Integrated Data for Population Statistics Conference 2018

The Office for National Statistics (ONS) held the Integrated Data for Population Statistics Conference on 9 and 10 July 2018 at the University of Surrey. At the conference, we, together with a number of external speakers and organisations, shared updates on our progress towards the 2021 Census. The conference was an opportunity for a range of stakeholders with different interests to comment on our progress so far. We received useful feedback to incorporate into our future plans for integrating non-survey data into our statistics.

The theme of the event was “Integrated data – Combining data to answer society’s biggest questions”. The event was aimed at those interested in:

  • integrated data and the 2021 Census

  • statistics using alternative data sources

  • methods to combine and link different types of data for population statistics

The slides from the conference are now available, along with a list of the talks that were held at the event.

Integrated Data for Population Statistics Conference 2018 – Day 1 (morning sessions)

Integrated Data for Population Statistics Conference 2018 – Day 1 (afternoon sessions)

Integrated Data for Population Statistics Conference 2018 – Day 2 (morning sessions)

Integrated Data for Population Statistics Conference 2018 – Day 2 (afternoon sessions)

Conference goals

The following table presents the goals of the conference together with an assessment of how we met them.

  Goal   How we met it
  1) To update stakeholders on the research we’ve carried out
  into the design of the 2021 Census and the parallel research
  on the administrative data census.
  We gave a variety of presentations. They showed our plans, the work we’re
  doing for the 2021 Census and our research into the use of administrative data
  for population statistics. We’ve now published all the slides and content. For
  further information please email us at
  2) To discuss the opportunities and challenges of integrating
  This was the focus of the second day of the event. There were many discussions
  on this topic, which are summarised in the Day Two section.
  3) To seek feedback from stakeholders to inform our plans.   We held a number of interactive sessions asking for input from delegates. We’ve
  shared their comments with the relevant project teams, who will incorporate
  feedback and suggestions into their plans, where appropriate.
  4) To inspire confidence in stakeholders that we’re on target
  to deliver the 2021 Census and a transformed administrative
  data-led population statistics system that better meets user
  needs in measuring our increasingly dynamic population.
  Relevant feedback from delegates included that they were:
    · “getting to understand the great work going behind the scenes”
    · “building connections between ONS and local authorities”
    · “keeping up with developments with the census”
    · “understanding the use of admin data in the 2021 Census, and the various
    methodologies still under investigation”

We’d like to thank everyone who was involved in this event and helped to make it a success. We’re grateful for all feedback and contributions received.

In particular, we’d like to thank our external contributors, as collaboration is critical in addressing some of the challenges highlighted. The feedback we received will be used in the planning of future events, as well as informing our ongoing work.

What did we cover?

Day one: Integrated data and the 2021 Census

The day kicked off with an introduction from ONS Director of Population and Public Policy Transformation, Frankie Kay. Frankie welcomed everyone to the conference and shared our plans to transform population and social statistics with administrative data at the core. She emphasised that the 2021 Census was integral in laying the foundations for the future.

Garnett Compton, Head of Statistical Design for the 2021 Census, then provided an overview of the 2021 Census design. This was followed by Pat MacLeod from the Office for Statistics Regulation outlining what the new Code of Practice for Statistics means for the 2021 Census.

The first set of parallel sessions covered:

  • the lessons we learned from the 2017 Test

  • updates on how we’re using new data science techniques to help identify hard-to-count properties in the census

  • an interactive session on the plans to quality assure the 2021 Census

After lunch, Bill South (ONS) and Justin Hayes (University of Salford) demonstrated their work to digitise the 1961 Census small area statistics. This included inviting the audience to quality assure the data using a crowd sourcing solution. You can have a go yourself on the Zooniverse website.

The afternoon parallel sessions included:

  • ways of integrating alternative data into census processing

  • the importance of developing a hard-to-count index for the 2021 Census

  • an opportunity to test out the new flexible dissemination tool that we’re proposing will be used to disseminate census (and potentially other) statistics

Dorset County Council rounded off the day. They talked about how they use a combination of census and administrative data to support service planning, policy formation and decision-making in the local authority. Their needs reflect the needs of many of our stakeholders – the right data, at the right time, in the right way. They provided a great insight into why the data collected in the census are so important.

Day two: Integrated data: opportunities and challenges

The focus for the second day was “What are the wider opportunities and challenges with integrating data to deliver a more responsive administrative data-led statistical system?”. As part of this, we asked delegates to tell us what these opportunities and challenges were. Some of these were then discussed in a panel session at the end of the day. The following table shows the opportunities and challenges that were discussed.

Opportunities with using integrated data Challenges with using integrated data
  Developing and sharing skills, resources and knowledge   Ensuring public acceptability
  Generational and longitudinal linkage   Assessing quality
  More data   Clear metadata
  Build relationships with suppliers   Need new skills
  Collaboration (with suppliers, internationally, and across the Government Statistical Service)   Assessing data
  Revisit user needs   An identifier to link on
  Having a flexible and scalable infrastructure

Frankie Kay again introduced the day and shared some examples of the potential of linking and combining data. This includes being able to answer important questions such as “What is the relationship between personal debt and mental health?”.

Stephen Penneck, Vice-President of the International Statistical Institute (ISI), then gave his thoughts on data ethics. This included the ethical concerns of linking data and using linked datasets in analysis and research.

The first parallel sessions of the day included talks on:

  • how Stats NZ is building public trust in using and linking personal data

  • addressing some of the quality issues that we see with integrated data

  • how ONS is looking at using integrated data to transform population and housing characteristics

In an international session, Becky Tinsley (ONS) and John Dunne (Ireland’s Central Statistics Office) shared some international perspectives of using integrated data to transform population statistics. Marina Pérez Julián from the Spanish Statistical Office (INE) then outlined how previous censuses in Spain have been carried out and presented their plan to move to a register-based census in 2021. Sharing experience and knowledge internationally has a real benefit and is something we’re continuing with at ONS.

The final parallel sessions included an example of cross-government working where data has been linked to understand the educational background of offenders. There were also discussions about acquiring and securing data, and how we are doing this at ONS using the “Five Safes” framework. We also held a session sharing our plans for transforming the population statistics system, placing administrative data at the core to estimate population stocks and flows.

We held a panel session to discuss some of the challenges and opportunities of using integrated data as identified by delegates (see table above). The panel members came from a range of backgrounds to reflect different perspectives in the discussions. The panel included Iain Bell (ONS perspective), Marina Pérez Julián (Spanish Statistical Office (INE) – international perspective), Rachel Leeser (Greater London Authority – local area perspective) and Chris Roebuck (NHS Digital – Data supplier perspective).

The topics discussed in the panel session

  1. Whether we can we do anything to influence public opinion on the use of their data.

    There was agreement that we need to be transparent about the use of data and to be clear about the benefits it can bring. This will help to bring the public along with us.

  2. How we can assess the quality of alternative data sources.

    While there is still a long way to go some steps have been made, such as working with those at the local level who collect the data to improve the quality at source.

  3. Whether it is possible to move from numerous data sources to a single indisputable source of truth.

    The truth is relative and we need to be able to reflect the changing and diverse society that we now live in. To do this, standard definitions and concepts may need to be reviewed.

Iain Bell, Deputy National Statistician, closed the event. His message was that traditional sources of data, and the methods associated with them, no longer match the pace of change that we’re experiencing. We need new insights into the data to inform decision-makers, ensuring they have sufficient evidence to make decisions. These decision-makers are not just the prime minister and local authorities, they are also the general public. Integrated data needs to be at the heart of this.

This event was a great chance for stakeholders to discuss how the new data becoming available to us give lots of opportunities, but also challenges. We need to work together to make the most out of these opportunities, to learn from each other and to collaborate. This will ensure we are producing statistics that reflect the changing society that we now live in.

If you’d like any more information about any of the sessions held on the day, please contact us. We’re also reviewing how we hold future events, so any feedback or suggestions would be gratefully received. Please email us and let us know your thoughts. You can also register for email updates about the 2021 Census and other areas of ONS.