Written evidence from Understanding Society (TEB03)

Public Administration and Constitutional Affairs Select Committee

Transforming the UK’s Evidence Base inquiry

Introduction

Understanding Society is a world-leading longitudinal survey. By tracking individuals and households over time, and across multiple aspects of people’s lives, the Study is a vital data infrastructure resource for providing insights into the causes and consequences of social, economic and cultural change, enabling the advancement of social science research, policy learning and evaluation. It builds on the long-running British Household Panel Study (BHPS), started in 1991, and so for some panel members and common questions across the two studies, data extends across 30 years.

The Study is funded by UKRI’s Economic and Social Research Council (ESRC) to underpin social science research in the UK and beyond. It is also supported by a number of government departments. It is based at the Institute for Social and Economic Research (ISER), University of Essex. Understanding Society is widely used by government civil servants and statisticians to better understand social issues and inform and monitor policies.

Our broad view is that, even in an era of more comprehensive administrative data, surveys have a vital role to play. Combining the two sources of data can help overcome the limitations of each, and enhance the quality and accuracy of the information available to policymakers, researchers, and the public. Our response covers the following three specific questions of interest to the Committee:

• On the changing data landscape, we comment on the continuing importance of social surveys. This section also provides examples of how Understanding Society data is used in the production of official and other statistics which are vital to understand the nature of change taking place in society.
• On protecting privacy and acting ethically, we explain how Understanding Society collects the data ethically, and makes this available ethically in the context of research and policy analysis.
• On understanding and responding to evolving user needs, we comment on the relevance of combining administrative and survey data and innovating in new forms of data collection, and how we can make sure users in the Civil Service, Parliament and beyond have the ideas and skills they need to make effective use of data.

The role of surveys

The ESRC continues to regard Understanding Society as a vital part of UK’s changing data infrastructure for social science research. The ways that surveys are carried out (for example, using mixed modes to collect data) and the purposes for which they are used continues to evolve, with Understanding Society playing a vital part in survey innovation, linking to administrative data sources and responding to emerging societal needs.

A unique advantage of surveys like Understanding Society is that, by collecting repeated data, they can provide a wide range of contextual and explanatory information which extends beyond being able to estimate numbers, or proportions, of people in different situations – whether one is seeking to understand the causes and consequences of rising mental health problems, cost of living crisis or climate change. The role of Understanding Society is to enable researchers to investigate these societal challenges over the short and long run by providing high quality consistent longitudinal annual data on the central domains of people’s lives – education, employment, family, health, and income and wealth. As a household study, we aim to collect data from all the members of the household, facilitating research into children, families and intergenerational differences, similarities and mutual dependency.

Government departments also make use of these data to understand policy imperatives. The Department for Work and Pensions, for example, produce official statistics on persistent low income in the UK over time using Understanding Society data. The Office for National Statistics use the Study to publish annual statistics on topics such as wellbeing and social capital. Other organisations outside of government depend on such data to understand societal needs and influence policy and practice. Administrative data do not provide the wealth of information available on households and individuals over time that enable such data analysis.

Government departments are required to collect data on protected characteristics under the Equality Action 2010, but there will be gaps in information where take-up or use of services varies across different groups. This can lead to gaps in administrative data, and to not being able to get a full picture of population needs, e.g. the mental health consequences of poor work, or transport challenges faced by people with different types of disability. By capturing broader contextual information, surveys play an important role in investigating the needs of specific groups, and more generally into distributional impacts of economic and social change.

In particular, as we move beyond discrete goal-focused policy interventions and into an era of more complex challenges such as spatial disparities, public health, climate change, an ageing society, impact of technology, etc., we can expect to see demand for more multi-dimensional data, the use of more complex methods in research, and growing calls for cross-departmental data sharing and working. While bringing together administrative data across departments will provide some evidence on these complex issues, there will be many gaps if one wishes to investigate issues in a comprehensive way. Unsurprisingly, tackling ‘wicked social problems’ usually depends on complex interventions, which in turn require information on different dimensions of people’s lives. When there is interaction between the intervention and its context that determines outcomes, attention has to be given not only to the intervention but the conditions needed for the mechanisms of change to work.

Surveys like Understanding Society can also be highly responsive to major events, in ways that administrative data cannot do because systems take time to adapt. During the pandemic, Understanding Society was able to quickly roll out an additional online survey to collect information on how people were responding to the pandemic and what impact it was having on their lives. For example, Public Health England’s regular reporting on mental health during the pandemic included use of Understanding Society data, and the data provided vital evidence in understanding which groups were hesitant about vaccination.

Protecting privacy and acting ethically

Understanding Society’s long-term prospects depend on building relationships with panel members (the individuals and households that provide information about their lives) and users. At the core of that relationship is making sure we protect their privacy, and that data provided by the public are used ethically.

For Understanding Society, protecting privacy and acting ethically means:

• transparent communication with panel members which explains the reasons for data collection, how the information will be used, and reporting back to panel members on research findings
• different levels of data access to minimise the risk of disclosure, with the highest level of assessment applied to the most sensitive data
• all staff members being trained in information security – the organisation is accredited with ISO 27001, an externally audited information security standard
• making sure the principle of public good is applied to the use of the data.

Understanding and responding to evolving user needs

Our reflections on this question are based on how we are responding to evolving user needs which includes government social researchers, statisticians and economists, charities and business, as this may have wider relevance and application. As a data infrastructure project, we also set out the challenges we face in realising the benefits of linking survey and government administrative data.

The dataset provided by Understanding Society is large and it can be complex to use, and while we have regular users who have invested time and effort in learning how to use the Study, others are occasional users, particularly ‘policy users’ who want to access data primarily when a specific need arises. Segmenting our user base by type of need and skills level, we are improving access for occasional users in a number of ways:

• It can be time consuming for users to hone in on the data they need. Our ‘code creator’ enables user to efficiently build a data-file by searching for variables of interest and extracting data from the main End User Licence dataset.
• We either produce, or commission research fellows to develop, specialised datasets from the main survey for wider use, and we encourage researchers to share the code written for their analysis. Our PEACH dataset, for example, provides information on pregnancy and early childhood as a specialised dataset.
• Our data dashboard, being built around our five core topics, provides a selective set of variables and socio-economic indicators that enables policy users to plot trends, undertake basic cross-sectional analysis and download charts for use in briefings and reports.

We welcome the creation of the Government Analysis Function (part of the Civil Service), with its focus on learning and development and championing impactful analysis, as it could potentially be a vehicle for building data skills and capacity across government through collaboration with academia.

Realising the benefits of linking administrative data to participants’ questionnaire (and other) data remains a core objective for Understanding Society. This will provide broader context and outcome measures to the government than are available in administrative data. It can also identify issues with the quality of admin data or missing groups. While the challenges of linking survey and administrative data are not new, and it is a complex logistical exercise, we think there is further room for improvement and innovation. Bringing together longitudinal survey data, administrative records, and new forms of data, for example, collected through apps or different types of devices, will generate a substantial resource of greater value than the sum of its parts. Linking the two sources of information more easily will further enable surveys to adapt and innovate.

August 2023