STATISTICS ON SUBNATIONAL POPULATION PROJECTIONS
I am writing to you following our recent review against the Code of Practice for Statistics of the four sets of subnational population projections published by the Office for National Statistics (ONS), National Records of Scotland (NRS), Welsh Government and Northern Ireland Statistics and Research Agency (NISRA):
- Subnational population projections for England
- Population Projections for Scottish Areas
- Local authority population projections for Wales
- Population Projections for areas within Northern Ireland
I am pleased to confirm that all four sets of subnational population projections should continue to be designated as National Statistics.
Subnational population projections are widely used for local planning purposes and the outputs are subject to high user interest. This review focussed mainly on the accessibility of the data, statistics and supporting information, and supplements our previous compliance check of the national population projections in April 2019. We welcome the way that the teams have responded to our earlier letter, and in particular how they are planning to apply the findings there to the Subnational projections.
In reviewing the subnational projections, we found several examples of good quality assurance, clear analysis and explanations, and demonstrable trustworthiness, which we detail separately for each country later in this letter. However, we have identified some key areas for improvement across the four producer teams, some similar themes have emerged, and therefore there are some common recommendations. In order to improve the public value of these statistics, the teams should:
- Continue to apply the findings from the previous review to ensure the positive changes to presentation and supporting documentation in the national publications are applied for the subnational publication planned for later in 2020.
- Consider how existing channels can be used to enhance collaboration between all the producer teams. Cross-producer working arrangements and collaboration for national projections has been effective. However, discussion and consideration of issues relating to subnational projections are not normally included. While this is primarily due to subnational projections being produced separately by different producers with some differing methods, potential opportunities for collaboration and sharing of best practice at subnational level are being overlooked.
- Be transparent about proposed changes to the timings of the publications, as well as future developments, considering the 2021 Censuses. The teams should continue to consider how best to engage with their users in relation to planned developments and changes in the wider statistical landscape, such as any impacts from the ONS migration and population transformation programme.
- Refresh and update the comparison document which sets out the differences in methods used to produce the four sets of subnational projections, to make it more accessible and easier for users to identity differences in approach and the reasons for any differences
ONS’s Subnational population projections for England are presented clearly and objectively. The graphs and interactive graphics do a good job of aiding understanding, and there are clear explanations on the accuracy and use of the statistics. We welcome the team’s commitment to mirror the positive changes to the national population projections. For example, following our previous recommendation, future developments are explained fully, and the relevant quality information is included in the main bulletin. The bulletin published in November 2019 demonstrates a commitment to users and flagged method changes well. As in the national publication, in the next subnational release we recommend that information be provided on the future of the outputs, especially in relation to proposals for the timings of the next publication and seeking user feedback on the proposal.
NRS’s Population Projections for Scottish Areas bulletin is engaging and informative. NRS sets out the key messages clearly and objectively and make good use of charts and graphs. The methods information is well set out in the methods guide, as it is in the national population projections, NRS could improve this information by bringing some of the key information on quality into the main bulletin. The Quality Assurance of Administrative Data (QAAD) documentation is available to support this but could better describe the strengths and limitations of the data sources to give users a clearer picture of the processes in place. We were pleased to hear from the statistical team about its future communication plans and a review of the quality information. As with the national population projections, NRS should ensure future plans are clearly communicated with users, including the impact of any GSS-wide programmes such as the ONS migration and population transformation programme.
Welsh Government’s Local authority population projections statistics and analysis are presented clearly and objectively. There are good descriptions of data sources and data quality, with clear links to the accompanying quality report. The sections on comparisons with earlier data, mid-year estimates, and ONS data are useful and appropriately detailed. Some supporting documentation could be linked more clearly to help users navigate to it, for example, statements on data governance and pre-release access, as well as guidance on comparing with NRS and NISRA data. After speaking with the team, we are pleased to hear that it is considering aspects of accessibility by planning a move to HTML publication as well as PDF. The publication has recently been delayed due to quality issues. This has been clearly explained in an accompanying blog and we look forward to hearing more on this and future plans in line with the 2021 Census in the next publication.
NISRA’s Population projections for areas within Northern Ireland are comprehensive, accessible and the analysis and data are demonstrably valuable to users. NISRA’s supporting material on methodology and data quality are thorough and assure users of the trustworthiness of the publication. The team regularly engage with users through surveys and is planning an in-house user engagement event in 2020 which we welcome and encourage. NISRA could improve by mirroring changes successfully made in the national population projections, such as improving the presentation and readability of the publication.
We appreciate each of the teams’ willingness to engage with us and take on board our recommendations. We recommend that the teams continue to engage users in the proposed changes to the publications and timings and clearly communicate the impact of changes to similar publications or datasets such as those involved in the ONS migration and population transformation programme. Our population and society domain team will continue to engage with your teams over the coming months to discuss progress ahead of the next publication of the subnational population projections.
I am copying this letter to Andrew Nash (ONS), Denise Patrick (NRS), John Morris (Welsh Government) and Jonathan Harvey (NISRA), the lead statisticians.
Assessment Programme Lead