2021 Censuses with the aim of moving as many survey and administrative measures to these new classifications - once the Census Order is approved, we will encourage those responsible for administrative systems that record ethnicity to set out their commitment to using the 2021 Census classification (we will report our progress on this)

• if necessary in the interim, and as a minimum, working with departments and GSS Harmonisation Champions to make sure as many measures as possible present the Census 2011 5+1 classification (we will report our progress on this)

• continuing to add information classified to the harmonised Census 2011 5+1 categorisation to the downloadable data files that form part of each measure

• providing advice to departments collecting data for the different countries of the UK that use different harmonised categories

• systematically engaging with owners of administrative systems to explore the practicalities of making changes to the way ethnicity is recorded

• focusing in particular on the harmonisation priorities in Appendix 1, offering advice for departments on how they deal with changes to classifications over time, to maintain time series, or reduce the impact of discontinuities in the data

• working with departments to maintain a harmonised approach to collecting data about Gypsy, Roma and Traveller people using the classifications proposed for the 2021 Census

• publishing a methods and quality report about the treatment of data which includes people whose ethnicity is not known, and using this to encourage departments to present the level of ‘unknown ethnicity’ in datasets and explore reasons for low declaration rates - this will allow users to be more informed about the quality of statistics and the conclusions that can be drawn from the data, and help departments to increase the declaration rate

• working with the ONS Equalities Data Audit team to produce a comprehensive set of information about datasets that record ethnicity

Harmonisation and consistency of data

‘Harmonisation’ is about making statistics more comparable, consistent and coherent. Read more about what harmonisation means.

The topic of consistency of ethnicity data includes a number of factors. One of the most important is the harmonisation of categories used in the collection and presentation of ethnicity data. This also includes providing relevant information on the number of people who don’t declare their ethnicity in a survey or administrative process. This might be different from response rates (read ‘robustness’ below) as the person may otherwise be willing to fill out the survey or complete the administrative process.

RDU’s preferred approach to the collection and presentation of ethnicity data is described in the harmonised principles for ethnicity. These were used in the Census 2011 (Appendix 2).

RDU will work with ONS and others to produce a harmonised set for the Census 2021. However, RDU is mindful of the challenge of balancing the need for comparability over time, the data needs of users, and trying to make as much data available as possible.

Levels of harmonisation on Ethnicity facts and figures

Ethnicity data published on Ethnicity facts and figures is taken from many different sources, including social and economic surveys. Some datasets are derived from administrative systems. On Ethnicity facts and figures in March 2020:

• measures currently use 21 different classifications of ethnicity
• the Census 2011 18+1 classification is used on 31 pages
• the detailed classification (or a small variant of it) used in the 2001 Census (the 2001 16+1) is used on 28 pages
• the harmonised summary approach used by ONS for the Census 2011 and applying to each country of the UK (the 2011 5+1) is used on 72 pages
• an alternate classification - these might be other 5-fold, 4-fold classification, or binary classification - is used on 91 pages

A binary split (White/Other than White, or White British/Other than White British) was used on 33 pages. Where this classification has been used, it often reflects the need to combine minority groups because sample sizes are too small to present robust data for these groups separately.

These alternate classifications limit the extent to which data can be compared and the binary classifications have very little analytical value. Indeed, for many purposes and most users even 2011 5+1 is not sufficiently granular. For example, the experiences and outcomes of people of Pakistani, Bangladeshi, Indian or Chinese origin may well all be markedly different. However, they are commonly aggregated into a single ‘Asian’ group.

These figures relate to pages on the website. A page can use one or more classifications.

Coverage across the UK

The geographical coverage of most of the datasets is England, but some data covers 2 or more UK countries, and some covers the whole of the UK. This presents a further complication in that there are 3 Censuses of population:

• England and Wales
• Scotland
• Northern Ireland

Each Census is governed by separate legislative arrangements, and collects data about ethnicity in different ways, and using different classifications. As a consequence, there are different harmonised (output) classifications for England, Wales, Scotland, and Northern Ireland.

Unknown ethnicity

Ethnicity is usually self-reported in the UK. Sometimes people don’t report their ethnicity in surveys or administrative processes. For example, they may choose to ‘prefer not to say’ or simply not give any answer at all. They may do either of these things even if they are otherwise completing the survey or administrative process. Sometimes ‘prefer not to say’ is included as a valid response and included in the declaration rate, even though it has no analytical value.

Low declaration rates can be more obvious in workforce statistics or in administrative data sources, such as Universal Credit statistics where it’s not mandatory for ethnicity data to be collected.

2. Robustness

Objective: Improve the robustness of ethnicity data, especially for smaller ethnic groups and in relation to specific datasets and topics.

RDU will:

• analyse the 2021 Census results when released, the best source of robust data for ethnic minority groups, especially groups with small populations, such as Gypsy, Roma and Travellers

• show how linking data sources, for example the Census and benefits data, could provide statistics that are higher quality, especially for ethnic groups with smaller populations - this will be through best practice examples or a demonstration project where a ‘small’ ethnic group has been analysed in more detail

• provide advice on how the robustness and coverage of data about Gypsy, Roma and Travellers, and other ethnic groups with small populations could be improved - this could be by combining data for more than one time period, and by exploring ‘ethnic minority boosts’ of existing surveys, for example through the new Integrated Population and Characteristics Survey (IPACS) currently being developed by ONS

• work with departments to explore what steps might be taken to reduce levels of non-reporting

• publish one or more methods and quality reports that show the implications for presenting and interpreting data when the geographical distribution of a disparity is particularly uneven, for example relative stop and search rates

• publish a methods and quality report on how ethnicity data that is provided by a third party compares with data provided by the individual themselves

Ways to increase robustness

The release of the 2021 Census results will give access to the best source of data for robust analysis of ethnic minority groups. However, there are a number of other ways to increase the robustness of these estimates. Some social surveys have complex sampling designs, and include ‘boost’ samples that are designed to increase coverage of selected minority categories. Usually these boosts would be within a specific location. In these cases, researchers should take into account the complexity of the survey design when they produce summary statistics from surveys that “boosted” their data sample. There might be a risk of the inaccurate estimation of statistical measures, or the drawing of wrong conclusions based on those measures, especially when there is clustering of the ethnic minority groups.

Similarly, it might be possible in some surveys and administrative sources to aggregate data over a number of time periods. This might allow estimates for smaller ethnic groups to be produced, and more generally potentially increase the reliability of estimates for all ethnic groups. One trade-off in this instance is that a time series of data might not be possible. Also, changes in the categorisation of ethnic groups over time might mean that it is not always possible to group years together in a consistent way over time.

Finally, linking data between administrative sources, or between surveys and administrative sources, could provide more robust statistics or offer the facility to provide more in-depth analysis for certain ethnic groups or topics.

Response rates

Low response rates will also affect the robustness of the data because the profile of respondents can differ from that of non-respondents. There is considerable evidence to suggest that response rates are lower among certain groups. Studies have shown repeatedly that people from Black and minority ethnic groups are less likely to respond to postal surveys than people from some other groups. Response rates are also generally lower among people living in areas with more residents who were from ethnic minorities.

Gypsy/Roma/Traveller

There are some cases where small ethnic groups such as the Gypsy, Roma and Traveller groups are not captured in surveys, or where small sample sizes cannot support robust results.

The April 2019 House of Commons Women and Equalities Select Committee report on tackling inequalities faced by Gypsy, Roma and Traveller communities noted that there was a scarcity of data on these groups. This created an obstacle to developing targeted evidence-based policy to reduce inequalities compared with other groups. In many data collections, the option for people to identify as Gypsy, Roma or Traveller is not available. Where it is available, people may be reluctant to self-identify as they mistrust the intent behind the data collection and feel it might be used to discriminate against them.

3. Granularity

Objective: Increase the granularity of data we publish, reducing the number of measures that use the binary (White/Other) classification, with a view to moving to new Census 2021 harmonised outputs or (as a minimum and as an interim measure) the Census 2011 5+1 classification.

RDU will:

• work with departments to provide more detailed data on ethnicities that provide alternatives to the binary classification and discuss these with users - this will include investigating a harmonised output standard for the UK that will be more granular than in 2011, for example consisting of 8 to 12 groups - the 2021 Census will provide an opportunity to release the most up-to-date information for different ethnic groups

• provide advice to departments on moving towards 2021 Census harmonised classifications for input and output and away from binary classifications in particular - where this is because of small sample numbers, RDU will provide advice and guidance on how to provide more robust samples (read the Robustness section for more)

• as RDU publishes new and updated measures, it will explore the potential to increase the granularity of the data by adding factors such as gender, age, region, and socio-economic status, recognising that many of the existing measures include at least some such analysis - the Census 2021 results will provide the best source of data for this when they are released, however, these new analytical resources might be produced by linking existing datasets, or by further analysis of existing individual datasets

• explore with users the extent to which they are interested in analysis by religion

Binary classifications

33 pages on Ethnicity facts and figures have small sample sizes which don’t allow the presentation of ethnicity classification into the 2011 5+1. These measures are listed in Appendix 3, and they use a binary classification for at least one disaggregation - there may be other disaggregations for the measure that use more detailed classifications.

Binary classifications are often used in household surveys, or where data is being disaggregated further by geography or gender. As a result, the data is sometimes presented as a binary classification (for example White and Other than White). This classification has little analytical value as it doesn’t give a comprehensive picture of the issues that different ethnic minority groups can face. RDU would prefer a move towards the 2011 5+1 classification as a minimum.

In the ‘Consistency’ section, we highlight how RDU will help departments move towards using harmonised ethnicity classifications for data collection and reporting.

In the ‘Robustness’ section, we explain that sample data boosts, aggregating data over more than one time period (usually aggregating years together), and linking datasets could help reduce the number of datasets using the binary classification.

4. Data gaps

Objective: Work with government departments to provide data to address important gaps in the evidence base.

RDU will:

• work with data providers to fill the data gaps identified in Appendix 4 and report on progress towards filling those gaps

• monitor new data collections to make sure they include ethnicity as a variable

• make use of user research to inform future data gaps, either where datasets exist but no ethnicity variable is collected, or where it’s important to improve the evidence base on ethnicity for a certain topic but no data or research exists

• work with departments to provide best practice on linking datasets to improve the amount, or quality, or ethnicity data

Prioritising the gaps

RDU has over 170 measures that analyse and display ethnicity data. RDU has also been working on further disaggregation of statistics by providing data (where available) by:

• ethnicity and sex
• geography
• socioeconomic status
• disability

RDU has established the list of data gaps using a number of criteria, including:

• the current and future policy needs
• the quality of the data
• whether the ethnicity breakdown is harmonised
• public interest in the data

5. Subnational data

Objective: Improve the availability of regional data, in particular for local authorities.

RDU will:

• increase the geographical granularity of the data by working with local authorities, public service delivery organisations, and third sector partners to explore the potential of data that hasn’t yet been published on Ethnicity facts and figures, for example, the ‘quality of life’ data from Bristol - the 2021 Census will provide the best source of robust ethnicity data on certain topics for small geographic areas

• influence more local authorities to use the same, harmonised classifications - the baseline for this will be an audit of classifications used by local authorities

• analyse the existing information on local authorities from surveys or administrative data sources and the 2021 Census, presenting the information in maps to improve the accessibility of the outputs

• explore the use of linking datasets, for example the Census with administrative data from the Department for Work and Pensions, to provide data and analysis at the local authority level

Data for local authorities

RDU has started to provide information for local authorities where departments have supplied this data and where it allows for this disaggregation. RDU has already established a network with cities such as Bristol to discuss data collection on frontline services and social mobility.

6. Leadership and coordination

Objective: Provide leadership and coordination for cross-government work on the quality of ethnicity data, and provide support and help to analysts in other government departments who are collecting, analysing and disseminating ethnicity data.

RDU will:

• lead on providing quality guidelines for ethnicity data - this could include providing or overseeing training on quality for data providers and users, providing seminars on the content design of ethnicity statistics, helping departments with reporting on ethnicity statistics, including advising on the language used, and helping authorities interpret changes over time when classifications have changed

• publish a regular series of methods and quality reports about methodological issues relevant to users’ interpretations of our data, and to showcase RDU work to improve the quality of ethnicity data

• develop and implement a quality management system across RDU, and publish internal documentation for the quality assurance of ethnicity data

• help analysts to publish data on Ethnicity facts and figures by providing seminars on RDU’s publishing tool

• work with Reproducible Analytical Pipeline (RAP) experts in other government departments to find out if best practice can be shared with other analysts - RAP could be used within RDU to improve efficiency and reduce the scope for errors, or between RDU and data suppliers to improve the overall process of data collection and validation

• have an RDU representative on the cross-government GSS RAP champions network

• give opportunities to statisticians to influence the design and development of administrative data systems from which statistics about ethnicity are produced

RDU support

The main contacts will be the Deputy Director for Data and Analysis and Head of Quality and Standards.

Transparency of outputs

RDU uses data from many different government departments for Ethnicity facts and figures, and information is taken from both surveys and administrative systems. RDU is transparent on the quality of the statistics used on Ethnicity facts and figures.

Quality assurance

In order for RDU to publish data on Ethnicity facts and figures, the quality assurance process will involve establishing and monitoring against targets elements, such as response rates for surveys, and checking figures that fail internal validation checks (such as outliers) with departments.

Quality control and reviews

The Code of Practice for Statistics states:

“Scheduled revisions or unscheduled corrections to the statistics and data should be released as soon as practicable. The changes should be handled transparently in line with a published policy.”

It also states that:

“Scheduled revisions, or unscheduled corrections that result from errors, should be explained alongside the statistics, being clear on the scale, nature, cause and impact. (Q3.4)”

RDU is committed to the principle of transparency and making sure that the information we present is consistent with published data from other departments. Our correction policy states:

• if RDU makes a mistake in handling data provided to us by a department we will correct it as soon as we can
• if a department corrects an error in its published statistics we will follow the department’s practice in making a correction
• we will publicise corrections transparently and openly - this is both on the measure page in question, and in a separate ‘corrections log’

Reproducible Analytical Pipelines (RAP) and other technologies

RDU will support the use of RAP and will establish whenever they can be applied within the work of the unit by training and upskilling its internal analysts. We will also have an attendee from RDU on the GSS RAP Champions Network.

Appendix 1: Harmonisation priorities

Department for Education (DfE)

The ethnicity codeset for DfE reflects categories used in the 2001 Census of England and Wales with some additional categories, for example Traveller of Irish heritage.

DfE uses Chinese as a separate category from Asian. This inconsistency prevents researchers from doing cross-cutting analysis, such as on education and employment. RDU is discussing changes to the codeset based on the 2021 Census ethnic categories with the aim of addressing the inconsistency with the Chinese ethnic group, and deriving the Arab category.

NHS Digital/Department for Health and Social Care

NHS organisations are mandated to use ethnic monitoring questions and response codes based on the Census. The NHS currently uses the ethnic category codes from the 2001 Census.

Additional codes can be included as appropriate at a local level to reflect the demographic make-up of the local population. This allows local monitoring to take place in a way that supports service planning, decision-making, and processes such as the Joint Strategic Needs Assessments. It also allows national comparisons to be made.

RDU is working with NHS Digital and is supporting its progress to harmonise the ethnicity data collection with Census 2021.

Ministry of Justice (MoJ) and the Home Office (HO)

There are 2 measures of recording ethnicity used in MoJ and HO statistics:

Officer identified ethnicity

This is as recorded by a police officer or a member of the administrative or clerical team and is based on visual appearance. The Police National Computer (PNC) ethnicity categories are then aggregated to 4+1 classifications:

• White
• Black
• Asian
• Other
• Unknown

This is the ethnicity information presented for information from the PNC and includes data on:

• cautions
• first-time entrants
• reoffending
• criminal histories

Self-identified ethnicity

This is defined by the individual. Categories are based on the 16+1 ONS classifications from 2001 and aggregated into the 5+1 classification as follows:

• White
• Black
• Asian
• Mixed
• Other
• Unknown
• not recorded

This is the ethnicity presented for information from the Home Office (data on arrests and stop and search), the Youth Justice Board (characteristics of children according to Youth Justice Application Framework and outcomes following remand), and the Youth Custody Service (children in custody in the secure estate and behaviour management).

The ONS introduced 2 further categories to the Census in 2011:

• White – Gypsy or Irish Traveller
• Arab

It moved Chinese to the Asian category.

To allow for comparability within the time series, Chinese is placed alongside the Other ethnicity category following the 2001 Census.

The Home Office has requested data from forces in the 2011 Census format on a voluntary basis from April 2020, with the view of making reporting in this mandatory from April 2021. We will continue to work with the Home Office on moving to the 2021 Census classifications in the future.

Appendix 2: Census 2011 harmonised ethnicity classification

The recommended Census 2011 harmonised lists for questions used in England, Northern Ireland, Wales and Scotland.

England

White

• English/Welsh/Scottish/Northern Irish/British
• Irish
• Gypsy or Irish Traveller
• Any other White background, please describe

Mixed/Multiple ethnic groups

• White and Black Caribbean
• White and Black African
• White and Asian
• Any other Mixed/Multiple ethnic background, please describe

Asian/Asian British

• Indian
• Pakistani
• Bangladeshi
• Chinese
• Any other Asian background, please describe

Black/ African/Caribbean/Black British

• African
• Caribbean
• Any other Black/African/Caribbean background, please describe

Other ethnic group

• Arab
• Any other ethnic group, please describe

Wales

White

• Welsh/English/Scottish/Northern Irish/British
• Irish
• Gypsy or Irish Traveller
• Any other White background, please describe

Mixed/Multiple ethnic groups

• White and Black Caribbean
• White and Black African
• White and Asian
• Any other Mixed/Multiple ethnic background, please describe

Asian/Asian British

• Indian
• Pakistani
• Bangladeshi
• Chinese
• Any other Asian background, please describe

Black/African/Caribbean/Black British

• African
• Caribbean
• Any other Black/African/Caribbean background, please describe

Other ethnic group

• Arab
• Any other ethnic group, please describe

Scotland

White

• Scottish
• Other British
• Irish
• Gypsy/Traveller
• Polish
• Any other White ethnic group, please describe

Mixed or Multiple ethnic groups

• Any Mixed or Multiple ethnic groups, please describe

Asian, Asian Scottish or Asian British

• Pakistani, Pakistani Scottish or Pakistani British
• Indian, Indian Scottish or Indian British
• Bangladeshi, Bangladeshi Scottish or Bangladeshi British
• Chinese, Chinese Scottish or Chinese British
• Any other Asian, please describe

African

• African, African Scottish or African British
• Any other African, please describe

Caribbean or Black

• Caribbean, Caribbean Scottish or Caribbean British
• Black, Black Scottish or Black British
• Any other Caribbean or Black, please describe

Other ethnic group

• Arab, Arab Scottish or Arab British
• Any other ethnic group, please describe

Northern Ireland

• White
• Irish Traveller

Mixed/Multiple ethnic groups

• White and Black Caribbean
• White and Black African
• White and Asian
• Any other Mixed/Multiple ethnic background, please describe

Asian/Asian British

• Indian
• Pakistani
• Bangladeshi
• Chinese
• Any other Asian background, please describe

Black/African/Caribbean/Black British

• African
• Caribbean
• Any other Black/African/Caribbean background, please describe

Other ethnic group

• Arab
• Any other ethnic group, please describe

Appendix 3: Measures using binary classifications

These are measures that use a binary classification for at least one disaggregation - there may be other disaggregations associated with the measure that use more detailed classifications.

• Young people in custody
• Use of force on young people in custody
• Single separation incidents for young people in custody
• Restrictive physical interventions involving young people in custody
• Assaults carried out by young people in custody
• Visits to the natural environment
• Honours recipients
• Consent for organ donation
• New social housing lettings
• Fuel poverty
• Fuel poverty gap
• Regional ethnic diversity
• Employment
• Unemployment
• Length of time spent in unemployment
• Economic inactivity
• Jobseeker’s Allowance
• Employment and Support Allowance: sanctions
• Armed forces workforce
• NHS workforce appointments from shortlisting
• NHS trust board membership
• NHS staff experiencing discrimination at work
• NHS staff believing career progression is fair at work for all staff
• Leadership of small and medium enterprises
• Time spent living in current home
• Home ownership
• Age of first-time buyers
• Deposits paid by first-time buyers
• Renting from a private landlord
• Renting from a local authority or housing association (‘social housing’)
• Overcrowded households
• Households under-occupying their home

Appendix 4: Data gaps

Crime, justice and the law

• Data for other types of crime (for example, knife crime)
• Young offenders background by ethnicity
• Access to legal aid

Health

• Cause of death
• Infant mortality
• Life expectancy
• Health inequality
• Low weight in babies during pregnancy
• Low birth weight and infant mortality
• Children and young people’s mental health

Work, pay and benefits

• Total wealth
• Sources of wealth
• Wealth distribution
• Debt
• Household expenditure
• Income Tax
• Access to childcare
• Employment by household type
• Universal Credit

Culture and community

• Loneliness
• Digital inclusion

Housing

• Mortgage arrears
• Repossessions
• Access to mortgages
• Smoke alarms and fires at home
• Social housing waiting list
• Social rent prices
• Right to Buy and Right to Acquire
• Rent and rent arrears
• Private rent prices
• Evictions
• Property ownership
• Protection from homelessness

Workforce and business

• Civil Service People Survey
• Armed forces workforce by age and gender
• Company directors and board members
• Prison population by prison category

Education, skills and training

• (Schools) Cross-border movement between local authorities (living vs studying)
• Highest study aim (for example, check the Annex of this DfE report)
• University league tables
• Graduate outcomes for postgraduates
• Behaviour/discipline
• Teacher qualifications/teachers professional development
• Teaching of British values
• Mental health training for teachers
• Support for children in need
• School funding
• Protection of pupil premium
• Participation in the arts
• Any other additional cross-tabulations of different measures (for example, absences by attainment as per the Timpson review)