Annex B: Our research evidence

The research for this review was based on conversations with a range of stakeholders and a review of published literature relevant to statistics and data about children and young people. Where relevant, we have drawn on our other recently published work and highlighted where issues have been raised through our wider regulatory activities. The examples highlighted in this research annex are not exhaustive and have been used to illustrate the points raised with us and the good practice that does exist.

B.1 Visibility lens

B1.1 Availability of statistics on children and young people

We found that children and young people are not as visible in official statistics as they should be. Importantly, many national statistical surveys do not include young people under the age of 16. In addition, the experiences of young people in the transition to adulthood are often not analysed separately due to small sample sizes.

During our engagement activities we heard that, whilst there are statistics around children and young people on most topics across the statistics system, these are more limited outside of topics specific to children, such as education. For other topics data on adults is often taken as a proxy for data on the whole population. For example, The Children’s Society told us that ‘Currently, data collection involving children and young people is fragmented, and their experiences of important issues (e.g. children’s well-being) are often not routinely and systematically measured in the way they are for adults’.

In some surveys children and young people are simply not included. Users of statistics told us they are concerned that this may be having an impact on policy, as the experiences of children are likely to differ from those of adults. They were concerned that the current lack of data for children in some national surveys means that policy interventions and funding efforts are being based on the experiences of adults rather than children. Therefore important opportunities to improve the lives of children and young people may be being missed.

The UK Statistics Authority Inclusive Data Taskforce has recommended that data producers should review the representativeness of key surveys and administrative datasets (initially by benchmarking against the UK 2021/22 Censuses) and take swift action to address identified issues, particularly as they relate to historically under-represented populations (for example young black men) or more marginalised groups (for example children).

A particular area of concern we heard from users of statistics is the lack of consistent data on spending on services aimed at children and young people across government. For example, it is not possible to estimate total government expenditure on children across the UK.

We heard that this can make it difficult to assess the effectiveness of funding on children, or the impact of changes to budgets on children’s services. This issue is addressed in the Report of the Children’s Commissioners of the United Kingdom of Great Britain and Northern Ireland to the United Nations Committee on the Rights of the Child, which found that ‘Across the UK, there is a lack of transparency and clarity in budget processes, allocations, and expenditure concerning children.’

The definition of a child or young person can vary between topics. The UN Convention on the Rights of the Child defines a child as everyone under the age of 18, in line with other legal frameworks. However, for children leaving care there remain responsibilities for the state until the age of 25.

Users reported to us that understanding outcomes for young people aged 18 and beyond is important, as these are needed to understand how to improve the lives of the current and future cohorts of children and to inform an understanding of the economic and societal impacts of childhood experiences.

Analyses relating to young people over the age of 18 are important in helping children and young people make decisions affecting their lives, such as on university courses, or in giving insight into outcomes such as labour market participation. The Key Data on Young People 2021 report, published by the Association for Young People’s Health, brings together a range of data sources relating to young people’s health and lives across the UK. It highlights the need for data on 10-25 year olds to understand transitions to adulthood, but that the data are frequently combined with other age groups, rather than being available separately.

The Office for National Statistics (ONS) Crime Survey for England and Wales was praised by some we spoke to for the direct surveying of children aged 10-15 years. However, as the experiences of young people are likely to differ from older adults, we heard concerns that data for young people aged 16 and 17 are included with adults, rather than analysed separately.

Achievable sample sizes however can limit the ability of producers to include children and young people in some surveys. As an example, a recent study exploring the feasibility of surveying 10-15 year olds in the Northern Ireland Community Safety Survey found that the likely sample size that would be achieved through responding households would not provide robust estimates. Additional questions were instead included in the Young Persons’ Behaviour and Attitudes Survey.

B1.2 Gaps in the available statistics

Children and young people are visible across a range of topics in official statistics, but there are some data gaps.

In this review, and through our wider regulatory work, we have identified several gaps in official statistics relating to children and young people. These include, but are not limited to, gaps in the topic areas detailed below. Though not an exhaustive list, we have drawn on examples to illustrate some of the types of gaps we found.

Mental health statistics: As we noted in our Review of Mental Health Statistics in Northern Ireland, the Northern Ireland Commissioner for Children and Young People concluded in their 2018 ‘Still Waiting report’ that there were ‘alarming gaps’ in operational data to efficiently plan, commission and deliver mental health services for children and young people in Northern Ireland. The Youth Wellbeing Prevalence Survey 2020 has since provided the first prevalence estimates of common mental health problems in children and young people in Northern Ireland.

Well-being statistics: An important source of data on children’s subjective well-being is the annual Children’s Society’s Good Childhood report. Historically there has been a gap in evidence around children’s well-being in official statistics. In the last few years there have been important improvements in these statistics.

These include international data on well-being from the Programme for International Student Assessment, the latest data from the Health Behaviours in School Age Children survey, and the Children’s Worlds international survey of Children’s well-being. In England, the Department for Education has published a State of the Nation 2021: children and young people’s well-being report which presents data from a range of sources. In 2019 Welsh Government published a report ‘Well-being of Wales: What do we know about children’s well-being’. The Active Lives Children and Young People survey in England, which was established in 2017, and Youth Wellbeing Prevalence Survey in Northern Ireland in 2020 included questions on well-being.  Although there have been positive developments, there remain difficulties in gaining a wider UK perspective on children’s well-being.

Homelessness statistics: There are limitations in the data that are available on homeless children. Gaps in data on homeless families were highlighted in the Children’s Commissioner for England’s Bleak Houses report. Users who we spoke to as part of our National Statistics assessment of Statistics on Statutory Homelessness in England highlighted the lack of data on some types of homeless children. Estimates of the number of statutory homeless young people and households with dependent children are published by the Department for Levelling up, Housing and Communities. These numbers do not include young people or children in households who are ‘sofa surfing’ or have not asked the council for help. Nor do they include children who are placed in temporary accommodation by children services rather than a council’s housing department. Data are available on the accommodation of care leavers in the Department for Education Children Looked After in England statistics.

Data from the Scottish Government covers the number of households with children, and the number of children who are homeless as well as households where the main applicant is aged 16-17 and 18-24. Homelessness statistics in Wales include estimates of the number of homeless households with dependent children in temporary accommodation and the number of young people aged 16-17 and care leavers aged 18-21 who are placed in temporary bed and breakfast accommodation. Data from the Scottish Government covers the number of households with children and the number of children who are homeless as well as households where the main applicant is aged 16-17 and 18-24. The Northern Ireland Homelessness bulletin also provides the number of children from households accepted as statutorily homeless by age.

Home educated children: The House of Commons Education Committee report, Strengthening Home Education, highlights a lack of data on the number of children who are electively home educated and the quality of that education. The report outlines the need for reliable data on the number of children educated out of school, their educational attainment and outcomes, and that greater collection and analysis of data is needed on issues such as why parents choose to home educate. In Wales statistics are published on the number of Pupils Educated Other than at School who are funded by a local authority. The release includes statistics on numbers of pupils that local authorities know are electively home educated. However there is no legal requirement for parents to inform the local authority that they have a child of school age.

Bereaved children: There is no official estimate of the number of children who are affected by the death of a parent, sibling or friend. In 2019 the Office for National Statistics analysed longitudinal studies to estimate that around 1% of children have experienced the death of their mother. The Childhood Bereavement Network have used a range of sources to provide further estimates of the extent of childhood bereavement.

Users told us that they have to rely on data from charities or infrequent data collections for some topics, but that they felt it was right that not all statistics on children and young people are provided only by official statistics producers. We heard it is important that civil society produce their own data, re-analyse that provided by official sources and provide challenge by providing alternative data or analyses. However, users also reported to us that there are some topics, such as those listed above, where there is little or no data available from official sources. These topics are of sufficient importance in understanding the lives of children and young people that users believe that official statistics should be available on them.

B1.3 Data on very young children and babies

We found that whilst children and young people are visible across a range of topics in official statistics, there are some data gaps on very young children and babies.

Many of the data collections relating to children and young people relate to children in a school setting. These include data sets such as the Department for Education National Pupil Database in England and collections such as the pupil level school censuses which collect data for every child in a state school. There are a limited number of collections for young children who do not yet attend school.

The main statistics available on very young children and babies include statistics on births, deaths, immunisations, breastfeeding and data relating to contacts with services such as health visitors or referrals to social services. These statistics alone do not give much information on the lives and development of this age group of children. Richer data are available through longitudinal studies which are discussed more in section B1.5.

In England, the community services dataset, published monthly, provides a rich source of information on this age group. These are currently Experimental Statistics. Official Statistics on assessments at age 2 to 2 and a half covering national level results are published on an annual basis. In Scotland, statistics are available in the early child development release on reviews at 13-15 months, 27-30 months and 4-5 years. Outcome indicators such as these are not available in Wales or Northern Ireland.

The UK government’s The best start in life: A vision for the 1,001 critical days highlights the importance of a child’s life up to age 2 in setting the foundations for an individual’s cognitive, emotional and physical development. The report raised concerns that the current data collections do not aid understanding of the variation and disparity of outcomes across the country.

B1.4 Seeing a holistic picture of children

It is difficult to see a holistic picture of children and young people’s lives through official statistics. Joined up approaches are needed to enable studies to collect, or link, data on children and their families for more than one aspect of their lives or over time.

To understand children’s lives, users told us it is important to see children and young people as whole entities. The Association for Young People’s Health Key data on Young People 2019 report states that, ‘The data are frequently compartmentalised into topics such as youth justice, obesity, or mental health, which may present information in different ways or relate to different age breakdowns. Drawing connections between the topics can therefore be challenging, yet we need to view young people holistically. This is the only way to obtain an overview of what young people need to reach their full potential and the services that need to be commissioned.’

Seeing a holistic picture of children and young people’s lives often requires linking of data from different sources to understand, not just the educational attainment of a child, but also their health, any contacts with support services or with the criminal justice system, the family situation, socio-economic environment and local area. Very recently one such data linkage has taken place. The ECHILD project, for the first time, joins together existing health, education and social care information for all children in England. It includes de-identified linked records for around 14 million children. This linked data will be used to better understand how education affects children’s health, and how health affects children’s education. An ADR UK report has been published which highlights the potential of this dataset.

The SAIL databank contains anonymised data on the population of Wales, including children, and enables linkage and analysis of the data. A recent study using data from SAIL explored the relationship between school absence and exclusion and diagnosed neurodevelopmental and mental disorders in children and young people.

Users described how children live in a more ‘multi-level’ environment than adults and that the key influences in their lives come from family, community and school. As such, child level data alone are not sufficient. For example, understanding the economic environment of the child requires data on the income of the parents and their claimant status linked to information on the child. Researchers reported to us that this data was often not available, so proxies at the geographical level (for example, postcode or Lower Layer Super Output Areas) are used, which risks introducing uncertainty and bias into the statistical analyses and reduces their robustness. Specifically in Scotland, researchers expressed concern that area-based measures such as the Scottish Index of Multiple Deprivation 2020 are inadequate as measures of individuals’ or families’ characteristics. Researchers want improved data on income, occupational class and cultural resources within the home to increase the data linkage possibilities to gather more insight on socio-economic characteristics within educational attainment and then through to post-school destinations.

These issues are reflected in the recommendations of the UK Statistics Authority Inclusive Data Taskforce that state that a set of measures of socio-economic background should be developed that can be collected in administrative data sources. It is recommended that these should include parental occupation and parental educational level.

The Nuffield Foundation report, How are the lives of families with young children changing?, highlights evidence gaps around separating families and paternal mental health. The report notes that ‘Many of the gaps in evidence about early childhood and family life reflect the fact that some national, official and longitudinal surveys have not been designed to collect data about certain groups.’

Relatedly, many children live in two separate households, for example due to parental separation. This is a challenge for official statistics, as many surveys are based on the concept of a household and official data collections tend to focus on the ‘main homes’. For children this may mean that the definition of a ‘household’ is getting outdated and their data are not accurately reflected in official statistics. We could not find an official estimate of the number of children in this situation.

A good example of where a more holistic picture of children can be seen is in the Health Survey for England which collects data on both adults and children. This enables analysis of the health of children in relation to the adults in their household. Analyses include obesity and health behaviours of children in relation to those of their parents. This demonstrates the benefits that greater linkage of adult and child information across wider topic areas could bring.

Another example is the Northern Ireland Young Persons’ Behaviour & Attitudes Survey a school – based survey carried out amongst 11-16 year olds, which covers a wide range of topics relevant to the lives of young people in Northern Ireland today.

B1.5 Birth and longitudinal cohort studies

Longitudinal cohort studies, which track children and young people over a period of time, require significant long term financial investment and a sample design that ensures the sample is representative and does not create bias in the survey data. Difficulties in non-government organisations accessing government administrative data to use as sampling frames have delayed or limited some longitudinal cohort studies. There has been a large time gap since the last UK wide birth cohort.

Birth cohort studies follow babies born around the same time throughout their lives through a mixture of direct surveying and linkage of administrative data. Historical birth cohort studies were cited as being helpful in understanding children’s lives and the impact of experiences through to adulthood.

The most recent UK wide birth cohort study is the Millennium Cohort Study which is following the lives of 19,000 young people born in 2000-2002. There have been seven waves of data collection, with the next wave planned for 2022. The study has collected information about cohort members’ physical, socio-emotional, cognitive and behavioural development over time, alongside detailed information on their daily life, behaviour and experiences.

In Scotland, Growing Up in Scotland (GUS) tracks the lives of thousands of children and their families through to the teenage years. Launched in 2005, to date GUS has collected a wide range of information on three nationally representative cohorts of children. An eleventh sweep of data collection is currently underway with the original birth cohort, who are now in their final year of schooling.

Several users expressed concerns about the lack of a recent UK birth cohort study, noting that there was due to be one in 2012, but that it did not take place. The Longitudinal Studies Strategic Review, published in 2017 for the Economic and Social Research Council (ESRC) noted that ‘The UK is recognised globally as having significant strengths in longitudinal data, thanks to historic investments by the ESRC and other funders over many decades. The investment in these resources gives the UK social science community a competitive advantage in understanding critical population trajectories over the life course and across changing context.’ but that there had been a ‘large gap in cohort studies on young children’. It recommended that a new UK population-representative birth cohort should be commissioned.

Such studies require significant long term financial investment and a sample design that ensures the sample is representative and does not create bias in the survey data. A new birth cohort study is currently in development, with a feasibility study taking place for a cross UK early years cohort study funded by the ESRC.

In addition to new birth cohort studies, there is a need for additional longitudinal cohort studies of children and young people to understand their lives. The pandemic has created a specific need to understand its impact on the lives of children and young people. The Covid Social Mobility and Outcome (COSMO) study, commissioned by ESRC, began fieldwork in Autumn 2021 and will follow 12,000 young people in England to understand the short, medium and long term impacts of the pandemic on educational inequalities and social mobility. We understand that difficulties in accessing suitable sampling frames have restricted the current survey to England only. In England the Department for Education’s National Pupil Database was available as a sampling frame. Some similar questions to those used in the COSMO study will be asked on the Growing Up in Scotland study and on Schools Health Research Network studies in Wales, which may enable some cross-country analysis.

Education related cohort studies in England are funded by the Department for Education. An example is the ongoing SEED study, which has studied early education use and child outcomes for pre-school children in England and will be collecting further data from the cohort at ages 10/11. Also, the Children of the 2020s study, which is ongoing in England, and aims to understand early education and childcare outcomes for children born in 2021. The Department for Education told us they are also planning longitudinal studies of primary, secondary and post-16 education students to aid understanding of the factors that influence education and well-being outcomes. These will provide useful information on the educational outcomes of children and young people in the coming years.

B1.6 Comparisons between countries

There are limitations with the data that are available to compare across the countries of the UK and to compare the UK with other countries internationally. Many of the areas of responsibility relating to children and young people are devolved, there are differences in government policies and data collections, making comparisons across the UK difficult.

Users told us that there can be a lack of comparability in data on children and young people between the countries of the UK due to differences in the data collected or the metrics that are produced. This makes it difficult for researchers to understand the experiences of children in the different countries of the UK. For example, it is not possible to compare estimates of obesity in secondary school pupils across the UK, due to differences in the age groups and modes of data collection in each country.

In addition, users told us that there is a mixed application of the legal frameworks to sharing data between the different countries and different pathways for accessing data, which can make cross-administration research projects challenging.

The OECD report Measuring what matters for Child Well-being and Policies found that although the availability of cross-national child data has improved considerably in recent years, comparable data are still lacking and limited in scope. For example, there are limited comparable data on concepts such as children’s ‘social capital’, which relates to their perceptions and confidence in their social and cultural identities, their participation in group activities, their trust in institutions, and their knowledge of global and societal issues. Internationally, data often fail to represent certain age groups for specific topics, such as the well-being of children of school age, especially in comparison to adults.

In the education space, all four countries of the UK participate in the Programme for International Student Assessment, PISA, which measures 15-year-olds’ ability to use their reading, mathematics and science knowledge and skills to meet real-life challenges across more than 90 countries and administrations. This study typically takes place every three years, with the next wave taking place in 2022. This provides inter-country comparisons of education and related concepts and is widely used for comparing the education systems across the UK. England and Northern Ireland also participate in the Trends in International Mathematics and Science Study and Progress in International Reading Literacy Study. Wales and Scotland cite the costs and burden of participation in comparison with the benefits as the reason they do not participate.

B1.7 Availability of data at local area level

There are some very good examples of data collection and presentation at a local level. However, overall there are weaknesses with the data available to inform decision making at a local level. Many national surveys are not available at local authority level due to sample sizes, and locally run surveys do not provide estimates for neighbouring or similar local areas to enable comparisons.

Many of the key services and support available to children are funded, commissioned and delivered at a local level. Robust data are needed to help understand the needs of children and young people at a local level and whether local services are meeting this need.

We found that there are some data on children and young people that are available for local areas from official statistics producer organisations to help meet this need. These include the Office for Health Improvement and Disparities childhood and maternal health website, which includes a range of indicators at local authority level with comparisons to national data, the Welsh Index of Multiple Deprivation 2019: deprivation analysis relating to young children which provides some data at small area level and the Statistical Profile of Children’s Health in Northern Ireland 2018/19.

School level data are also available in all four nations of the UK. The school level data are available via Compare School Data in England, School Information dashboards in Scotland, My Local School in Wales and Schools plus in Northern Ireland and includes numbers of pupils, attendance, pupil teacher ratios and funding. Prior to the pandemic, attainment data for secondary schools was available in all four countries and attainment data for primary schools was available in England and Scotland.

There are also examples of where official data are used to provide insight for local areas. For example, The Child Welfare Inequalities Project App uses government data related to children’s services to provide visualisations and analysis that explore links between indicators and deprivation at a local authority level.

We found that, where they are available, data and statistics at a local level are predominately sourced from administrative or operational data collected by local authorities. These data are commonly derived from funding decisions and delivery of services. These data can provide useful insights, but are limited in depth and scope. For example, service level data gives limited insights into experiences and outcomes. National level surveys that provide these insights typically do not have large enough sample sizes to provide estimates for local areas, for example local authorities. Wider data at a local level, particularly on experiences and outcomes, would support a better understanding of the impact of interventions.

We found some examples of surveys of children and young people that are run at local level. For example, the Bee Well project is a collaborative project which includes the Greater Manchester Combined Authority, the University of Manchester and the Anna Freud Centre. It surveys all secondary schools in the Manchester region about aspects of pupils’ lives that influence their well-being and provides valuable insights and information for those working within the support sectors. However, these are not replicated across other local authorities which would support comparison of experiences and outcomes between local areas.

B1.8 Gaining access to data

Improvements to the sharing and linkage of data are needed to enable a more rounded picture of children and young people’s lives to be seen whilst protecting their data.

Several users reported difficulties in gaining access to data on children and young people for research purposes, which they felt was a complicated and time-consuming process. It should be noted that many of the issues that users told us that they face in gaining access to data on children and young people mirror those for gaining access to data more widely.

One example of the issues raised with us was included in our 2019 review of the Public Value of Post-16 Education and Skills Statistics in England. In that review users raised issues with access to Department for Education (DfE), learner level data across all levels of study. Similar issues were also raised with us during the early part of this review. Since September 2018, DfE has given access for research purposes to a number of datasets via the Office for National Statistics’ Secure Research Service.  These are detailed in the DfE external data share transparency data. Many of the issues raised with us, such as the requirement to visit a secure site to gain access to the data, have since been resolved due to changes in the way that data can be accessed via the Secure Research Service.

There are several legal gateways that allow access to data for research purposes. These include the Digital Economy Act which permits public authorities to share de-identified information with accredited researchers for the purposes of research in the public interest. Differences in the understanding of, and the application of, legal gateways between department and administrations were reported to us.

These have led to issues in gaining access to some datasets for research purposes. For example, there can be confusion around whether datasets can be linked, and identifiers then removed to enable linked datasets to be shared. Significant improvements to departments’ approaches to data sharing have been seen through the work of Administrative Data Research UK (ADR UK). For example, ADR UK has funded the Ministry of Justice Data First project which links data from across the justice system and enables accredited researchers to access the data in a secure way. ADR UK told us ‘We need to build better processes so that every child is counted and heard.’

B.2 Vulnerability lens

B2.1 Identifying groups of vulnerable children

We found that official statistics on vulnerable children and young people are limited and do not meet the wide range of user needs. There are many challenges in ensuring the experiences of children and young people who are vulnerable to poorer outcomes can be separately understood. Not least of these are the number of categories or experiences that can make a child vulnerable to poorer outcomes. We identified some significant gaps where children in vulnerable groups are not visible. It was reported to us that some proxies for vulnerability are not always fit for purpose.

In their report ‘Trends in Childhood Vulnerability’, the Children’s Commissioner for England found that ‘Data on how the prevalence of childhood vulnerability is changing over time is limited. It is not possible to explore this question reliably for many vulnerable groups, such as children in need of protection, with health issues, educated outside mainstream schools, or facing housing difficulties.’ In particular, the report highlights that there are data gaps around children who do not meet statutory thresholds for support.

The Human Rights of Children in Wales: An Evidence Review, commissioned by the Children’s Commissioner for Wales, found data gaps in relation to outcomes for some groups of children. In particular, relating to asylum-seeking and refugee children, disabled children and children from ethnic minorities living in poverty. Users also reported to us difficulties in understanding the lives of other groups of children who are vulnerable to poorer outcomes. These include children’s whose parents are in prison or the military, young carers or children being cared for by a sibling or grandparent.

We heard that where indicators that can help identify disadvantage do exist, they are not always fit for purpose. For example, users want to understand the impact of poverty on children’s education. Eligibility for free school meals is often included as an indicator in education related datasets. However, we were told that this an inadequate measure of poverty for identifying those most at risk of disadvantage. The Child Poverty Action Group has found that two in five UK children under the poverty line are not eligible for free school meals. Therefore, the targeting of this group may not adequately meet the needs of supporting children in poverty. In addition, as universal entitlement to free school meals in some primary school age groups is increasingly rolled out, the incentive to register children as eligible for free school meals is removed for some age groups. This may impact on the usefulness of this indicator of deprivation in the future. Our Review of Income based Poverty Statistics highlights the issues with measuring poverty more broadly, which impact on the measurement of children in poverty. The UK parliament Work and Pensions Select Committee has also published a report on Children in poverty: Measurement and targets which makes wide ranging recommendations for improvement to the measurement of child poverty.

Our interim report on children and young people statistics in the pandemic, noted the difficulties in recording which groups vulnerable children belonged to, and how different definitions and coverage can lead to different estimates. An example of this can be found in the Department for Education Attendance in education and early years setting during the Covid-19 pandemic release, which highlights limitations in counting children with a social worker. The NHS Digital Mental health survey of children and young people 2020 also highlights the limitations associated with the higher non-response rate identified amongst children and young people who were at a greater risk of mental disorders. A non-response adjustment was introduced to account for this, but the report still acknowledges that the estimates of mental disorders given may be an underrepresentation of the actual rates.

Another category of vulnerable children raised with us was those experiencing multiple adverse childhood experiences. Adverse childhood experiences (ACEs) are stressful and potentially traumatic events or situations that occur during childhood or adolescence. The Early Intervention Foundations report, Adverse childhood experiences: What we know, what we don’t know, and what should happen next, highlights research which has demonstrated the impact of adverse childhood experiences on health and other outcomes in later life. The report also describes the evidence around the prevalence of ACEs and the issues with measuring it. It highlights that the prevalence of ACEs, and certain ACEs happening together, is not known. The Youth Wellbeing Prevalence Survey has since provided estimates of the number of children who have experiences ACEs and the most common ACEs in Northern Ireland. Public Health Wales has published reports on the experiences of adults in Wales who had suffered ACEs in childhood.

Some forms of vulnerability may be temporary. The UKSA Inclusive Data Taskforce has recommended that: A wider range of methods should be considered for capturing those temporary experiences that are not often well recorded – but which may be important for inclusion. These include pregnancy, hospital stays, school exclusions, periods children spend looked after by the local authority, “sofa surfing”, and periods in prison or on remand.

B2.2 Inclusion of children in vulnerable groups in official statistics surveys

Most surveys of children and young people are either household or school based. Some of the most vulnerable children in society will not be captured by these data collections as they are less likely to be in mainstream schools and more likely to be living in institutions.

Where groups of children are not included in data collections their experiences and outcomes cannot be understood.

Vulnerable children are less likely to be in mainstream schools, either due to persistent absence, or due to attending special needs schools or pupil referral units. Similarly, vulnerable children are more likely to be living in institutions which will not be included in household surveys.

Similar issues face longitudinal statistical studies, as gaining access to the families of vulnerable children is more difficult. In part, this is due to the difficulties in identifying vulnerable children through sampling frames, and to the additional ethical issues in including such children in a study. The Economic and Social Research Council, who part-fund the main birth cohort and other longitudinal studies in the UK, told us that they are commissioning qualitative work to understand how vulnerable children and young people could be included in the longitudinal studies that they fund.

In our interim report we found that some key data sources used to understand the pandemic risked excluding the most vulnerable children and young people. For example, the ONS Schools Infection Survey excluded special schools, pupil referral units and further education colleges, due to the purpose of the study and the potential difficulties in administering it in these institutions. However, these are institutions where vulnerable children and young people are likely to be over-represented. In addition, enrolment for both the pupil and parent surveys was only available online. Whilst email is often the main method of schools communicating with parents, this approach risks excluding vulnerable children who do not have full access to digital technologies outside of the school environment.

An October 2020 research report by the Department for Education  identified that there has been substantial gaps in the evidence base for how experiences of remote education varied for children and young people with particular characteristics including sex, ethnicity, LGBT and economic or social disadvantage.

The UKSA Inclusive Data Taskforce also raised the issue of groups that are underrepresented in official surveys. They recommended that a mechanism was needed to regularly review groups that are under-represented in official statistics and data collection exercises and work done to improve this. Some of the priority groups highlighted relate to children, such as disabled children and parents, children in food poverty, the “hidden homeless”, and small groups such as Gypsy, Roma and Traveller communities, scheduled caste and tribe Asian and African groups. They also recommend that data producers should evaluate the coverage of non-private household population groups in UK data and take the necessary action to address those missing from the current data.

B2.3 Service level data versus data on vulnerable children themselves

We found that the statistics that are available on vulnerable children tend to be metrics of service use, rather than statistics about the experiences of the children themselves.

Much of the official data on vulnerable children are primarily service level data. For example, the number of referrals to social services, or the number of children in care. Often the primary aim of these data is to identify the pressure on the services and drive funding decisions, rather than to gain an understanding of vulnerable children’s lives. These service level data are often supplemented by data on the characteristics of the children receiving the service. An academic we spoke to explained that ‘In social care the state is the parent when children enter the care system. We should expect the same for them as for our own children. Data and statistics do not support us to understand if this is the case’.

Importantly, some of the most vulnerable children do not have any contact with support services. Users told us that more data are needed to understand the lives of vulnerable children beyond simply the number of contacts with services. This is one area where civil society has stepped in during the pandemic. The Royal College of Paediatric and Child Health has collated a list of Covid 19 – research studies on children and young people’s views, most of which are from non-official sources. Studies include research on chronic illnesses, experience of the justice system, young carers and those with experience of mental illness.

Our interim report highlighted one example of official statistics that have been collected and published on the experiences of vulnerable children. The NHS Digital mental health survey of children and young people 2020 study followed up responses from the 2017 survey to ask about experiences during the pandemic. Results were separately reported for those who possibly or probably had a mental health disorder. During the pandemic the Department for Education undertook a vulnerable children and young people survey which aimed to help understand the impact of the coronavirus (COVID-19) pandemic on local authority run children’s social care.

B2.4 Holistic view of vulnerable children

Vulnerability is often a cumulative issue. Having one characteristic or experience that could lead to poorer outcomes may have minimal impact on children and young people’s lives, but having multiple is likely to increase the impact. This makes the need for statistics that give a holistic picture of children’s lives even more important. It can be difficult to gain a holistic understanding of the lives of vulnerable children and young people through official statistics.

In a similar way to the difficulties of getting a holistic view of children and young people, users reported specific issues with gaining a holistic view of vulnerable children. For example, there are limitations in the data collected on referrals to social services – in England only one category can be recorded for reason for referral, when the reasons for referral are often complex. For children in care there is also little information on their family of origin.

There are a number of reasons why a child may be vulnerable to poorer outcomes and for a particular child, there will often be more than one reason. As an example, the latest statistics on outcomes for children in need in England show that just over half of children are eligible for free school meals, nearly half have a special educational need and one in eight looked after children are persistently absent from school. Small numbers of children in some vulnerable groups, and the sensitivity of the data, increases the risks around the sharing and linkage of data for research purposes and limits analysis. These constraints limit the ability of statistics to provide a holistic view of vulnerable children’s lives.

Users told us that understanding the family situation of vulnerable children is important, as these influences can affect life chances. In addition, some of the experiences that can make a child vulnerable to poorer outcomes arise from the family situation, for example, children living in poverty. As noted in the previous section, users told us that indicators of poverty are either only available at an area level, are poor proxies for true poverty, such as free school meals, or are family level indicators that do not allow analysis of the children themselves.

The multidimensionality of vulnerability makes it even more important that a holistic view of vulnerable children’s lives can be understood. This will help identify the groups that are most at risk of poorer outcomes and the levers that can help improve their lives. It will help direct funding to where it can make the biggest difference.

B2.5 Sharing data on vulnerable children

Small numbers of children in some vulnerable groups, and the sensitivity of the data, increases the risks around the sharing and linkage of data for research purposes and limits analysis.

The Public Services Committee in their report on ‘A critical juncture for public services: Learning lessons from Covid-19’ state that ‘We are concerned that agencies do not share the data that they need to support vulnerable children and to determine which children need their help. The Government should issue new guidance on data sharing powers and duties to protect vulnerable children, and, if necessary, introduce legislation to ensure that such data are shared.’ Whilst this comment is aimed at data sharing at the micro level to support individuals, any such improvements will support better data sharing and linkage at the macro level.

Users raised with us that they recognise the high level of sensitivities around data on vulnerable children, and that the small numbers of children increase the risk of identification. However, access to these data is crucial to understanding and improving the lives of these children. One researcher suggested that additional arrangements should be put in place to enable a small number of specific approved researchers to access the data and derive the value from it.

In recent years there has been developments in this area. Children and Young People is one of ADR UK’s flagship research themes, and their Children and Young People research theme provides farther details of current projects.

B.3 Voice lens

B3.1 Surveys of children themselves

Children and young people have the most knowledge about their lives and what is important to them. Whilst there a number of examples of where children and young people are included in surveys, children and young people are currently not consistently given a voice through official statistics.

Most surveys that inform official and National statistics are of those age 16 or older.

There are some examples of where children and young people are surveyed directly in official surveys in specific areas. These include on topics such as health, crime and sports participation. In recent years there has been a notable increase in surveys of children that cover aspects of wellbeing, such the Youth Wellbeing Prevalence survey in Northern Ireland.

All the countries of the UK also take part in international studies relating to children and young people such as, Health Behaviours in School-age Children and the Programme for International Student Assessment. Both of these studies collect the views of children and young people.

In addition, longitudinal cohort studies of older children, such as the Longitudinal Study of Young People in England, survey children directly.

These surveys are providing important data on the lives of children and demonstrate that children’s views can be collected via official surveys and the benefits of doing so.

Our interim report found that a notable gap in official statistics produced during Covid-19 is the lack of data on social outcomes that give a voice to children themselves. We found that much of the data that did give children a voice came from non-official sources, outside of official statistics.

In our research we also found examples of official statistics producers working with others to collect data from children and young people. As an example, the Department of Education in Northern Ireland has commissioned modules in the Kids’ Life and Times, an annual survey of Primary 7 children which is carried out by ARK, a joint resource based in Queen’s University Belfast and the University of Ulster.

Organisations outside of government including charities, independent organisations and think tanks carry out numerous research studies on children and young people. Many of these organisations have also considered alternative methods of responding for children without access to electronic devices or the internet and come up with creative ways to encourage young respondents to express their thoughts and feelings. For example, the Coronavirus and Me survey run by the Children’s Commissioner for Wales which encouraged children under 7 to draw pictures to express their thoughts and feelings.

Whilst the role of civil society is important in providing evidence, and official statistics producers should not seek to replace these types of surveys, the non-official surveys demonstrate that there are many ways to collect this data that children and young people can engage with.

The UK Statistics Authority Inclusive Data Taskforce has recommended that data producers should ensure that data collection instruments are accessible to all, recognising differences in language, literacy, and the relative accessibility of different modes and formats. For example, using multi-mode surveys as standard practice and implementing additional adjustments to enable the participation of adults and children with a range of disabilities, and those who experience other forms of exclusion, including digital exclusion.

They have also recommended that data producers should explore opportunities to utilise more varied, innovative and flexible approaches to data collection and combination, where this will be of particular value for enhancing our understanding of the experiences of relevant groups and populations and/or for enabling the inclusion or voices of groups currently under-represented or missing from existing data sources – such as undocumented migrants, those with disabilities, the “hidden homeless”, and children.

The inclusion of children and young people directly in surveys of relevance to them helps ensure that their views are heard and their experiences understood.

B3.2 Use of proxies

Where children’s data are collected there can be a reliance on proxy responses by parents in some official sources.

These responses can include surveys, that ask parents any questions that relate to those under 16, or socio-demographic data collected in administrative systems such as GP registrations or school enrolment.

The UK Statistics Authority Inclusive Data Taskforce has recommended data producers should avoid the use of proxy responses and ensure that the default approach is for self-reporting of personal characteristics, including, where appropriate, collecting information directly from children.

Children’s experiences should be included in data collection in order to fully understand them. There are examples both within the official statistics producer community and more widely that demonstrate the power of data collected from children and young people and the feasibility of doing so.

Users highlighted that there can be significant variation between responses about a child given by an adult compared to those that the child themselves might give. For example, the Children’s Society’s Good Childhood Report 2018 found that a single question about life satisfaction answered by children was a better predictor of self-harm than a list of 20 questions about children’s emotional and behavioural difficulties answered by parents.

One example of where children’s views have been gathered alongside those of parents is the OfCom Children and parents: media use and attitudes report 2020/21. The report looks at media use, attitudes and understanding among children and young people aged 5-15, as well as media access and use by young children aged 3-4. It also includes findings on parents’ views about their children’s media use, and how they monitor and manage it.

B3.3 Statistics that children and young people need

Children and young people are users of statistics, albeit often through an intermediary, and are likely to have different requirements to adults. In general, children’s data requirements are not considered in the development of statistics.

An important theme we found was the need for statistics to be available that meet the needs of children and young people. In order to uphold children’s rights to express their opinions in matters that relate to them, it is important that official statistics producers understand what data children need and how children will want to access them. However, we found there was a lack of research on what their data needs are.

The Census teams in the four countries of the UK have sought to engage children with the Census, particularly to encourage responses by households. For example in England and Wales, ONS has developed education resources programmes aimed at educating children and young people about the importance of the census and the information that it collects. They engaged with teachers to ensure that the resources are accessible and engaging for children. We are not aware of efforts to actively engage with children and young people themselves over the data that would be collected during the census collection. However the consultations undertaken were public, so children were not excluded from participation.

An example of official statistics producers engaging with children and young people around their data needs is set out in the UK review of information about higher education in 2015. Primary research was carried out with prospective students and their parents and advisors to understand their information use in deciding on university. In addition, research was carried out with first year students to explore the information that they had used in making decisions about undergraduate courses. This informed design decisions in the development of the Discover Uni website.

B3.4 Participation in the official statistics

In general, the official statistics system does not seek the views of children and young people in designing official statistics. They are often not given an opportunity to influence the official data that is collected about them.

Children and young people have the most knowledge about their lives and what is important to them. We found few examples of official statistics producers engaging with children and young people to develop statistics of relevance to them. As part of their children’s well being indicator review, ONS carried out focus groups with children aged 10-15 to find out what they need to have a happy life. These discussions enabled ONS to identify some additional potential factors that impact on children’s well-being that were not included in the previous set of indicators.

For their 2021 review of national indicators, Welsh Government worked with a panel of young people to hear their ideas on what should be reflected in the national indicators. Feedback from the wider consultation also highlighted the need for greater visibility of children in the national indictors, and Welsh Government have told us that they are committed to improving this.

The Observatory for Monitoring Data-Driven Approaches to Covid-19 (OMDDAC) report Young People, Covid-19 and Data Driven Decision Making found that young people’s views had not been considered in the development of data driven decisions during Covid-19 and that ‘Government could do more to ensure young people can access information, particularly information about policies which will impact directly upon them’. The report recognises that young people are not a homogenous group and that a variety of mechanisms are needed to engage with them effectively.

B3.5 Engagement over how their data are used

Children have the same rights as adults over their personal data. We found that a greater understanding is needed of children and young people’s views on how their data are used in official statistics. The official statistics system could learn from health research studies in how to effectively engage children and young people in statistics that use their data.

Children’s right to privacy of their data is enshrined in the UN Convention on the Rights of the Child. It is important that the sharing of children and young people’s data is done so with their knowledge to uphold their rights and ensure public trust. The Information Commissioners Office (ICO) has published general guidance on how General Data Protection Regulation applies to children.

During this review a number of issues with how children and young people are engaged with over how their data are used were raised to us.

For example, the ICO Department for Education (DfE) Data protection audit report  highlights the importance of ensuring the information young people receive is clear on how their data will be used. The audit found that ‘The DfE are not providing sufficient privacy information to data subjects as required by Articles 12, 13 and 14 of the GDPR.’ In the Official Departmental Response to the 2020 ICO Audit, DfE stated they have developed a data sharing policy and data sharing strategy and taken steps to improve how they share privacy information with schools as well as simplifying the information they pass on to parents about what DfE do with their data. They state that the changes introduced during Covid-19 have been welcomed by schools.

The Defend Digital Me The State of Data 2020 report highlights the amount of data that is collected via the education system in England through both statutory data collections and use of educational software provided by the private sector. They call for better protections and oversight of children’s data. In addition, their The Words We Use in Data Policy: Putting People Back in the Picture explains how children are frequently missing from the discussion around data use. The report finds that children are often represented in UK data policy either as victims or criminals, rather than as rounded young people who are also data subjects in their own right and given a voice around the use of their data. The report details the importance of the language used around data, arguing that it influences the way we reason about it and the data policy choices we make. It proposes better ways to talk about personal data, ensuring that children’s rights to control data on themselves are prioritised at multiple levels.

The OMDDAC Data Driven Responses to Covid-19: Lessons Learned report also highlights the need for greater transparency and better public engagement around how data are used. The report states that ‘it is imperative that children and young people are included in these conversations’.

Whilst we heard a strong interest in ensuring the children’s rights around their data are respected, proponents for better adherence to children’s rights also recognise the societal benefits that can be achieved through appropriate and safe data sharing. Indeed the Information Commissioner’s Office Data Sharing Code of Practice highlights the societal and economic benefits of data sharing provided that that data are shared in fair, safe and transparent way.

There are examples of where public bodies have sought the input of children or young people in how their data should be used. One such example is the UCAS survey on data use. In 2015 UCAS carried out a survey of applicants to understand their views on how the data that they provide to UCAS can be used. This included questions around the onward sharing of data.

The ADR UK/ Diabetes UK report on Public Views on the use of personal identifiers for linking diabetes and education data for research offers useful insights on young people’s views around the use of their data. The data being used in this project are official data and we heard that the approach taken could usefully be repeated for other projects. The project found that the children and young people involved were happy for their data, including personal identifiers, to be shared for the purposes of the research and that they favoured an opt-out approach to ensure that the findings were representative. However, the project also found that the way that children’s data is used for research purposes is not well communicated, particularly education data.

ADRUK commented that there is a difference in the approach to participation between health research studies and studies from other areas. Patient participation has long been an important feature of health research studies and there could be large benefits to wider research studies of sharing this approach.