Effective communication

Communicating effectively

Good communication is essential for statistics to be used and understood effectively. When good-quality statistics are communicated well, public good is increased, people can make better data-based decisions, and trust in the system is enhanced. We see producers communicating statistics effectively and accessibly, drawing out key insights from the data. We now want to see producers progress to better communicating the quality of the statistics – particularly with regard to uncertainty. An increasing body of evidence, including research by ESCOE, shows that communicating information about uncertainty improves the public’s understanding of it and does not reduce trust in the data.

Producers are demonstrating good practice. In November 2023, ONS published a series of international migration publications which brought together a range of materials, including information on the uncertainty in the estimates and the methodology used, to support users. This comprehensive series represented an excellent example of ONS’s communication of its migration estimates.

We have also seen good practice from the Scottish Fiscal Commission (SFC), who voluntary apply the Code of Practice for Statistics, finding that analysts make complex concepts easy to understand using visual guides and explainers. Additionally, SFC takes an open and transparent approach to explaining its methods and assumptions to both expert and non-expert users. SFC also delivers media presentations to journalists to promote correct interpretations of its work among this group, supporting accurate reporting by the media.

The publication of the Census results highlighted examples of good practice in the communication of statistics. The presentation of the results gave users unprecedented freedom and flexibility in producing their own analysis and cuts of the data. In Scotland, the National Records of Scotland promoted the transparency and understanding of the results by explaining relevant quality information to help users understand, interpret and use these important statistics.

However, for some other statistics, more could have been done to improve user understanding around uncertainty and revisions to enable appropriate use of the statistics. In Autumn 2023, ONS made larger-than-usual revisions as part of its annual Blue Book (UK National Accounts) revision cycle, which generated a range of criticism from users and the media. In November 2023, we published a review of ONS’s approach to revisions to GDP. We found that while ONS’s approach to revisions was appropriate and well managed, it could have improved its communication, for example, by enhancing the analysis of uncertainty and revisions using both qualitative and quantitative approaches and better explaining revisions during economic turning points. We welcome the quick action ONS took to improve its communication following this review. In our rapid reviews of labour market statistics and gender identity statistics, we also recommended that ONS be clearer with users about its research and development work and provided additional supporting quality information.

A gender identity question was asked, on a voluntary basis, for the first time in the 2021 Census for England and Wales conducted by ONS. This question is also the Government Statistical Service’s harmonised standard in development for collecting data on gender identity. Those whose gender identity differs from their sex represent a small proportion of the population. The concept may be unfamiliar to many people and there are few robust alternative sources of data to cross-check against. This creates challenges for data collection.

In April 2023, we commenced a review to look into concerns raised with us directly and publicly in the media. These concerns included how well the census question was understood and the effects of possible misinterpretations on the quality and usability of the gender identity data from the 2021 Census for England and Wales. In our interim report, we found that ONS should have developed its quality information more fully, including clearly highlighting areas of uncertainty in this new collection, before releasing the data to support appropriate use. We have recommended that ONS carry out further testing of the question and work with NRS on any joint learning efforts across the different censuses. We are currently working on our final report on the gender identity census data in England and Wales, which we will publish later in the summer.

Intelligent transparency and misuse

We continue to see widespread use of official statistics to inform public debate and, in most cases, statistics are appropriately used. While we continue to receive a significant amount of casework (when someone raises a concern with us, or when we identify a concern ourselves, regarding the production and use of statistics), rarely do we find a blatant misuse of statistics. Quite often, the root cause of concern is a lack of broader context about the statistics that might lead to their misinterpretation.

A lack of transparency around data and statistics can result in confusion about where numbers have come from or lead to accusations of manipulating the data – both of which are relevant to supporting public confidence and maintaining trust in statistics and those producing them.

Our overarching ambition is that when data and statistics are used publicly to inform parliaments, the media and the public, they are always published with appropriate explanation of the context and sources. To support this aim, we publicly promote the importance of taking an open, clear and accessible approach to the release and use of data and statistics – we call this ‘intelligent transparency’.

Applying the principles of intelligent transparency across government supports public confidence and maintains trust in data and statistics and those producing them. When these principles are upheld, public conversations can focus on the important issues rather than the validity and transparency of the data. This also benefits those speaking on behalf of government by reducing the need for corrections in the media and in parliaments.

In September 2023, we published an update on our intelligent transparency guidance. This provided questions that can be used to ensure transparency and new content on proportionate approaches to ad hoc releases. We also set out common misuses of statistics, particularly where misleading claims are repeated.

Heads of Profession have continued to play an increasingly important role in championing the principles of intelligent transparency across both their work and within government departments. Specifically, we have seen closer working with ministers, special advisors and policy and communication colleagues to ensure statistics are used in a way that is consistent with these principles and supports the public good. Producers have told us that the development of these principles has promoted the independence of statistics across government.

While there is now a greater understanding of the need for intelligent transparency among analysts, more work is needed to fully embed its underlying principles consistently across government, especially in relation to equality of access. We have been delivering teach-ins to analytical and communications teams within departments and are encouraged by their response and engagement. However, we are still seeing concerns raised with us via our casework process in relation to ministers and other government officials quoting unpublished figures in the public domain. This often occurs when figures are based on ad hoc management information, rather than official statistics. We aim to continue to share the principles of intelligent transparency with new audiences, including ministerial private offices and special advisors.

The UK 2024 General Election campaign period saw significant use of statistics to inform democratic debate and by the media. We encouraged producers and leaders of political parties to continue to present appropriate contextual information to limit the risk of misuse and to engage widely with those who might use the data to identify common pitfalls. Our dedicated election section on our website brought together our new and existing guidance to help users navigate statistical claims and support those working with statistics in upholding the Code of Practice for Statistics.

Producers are increasingly using online platforms, including social media and blogs, to publicly share their statistics. Social media can quickly amplify messages, which can be advantageous for reaching new audiences but also have a negative impact when the messages amplified have the potential to mislead. The structure of the internet and social media means statistics can gain a life of their own, even when the producer takes all the right steps to prevent misuse. Social media flattens information to the same level of importance – for example, a post by a producer may appear in someone’s feed alongside a video of a cat and a post from a friend. When misuse is seen among a long list of information, it makes it easier for subtly misleading claims to slip through. We have seen an increase in casework where an abbreviated claim made on social media has changed the meaning of the underlying data and led to misinterpretation. We have brought together philosophers with an interest in the use of data in online communications to discuss these challenges and have used the insights from these discussions to shape how we respond to casework for best effect.

For many who see publications of analytical information by government, the distinction between official statistics and other data may not be clear. To support public accountability, government should publish its analytical evidence wherever possible, and every time analysis is used or quoted publicly. This includes when management information (information that is used in the normal course of business to inform operational delivery or policy and management decisions but is not subject to the same quality assurance as official statistics) is quoted publicly.

Publishing more of the analytical evidence that government produces in a transparent and accessible way supports accountability, evaluation and improved outcomes. It builds public confidence in an organisation’s commitment to evidence-based decision-making and the appropriate use of data more generally. A good example of this approach is the Welsh Government’s published data on homeless people and rough sleepers, which provide valuable insights to users and decision-makers about some of the most vulnerable people in society.

We never want to inhibit public debate or discourage the use of analytical information, and we recognise that ministers and other officials get up-to-date management information about fast-moving situations. But the use of this information in public statements should be by exception rather than a regular occurrence. When data are used in this way, we expect the principles of intelligent transparency to be met, namely that the data are accessible to all and enhance public understanding, and that analytical expertise is sought for their use in public to ensure adherence to the Code.

We intervene when we see these principles are not upheld; recently, we asked the Welsh Government to improve the transparency of its journey time estimate in relation to the 20-mph speed limit roll out. When we step in, departments generally respond positively, and the analysts work with policy and communications colleagues to make the evidence available. Indeed, the Welsh Government responded quickly to our recommendation and issued further explanation for users on the evidence to support this claim.

During the 2024 election, we stepped in to promote and safeguard the use of statistics in public debate. We do not act to inhibit or police debate. However, we were willing to publicly highlight cases where campaigning parties made statements that draw on statistics and data that are not published or presented in a misleading way. This included on claims made by different political parties about opposition party spending plans.

In a statement on 6 June 2024, responding to concerns raised with us on a claim made in the ITV election debate on 4 June 2024 by the Conservative party regarding Labour party tax rises, we stated that:

“when distilling claims into a single number, the context should be sufficient to allow the average person to understand what this number means and how significant it is”.

In a letter on 24 June responding to concerns raised with us about the Labour Party’s analysis of Conservative Party commitments, we stated the following:

“Future costings are always subject to uncertainty and dependent on choice of methodology. To help people understand the assumptions that have gone into costing models, it is essential that the underlying calculations, data sources and context are provided alongside the figures. When distilling these claims into a single number, there should be enough context to allow the average person to understand what it means and how significant it is. Omitting this information can damage trust in the data and the claims that these data inform.”

As the statistics regulator our role in such claims is to determine whether they are sufficiently transparent for people to come to their own view, rather than to comment on whether the calculations themselves are reasonable.

Our recommendations