UK Statistical System UK Statistics Authority Code of Practice
Casework

Ed Humpherson to Alison Byrne: 2022 Census in Scotland, phase 3 assessment report

Dear Alison

Today we have published our phase 3 assessment report of the 2022 Census in Scotland. This report, our third and final report, focussed on assessing how well NRS’s census outputs have met users’ needs and revisited the requirements outlined in our phase 2 report published in 2023. On behalf of the Board of the UK Statistics Authority, I am pleased to confirm their status as accredited official statistics having been found to comply with the standards of Trustworthiness, Quality and Value in the Code of Practice for Statistics.

NRS has been dedicated to delivering high quality and accessible census outputs which have met the needs of a wide range of users. Following quality challenges caused by lower-than-expected response rates during census data collection, NRS conducted rigorous quality assurance to understand and assure the reliability of the census results and put in place appropriate mitigations to ensure the statistics’ quality. NRS ensured its outputs were informed by input from users and proactively communicated with the public, including diligence in effectively communicating uncertainty where needed. NRS also developed a range of innovative tools that have generally enhanced users’ experience of using the 2022 outputs. This approach to recognising user needs, alongside NRS’s rigorous approach to quality assurance, were reflected in the high level of satisfaction in the outputs reported to us by users.

NRS has responded thoroughly to the requirements outlined in our phase 2 report, such as by providing accessible information on data quality and strengthening engagement with users. NRS has clearly explained definitions and methods, and strongly evidenced working with the other UK census offices to address the comparability of its outputs to the 2021 censuses in the rest of the UK.

To inform future population work, our assessment report includes five recommendations for NRS to address. The recommendations are largely focused on carrying forward lessons from the 2022 Census into the design and delivery of the 2031 Census. The recommendations include building on lessons learned, working with users requiring data-zone level outputs, and continuing collaboration with other UK census offices to support comparability of outputs. NRS should also continue to build on feedback regarding the useability of the flexible table builder, as well as working with expert user groups during question design and development ahead of the next census. NRS should aim to address these recommendations within six months, or demonstrate how they have been incorporated into planning for the 2031 Census.

I would like to thank your team for their positive contribution and engagement throughout the assessment process. If you would like to discuss any aspects of this letter or the accompanying assessment report, please get in touch.

I am copying this letter to Alastair McAlpine, Chief Statistician, Scottish Government.

Yours sincerely

Ed Humpherson
Director General for OSR

Rob Kent-Smith to Peter Connolly et al: Review of Assessment of salmon stocks and fisheries statistics

Sent to:

Peter Connolly, Environment Agency 

Simon Toms, Environment Agency 

Henry Knowles, Natural Resources Wales 

Ben Wilson, Natural Resources Wales 

David Maxwell, Centre for Environment, Fisheries and Aquaculture Science 

Alan Walker, Centre for Environment, Fisheries and Aquaculture Science 

 

Dear all,

We have recently completed our review of the statistics from the Assessment of Salmon Stocks and Fisheries in England and Wales. This publication is jointly produced by the Environment Agency (EA), Natural Resources Wales (NRW) and the Centre for Environment, Fisheries and Aquaculture Science (Cefas). Although these statistics are not labelled as Official Statistics, we agreed to carry out a short review in light of proposed changes to methods and reporting, in order to offer advice, support and recommendations to help you best manage these changes in line with the Code of Practice for Statistics.

Our review looked at how the proposed changes to methods have been handled, how the report is structured, the status of the output and how it is published, and the general statistical and Code culture. We found that producers are developing more transparent, robust and reproducible methods and have begun improving data timeliness. However, unclear timelines for these method developments and inaccessible reporting formats as well as weaknesses in statistical leadership and compliance with the Code risk damaging user confidence in the statistics. We have made several recommendations to support improvements in these areas and expect you to provide a public update on progress against them by November 2026.

I am grateful for the constructive engagement from you and your teams throughout this review and look forward to receiving updates on progress against the recommendations.

I am copying this letter to Ian Lonsdale, Head of Profession for Statistics, Defra.

Yours sincerely

Rob Kent-Smith

Deputy Head, Office for Statistics Regulation

 

Ed Humpherson to Alastair McAlpine: Request for assessment of the new fraud and computer misuse questions in the SCJS

Dear Alastair, 

Thank you for your letter of 22 April requesting an assessment of the new fraud and computer misuse estimates from the Scottish Crime and Justice Survey (SCJS), produced by Scottish Government. 

Given the public value of information on fraud and computer misuse victimisation, I welcome your aspiration to publicly demonstrate compliance with the Code of Practice for Statistics. I am pleased to confirm that our teams have agreed the assessment will start in May 2026. 

Yours sincerely,  

Ed Humpherson
Director General for OSR 

Related links: Alastair McAlpine to Ed Humpherson: Request for assessment of the new fraud and computer misuse questions in the SCJS

Alastair McAlpine to Ed Humpherson: Request for assessment of the new fraud and computer misuse questions in the SCJS

Dear Ed,

I am writing as the Chief Statistician for the Scottish Government.

For the first time in 2023/24, the Scottish Crime and Justice Survey (SCJS) was expanded to collect new information on adult experiences of fraud and computer misuse. This was developed following interest from users, to keep the SCJS in line with the changing nature of crime and to further align with the Crime Survey for England and Wales (CSEW).

The new victim form that underpinned this move was largely developed, with some modifications, from the one used by the CSEW. Throughout this process, colleagues from ONS provided ongoing advice and guidance, which added significant value. Once the questionnaire was finalised, a panel of Ipsos and ScotCen interviewers conducted a pilot exercise. Their feedback was helpful in improving the questionnaire and adding additional explanations to aid respondents.

For the first year of these results, Scottish Government analysts worked with colleagues at ScotCen to conduct a thorough review of all victim forms collected for the purpose of quality assurance and categorisation. Aside from these enhanced quality assurance checks, the data is processed in the same way as has been established for the existing property and violent crime victim forms.

The first findings from these questions were published in June 2025 (1) with the second published in March this year (2).

In accordance with Section 12(1) of the Statistics and Registration Act 2007, I would be grateful if your Assessment Team can undertake an assessment of the recently introduced fraud and computer misuse questions within the Scottish Crime and Justice Survey (SCJS).

We look forward to working with the assessment team on this matter.

Yours sincerely

Alastair McAlpine
Chief Statistician

James Benford to Ed Humpherson: ONS Reports on Progress with Economic Statistics, April 2026: OSR response

Dear Ed,

Thank you for your comments on the second progress report on the Economic Statistics and Survey Improvement and Enhancement Plans.

I welcome your recognition of the progress we are making and of our transparent approach with regards to the challenges that we are facing and how we are adjusting priorities in response.  As I noted in my previous exchange with you, which followed the first quarterly update, this is an iterative process.  The Office for National Statistics is currently finalising its high-level business plan, including the allocations of budgets across the office.  The next update in July will be the point when we are able to set out a set of key performance indicators, in the context of agreed organisation-wide objectives, the specific milestones we will be working towards on our improvement plans, as well as the proposed initiatives that we will be holding in a ‘waiting room’ to protect delivery confidence in those milestones.  It is critical that we learn from the challenges we’ve had over the past year and, as the organisation also steps up to deliver the next census including a test in 2027 and implement a new set of macroeconomic statistical standards, to ensure that we do not take on more than we have capacity and capability to deliver.

You note in your letter the achieved response level for the Labour Force Survey is now close to its pre-pandemic level, but there are challenges in other social surveys.   Whilst the period of double-running two labour force surveys is constraining what we are able to do, we are doing what we can to stabilise and then increase the achieved sample sizes.  For the Living Costs and Food survey, we are forecasting an achieved sample size of 5,200 at the end of the Fiscal Year 2026, compared to 5000 in 2025 and 4200 in 2024, and a consistent 5,000 in the years before the pandemic.  On the Wealth and Assets Survey, our achieved sample in ‘Round 9’ for the period 2022-24 was 11,708 households, ‘Round 10’ for 2024-2026 will be 11,912 households and for the next ‘Round 11’ we are targeting 18,000, in line with the pre-pandemic position.  Our planned survey response tracker will bring together information on achieved sample sizes and response rates to allow users to track these performance indicators.  You will have also heard at the UKSA Board last week that the average response rates on business surveys are approaching an historic high, supported by the move to digital surveys and a changed approach to businesses who initially do not respond to our surveys.

You also note the challenges that we have laid out with the Transformed Labour Force Survey (TLFS) and the Statistical Business Register.  On the former, we set out in our first update in December that work to introduce data rotation had been more complex than anticipated and would not complete until April at the earliest, rather than in January as had been planned for originally.  We have hit the delayed April milestone and have now implemented all agreed major design changes to the TLFS.  The later milestone means we will only have partial information on how the survey has evolved in light of those design changes and so, having liaised with stakeholders, we have confirmed we now expect that we will not transition to the TLFS as our main headline measure until 2027.

As we set out in our update, following a review, the Statistical Business Register project is working through a 10-week period to strengthen its organisation and approach to delivery.  This is focused on clarifying governance and accountabilities, tightening scope and success measures, and establishing an iterative delivery plan that releases value to register users in smaller increments, with earlier and more frequent opportunities for feedback and course-correction.  The outputs from this 10-week work will inform an updated plan and timetable for the SBR, including how we will manage dependencies and delivery risks, and we will update users and the Office for Statistics Regulation on our emerging plans at the appropriate point.  I welcome your plans to support our work by undertaking an OSR review this summer.

I look forward to working with you and your teams as we prepare and finalise our July update, including our more detailed milestones and plans, to close off the remaining recommendations you have made.

I would also like to thank you for the initiative you took to host with me a workshop bringing together suppliers of administrative sources of data underpinning official statistics and the producers of those statistics last week.  I found it to be a highly collaborative and productive event that identified a number of actions to strengthen the broader data eco-system that is vital to the quality of statistics.  I suggest we work these up into an action plan that can be shared with the Government Statistical Service and the UK Statistical Authority Board.  As I noted in our previous exchange, errors in source data were the root-cause of the majority of major errors to ONS economic statistics last year.  It is essential that we collectively take action to learn from these errors by strengthening the system and I really welcome your support on this.  I look forward also to learning of the findings of the review you announced last November of HMRC statistics against the Code of Practice.  This review is particularly important for improving the quality of public finance and trade statistics produced by the ONS and we will do whatever we can do to support the recommendations that follow from it.

Yours sincerely,

James Benford

Director General, Surveys and Economic Statistics Group

Related links: Ed Humpherson to James Benford: ONS Reports on Progress with Economic Statistics, April 2026: OSR response

Ed Humpherson to James Benford: ONS Reports on Progress with Economic Statistics, April 2026: OSR response

Dear James,

We have today published our response to the Office for National Statistics (ONS)’s second quarterly progress report on its Plan for Economic Statistics and Survey Improvement and Enhancement Plan. Our response assesses ONS’s progress in addressing the findings of OSR’s systemic review of its economic statistics, including our interim and final report, and the recommendations of the Devereux review.

We note that the ONS report shows both continuing progress and sets out ongoing challenges. We particularly welcome the increased transparency on progress against milestones, on prioritisation, and on the ongoing challenges. However, we believe transparency could be further improved, and we welcome the proposal to introduce a set of outcome-based Key Performance Indicators (KPIs).

On specific issues:

  • We acknowledge the restoration of the achieved sample size for the Labour Force Survey to near pre-pandemic levels but also note the continuing challenges faced in restoring response rates across social surveys as a whole.
  • We also note other key challenges, particularly delays to the Transformed Labour Force Survey and slower than anticipated progress with the Statistical Business Register, a key enabler of transformations and improvements to key statistics.

In our response to the first quarterly update, we recommended that ONS should:

  1. Provide more succinct and accessible quarterly summaries, focused on prioritised user-facing milestones, with clearer information on progress, risk and changes to planned delivery timescales.
    • We judge that this has been partially met
  2. Include more-disaggregated information on the costs of major outputs to support informed stakeholder engagement as part of its business planning processes.
    • We judge that some progress has been made, but more needs to be done
  3. Report further on progress with quality reviews, workforce recruitment and retention, and replacing the legacy systems that underpin core economic statistics in the next update.
    • We judge that this has been fully met

We will be in a better position to make a comprehensive assessment of progress after the next quarterly update, in July 2026.

We expect that our future responses to ONS progress reports will focus solely on delivery rather than also on presentational matters. We intend to review the periodicity of our responses to ONS quarterly updates as a regular rhythm of reporting and delivery are established.

Yours sincerely,

Ed Humpherson
Director General for OSR

Related links:

James Benford to Ed Humpherson: ONS Reports on Progress with Economic Statistics, April 2026: OSR response

John Wilkins to Ed Humpherson: Asylum Seeker Hotel Population Statistics

Dear Ed,

I am writing to inform you that I have become aware of the unscheduled release of an internal Home Office figure that appears similar to an upcoming official statistic from our Asylum Seeker Hotel Population Statistics series.

Despite not being released publicly through formal channels, a figure for the number of asylum seekers in hotels “this month” was quoted in two newspapers on Thursday 16 April:

  1. The Daily Telegraph: https://www.telegraph.co.uk/news/2026/04/14/labour-move-10000-hotel-migrants-into-communities/
  2. The Daily Express: https://www.express.co.uk/news/uk/2194138/labour-migrants-hotels-close-most-remain-open

The figure quoted was 20,800, “down by a third since December” (30,657). Other figures and Home Office contributions were also included in these articles, but arose from appropriate sources. At the time of writing this figure has not seen significant re-use beyond the above, although it has been picked up by a very small number of news aggregator websites.

Figures in this series are released quarterly, with the last release being of data as of the end of December 2025, published in February 2026. Our next, already pre-announced update to this series is scheduled for Thursday 21 May 2026.

I should be clear that the figure of 20,800 does not appear to be a regular official statistic in its final form. For example, the figure is rounded, when our usual estimates are unrounded. However, my main concern here in raising this matter is that it could be confused as being part our regular statistical series by users and thus – in line with relevant guidance – I believe this should be viewed as a breach of the Code of Practice for Statistics.

At the time of these articles being published, Home Office statisticians were not in possession of the required data to calculate the above figure. Work is already underway to understand the circumstances around the figure’s unscheduled release.

My team and I have had to consider whether to follow the above with our own orderly ad-hoc release of an estimate for this time point. After reviewing relevant guidance and consulting colleagues in the UK Statistics Authority – amidst the current absence of a National Statistician – I have concluded that an ad hoc release is not necessary. This is particularly noting:

  1. that the scheduled release of the actual Home Office official statistic is now relatively soon – scheduled for Thursday 21 May
  2. that there has not been significant re-use of the figure – for example, other news outlets have published stories on this topic whilst still referring to our last end of December 2025 estimate

I intend to keep particularly point (2) above under review, as I recognise this is a situation that could quickly change.

In addition to informing you of this matter via this letter I will arrange for an associated note to be placed on the ‘Statistics at Home Office’ website accessible via GOV.UK.

Yours sincerely,

John Wilkins

Chief Statistician, Home Office

Robert Kent-Smith to Liz McKeown and Alex Lambert: labour force survey transformation

Dear Liz and Alex 

Today, we have published an update report about how ONS is improving the Labour Force Survey and progressing its transformation. This current report reflects on ONS’s progress so far and brings together feedback shared by users in autumn 2025 and spring 2026.  

ONS has fully met four of the twelve recommendations from our 2025 update report, and has partially met the remaining eight, as detailed in the section Progress against recommendations. ONS has been increasing the response for the LFS and prioritising the design of the transformed survey. There has been encouraging progress in communication, engagement and development of the TLFS Core, and we recognise the effort invested in these areas. 

There remain challenges to overcome in continuing to deliver the TLFS programme. ONS is focusing on ensuring that methodological and design developments are delivering improvements. Extended dual running of the LFS and TLFS continues to strain resources. To strengthen user confidence in the transition to the TLFS, ONS should publish a plan outlining broader user engagement and communication as it nears the transition and continue its work to overcome the remaining methodological and processing challenges, completing the remaining recommendations in 2026. 

Statistics from the existing Labour Force Survey (LFS) remain as official statistics in development. We note that ONS is already progressing work to ensure full compliance with the updated Code of Practice for Statistics (Code 3.0), and we welcome the commitment to embed these standards as part of ongoing development of the statistics. 

ONS has agreed to continue to publicly report its progress through its quarterly updates on progress and plans for labour market transformation and updates on its Economic statistics and surveys improvement plans. 

I am copying this letter to Heather Bovill and David Freeman, Deputy Directors Labour Market and Household Accounts, in ONS. 

Yours sincerely 

Rob Kent-Smith 

Deputy Head of the Office for Statistics Regulation 

Mary Gregory to Ed Humpherson: Update to ONS’s Long-term international migration estimates

Dear Ed,

I am writing to you with an update on how we have developed long-term international migration estimates since the last OSR progress report in December 2023. There has been considerable improvement to the output, and this work is continuing. Our International migration research, progress update, published today, provides an update on the research underway that will further improve our admin-based migration estimates.

It is important to recognise in the last few years there has been significant research into administrative data for migration estimates. Our increased understanding of the Home Office Borders and Immigration (HOBI) data and the Department for Work and Pension’s (DWP’s) Registration and Population Interaction Database (RAPID), and how to use them, have helped us to improve long-term international migration (LTIM) statistics. We continue to review other potential data sources that could be used in the future to help improve these statistics.

In November 2025, we reached a major milestone that transformed this output. We published our work introducing new methods for British National migration and EU+ migration. This meant that for the first time, LTIM estimates no longer relied on International Passenger Survey (IPS) data. This was a significant step in improving the quality of these statistics, moving away from reported intentions to migrate and towards observable behaviour in admin data. When making any changes, we have internal and external reviews in place to ensure improvements to methods and data sources are researched and tested before publication. We also work hard to ensure users are well informed of any improvements.

Independent advice and input from users have been crucial to ensure the plans and benefits of innovating and improving these estimates are achieved, valued and communicated. To support this, we have refreshed our approach to engagement, including through the formation of the National Statistician’s Advisory Panel on Migration Statistics, Migration Statistics User Group, and the Methodological Assurance Review Panel International Migration subgroup. These provide us with independent advice on the development and use of our migration statistics.

Our Quarterly update on population and migration statistics provides a long-term look ahead, setting out our plans and progress. This is circulated via our regular newsletter to a wide network to ensure users are aware of any changes. We also follow each LTIM publication with a webinar.

We continue to look at how we can improve the quality and presentation of our statistics. This includes sharing more on measures of uncertainty. Uncertainty estimates for administrative data sources require novel methods, which have been developed and improved over time, making use of independent expert feedback. Our latest research update published today includes a summary of our work on uncertainty estimates.

For our upcoming release in May we plan to include more comprehensive estimates of uncertainty, incorporating more information about the sources that affect the stability of the estimates. We also aim to make uncertainty more prominent in the bulletin, helping our users understand the likely stability of the headline migration estimates.

We will continue to update OSR through our regular meetings and I would be happy to discuss any aspect of this update.

Yours Sincerely,

Mary Gregory

Related links: Office for National Statistics’ transformation of Long-Term International Migration Estimates: OSR progress report, December 2023 

Ed Humpherson to Sioned Cardew-Richardson: Assessment of statistics on higher education, further education, apprenticeships and adult community learning in Wales

Dear Sioned,

We have completed our assessment of the statistics on higher education, further education, apprenticeships and adult community learning, produced by Medr:

I am grateful for the positive contribution and engagement from your teams throughout the assessment.

We found that Medr has established a strong foundation as a new official statistics producer. We identified many areas of good practice, including collaborative engagement with users, constructive relationships with all data providers, clear quality information, and insightful statistics.

We have identified seven requirements to strengthen the statistics. These focus on publishing required statistical policies and statements, expanding quality information, improving the quality of data on Welsh speaking ability of students in higher education, being more transparent about methods where sex and gender identity data are combined, and prioritising website developments.

We judge that the statistics can be designated as accredited official statistics once we have confirmed that the requirements set out in our report have been met. Addressing the requirements will demonstrate that the statistics on higher education, further education, apprenticeships and adult community learning meet the Code of Practice for Statistics’ standards of trustworthiness, quality and value.

We have agreed that Medr should meet these requirements by March 2027 and keep us updated on progress.

I am copying this letter to Jon Ackland, Head of Further Education, Apprenticeship and Adult Learning Statistics; and Hannah Falvey, Head of Higher Education Statistics.

Yours sincerely,

Ed Humpherson
Director General for OSR

Related links:

Ed Humpherson to Sioned Cardew-Richardson: Accreditation for Medr’s statistical publications – Request for temporary suspension and request for assessment