Public confidence and use of statistics

Over the past year, the way statistics are communicated, used and understood has remained an important focus of regulatory activity. This section shows that there has been some improvement in the way producers explain uncertainty, limitations and comparability, and that OSR has continued to intervene publicly where statistics have been misused or presented without sufficient context. It also shows that public involvement and engagement is now a clearer part of the Code’s expectations for official statistics, but practice is still developing across the system.  

The section also considers progress following the inaugural Statistics Assembly. Progress has been slower than users might reasonably have expected, and users need a clearer public plan for taking forward the Assembly priorities.

Supporting appropriate use of statistics

The appropriate use of statistics depends not only on the quality of the underlying data, but also on how clearly statistics are communicated and explained. The Code of Practice for Statistics makes clear that producers should support understanding of what statistics can and cannot say, including their strengths, limitations, uncertainty and comparability across the UK. These expectations are reinforced by the Standards for the Public Use of Statistics, Data and Wider Analysis, which emphasise that statistics should be used and communicated transparently, and with integrity, clarity and accuracy. 

Supporting the appropriate use of statistics was a recurring theme across OSR’s regulatory work over the last year, featuring in ten assessments and compliance reviews. Our work has highlighted the importance of clearly explaining data quality and methodology, including revisions, provisional estimates and comparability, and of giving users clear guidance on how statistics should and should not be used. This was reflected in our July 2025 confirmation of accredited official statistics status for Earnings and employment from Pay As You Earn Real Time Information, where we recognised improvements by HM Revenue & Customs (HMRC) and the Office for National Statistics (ONS) that helped users understand the coherence of labour market data and improved the accessibility of the statistics. We also noted the importance of users understanding the comparability and coherence of the wider suite of labour market statistics. 

Over the last year, we have also seen positive examples of producers improving how they explain and present statistics in response to our regulatory work. This has included work by the National Records of Scotland (NRS), where we found that the quality of the 2022 census outputs was communicated thoroughly and effectively, with appropriate attention given to communicating uncertainty. We found that NRS had taken a proactive and comprehensive approach to user engagement and in supporting users to understand and make use of the census statistics. Our work has also found that Welsh Government communicated uncertainty and methods effectively in its estimates of additional housing need statistics. 

However, our regulatory work continues to identify examples where statistics are published or used without sufficient explanation of caveats, uncertainty or comparability. This has included NHS waiting times for planned healthcare, where differences in data collection and definitions across the UK created a risk of inappropriate comparisons between nations. In our review of NHS England’s Referral to Treatment waiting times statistics, we found that users needed clearer explanation of UK comparability and NHS structural changes to help them understand the limits of comparison and interpret the statistics appropriately. Beyond these regulatory examples, migration statistics also show why clear guidance on use matters. ONS’s latest long-term international migration bulletin explains what users can and cannot do with the statistics, including how to use the newest provisional estimates. Cases such as these show that appropriate use depends not only on the availability of statistics, but also on whether users are given enough explanation to interpret them correctly. 

While there has been some improvement over the past year, practice remains inconsistent. Our regulatory work continues to identify cases where clearer explanation of quality, uncertainty and comparability is needed.

Challenging the misuse of statistics

A core part of OSR’s role is investigating concerns about the production and use of statistics through casework. Over the past year, we have made public interventions where statistics were used without enough explanation or context or had the potential to mislead. The cases below are examples chosen to illustrate the different ways in which misuse and misunderstanding of statistics can arise in high-profile public debate. 

• Presentation of modelling: In our intervention on the presentation of modelling in the UK Covid-19 Inquiry, we raised concerns that the way a modelled estimate was presented risked over-interpretation without sufficient caveats. This highlighted the importance of communicating uncertainty clearly, particularly where modelled figures may be given undue weight in public debate. 

• Choice of measure: In relation to the Welsh Government’s apprenticeships target, we highlighted the need to be transparent about which measure was being used and why, particularly where more than one possible measure was available and previous reporting had used a different basis.  

• Lack of context: In our work on claims about changes to business rates announced in the November 2025 Budget, we found that fuller context was needed to support understanding, including clarity about whether statements referred to changes in the tax rates specifically (for example, the multiplier) or to business rates more broadly. 

• Unsupported comparisons: Misuse of statistics also featured in recent Scottish Parliament and Welsh Senedd election debates. High-profile claims comparing NHS planned care waiting times across the UK demonstrated how differences in definitions, coverage and limitations can be misunderstood or ignored in public debate. In response, we published a public statement setting out how these statistics should be interpreted and the risks of misleading comparisons. We also investigated similar claims in January and March 2026. 

These are all issues that the Standards for the Public Use of Statistics, Data and Wider Analysis are intended to address. 

Public involvement and engagement

The revised Code places greater emphasis on how producers engage with users and the wider public. It requires producers to put users at the centre of decisions about statistics and to consider evidence on the degree to which the collection and use of data for sensitive topics are viewed as acceptable by society. It also requires producers to publish a public involvement and engagement strategy. 

This is an important development for the statistical system. The Code is clear that official statistics should serve the public good. Public involvement and engagement supports this by helping producers understand the views of people who may or may not be direct users of official statistics, but who should have a say in what statistics are produced and how that is done. 

This is not the same as user engagement. User engagement helps producers understand what people who consume statistics need from them, including their data needs and their views about the published statistics. Public involvement and engagement may be part of a wider user engagement strategy, but it also goes beyond this to understand views of non-users as well, for example around public acceptability. This includes people whose data are used, people who are represented in statistics and people affected by decisions made on the basis of statistics. 

The toolkit for public involvement and engagement in official statistics, developed by Connected by Data in conjunction with OSR, provides practical support for producers in meeting these expectations. It sets out why, when and how producers can reach members of the public in a proportionate way. It is clear that it is not necessary to involve or engage the public in every decision and that the level of involvement or engagement should be guided by factors including the scale of the statistical work, the use of the statistics, the level of sensitivity attached to the statistics and the public interest in the statistics.  

During the year, we have seen examples of work that can support producers in understanding wider public attitudes to data. ONS published the outcomes of its ONS Data Debate: 65 plus and an updated summary of people’s attitudes to data. This work provides evidence that other producers and organisations can draw on, particularly where statistics involve sensitive topics or substantial change. Producers can also draw on system-level work when considering how people are represented in statistics and the language used in statistical surveys or outputs. The Government Statistics Service (GSS) Harmonisation Team’s work plan outlines further development of standards and guidance on ethnicity, disability, sex and gender identity. The National Statistician’s Inclusive Data Advisory Committee provides independent advice to support the ongoing improvement of the inclusivity of data and evidence across the UK statistical system. 

Public involvement and engagement does not replace user engagement. Our regulatory work over the last year has shown positive examples of producers strengthening user engagement, which remains central to the Code expectation to put users at the centre of decisions about statistics. In our assessment of statistics on higher education, further education, apprenticeships and adult community learning in Wales, we found that Medr was putting users at the centre of decision making and maintaining a dialogue with users, which helps ensure that the statistics continue to be relevant. In confirming accredited official statistics status for Northern Ireland Business Expenditure on Research and Development statistics, produced by the Northern Ireland Statistics and Research Agency (NISRA), we also noted that NISRA had conducted a user consultation to better understand user needs, revised its quality report in response and planned to publish a back series because users said it would be useful. 

At the same time practice is not yet consistent across the system. The Royal Statistical Society (RSS) – OSR roundtable on user engagement and the Code of Practice, held in February 2026, found that users often remain uncertain about whether their input is heard or acted on. It highlighted the importance of clearer feedback loops, more consistent approaches to engagement and more transparent routes into dialogue with producers.  

Public involvement and engagement is now a clearer part of the Code’s expectations for official statistics, and part of how producers show that statistics serve the public good. The system now needs to strengthen practice in a proportionate way, recognising the practical challenges producers face in deciding when and how to involve the public. Continued collaboration and sharing of knowledge will be important in supporting this. 

Reviving momentum following the Statistics Assembly

The inaugural UK Statistics Assembly, held in January 2025, provided useful insights into the current demand for data and statistics across the UK. It brought together a wide range of stakeholders and, through the independent report published in March 2025, identified four high-level priorities for the UK Statistics Authority and the Government Statistical Service: reinvigorating user engagement, improving more-granular statistics, scaling up the use of administrative data and strengthening UK-wide coherence. 

There is evidence of some progress in areas connected to those priorities. This includes the extensive topic consultations conducted by the three census offices ahead of the 2031 censuses, the July 2025 Family Finance Surveys User Conference, and the activity described in the UK Statistics Authority’s December 2025 update on granular statistics, administrative data and UK-wide coherence. More recently, the UK Government’s March 2026 response to recommendation 5 of the Lievesley Review has also set out further commitments on comparable UK-wide data and a strengthened strategic role for the Inter-Administration Committee (IAC). 

However, progress against these priorities has been slower than users might reasonably have expected, and it remains important that momentum is maintained. Without reinvigoration, the Statistics Assembly has the potential to become a significant missed opportunity. 

The Assembly generated a wide range of recommendations, not all of which can be taken forward at the same pace. Delivery will need to be prioritised alongside other pressures facing ONS, the GSS and producers across the system. The Authority’s December 2025 update stated that it had “not progressed development of the refreshed Authority user engagement strategy as quickly as we would have liked.” The same update said that activity would be reviewed again following the appointment of a permanent National Statistician in early 2026. In January 2026, OSR similarly warned that there was a risk of losing momentum from the Assembly and said that renewed momentum was needed to take its priorities forward. 

The Assembly was not intended to be a one-off event. It identified priorities that require co-ordinated effort across the statistical system. Users need a clearer public plan for taking forward the Assembly priorities. Without that clarity, the Assembly risks losing some of the value it created.