1. Making public statements about the dissemination and use of statistics has been an important part of the Authority’s work since it began.
2. The Authority is guided in this work by its overarching statutory objective to safeguard the production and dissemination of statistics. Section 8 of the Statistics and Registration Service Act 2007 provides for the Authority to monitor the production and publication of official statistics, and to report any concerns it may have.
3. The Board keeps its practices under regular review. At its October 2017 meeting, the Board concluded that the practices themselves remained appropriate, and that it would be sensible to make them more explicit publicly.
4. The Authority recognises that it is not an arbiter of political debate, and that each example of the use of statistics takes place in a different context.
5. We are guided by a series of long-standing practices. The Authority will intervene if:
i. official statistics are leaked before publication;
ii. the advice of professional statisticians is ignored; or
iii. official statistics in a document or statement are presented in such a way that, in the Authority’s opinion, they are liable to mislead the public or undermine the integrity of official statistics.
6. In considering whether documents or statements are liable to mislead, we consider a range of factors. We take into account:
i. the standing of the person responsible for the misleading statement (for example, whether that person, by virtue of holding a public office would be expected to make authoritative statistical statements);
ii. the context in which the statement is made (for example, distinguishing speeches and publications from off-the-cuff remarks in interviews or debate);
iii. the extent to which the statement might be misleading – for example, if a third party who relied on the statistical statement might reach an incorrect or unreliable conclusion – with any evidence that they are being misled;
iv. the extent to which the statement might undermine public confidence in the integrity of official statistics; and
v. whether this is the first time the issue has arisen, or whether it is a repeated issue.
Scope: official statistics and wider information
7. While it is the case that our formal statutory remit concerns “official statistics”, we take a broad view of what falls within our scope. Under the Statistics and Registration Service Act 2007, it is for each producer to define whether a publication is official statistics, and this results in a diversity of practice – for example that published data on the funding of police forces are not official statistics. Moreover, for a member of the public or Parliament who observes the publication of numerical information, the distinction between “official statistics” and other forms of data may well seem artificial. As a result, we will often state in our responses that, even though data are not defined as official statistics by the publisher, we will consider the case on the basis of the principles of the Code of Practice.
8. It is also important to be clear that official statistics generally relate to descriptions of a current state of the world. For economic forecasts, we recognise that there are a wide range of potential assumptions and methodologies, so that different forecasters can arrive at quite different conclusions, and therefore support different claims about, for example, the economic impact of leaving the European Union.
Nevertheless, we typically would restate our basic expectations around transparency of methodology and assumptions. We have also taken the view that published annual accounts of Government entities, which are prepared under International Financial Reporting Standards and are audited by the National Audit Office, should not fall under the “wider application” approach – because there are already sufficiently rigorous arrangements to establish the reliability of these types of statement. However, there may be a case for considering the unaudited performance information which accompanies the financial statements as official statistics.
Sources of cases
9. In terms of the sources of the cases, matters come to our attention in two ways. First, our own analysis may identify a concern. We will consider whether the issue is best resolved by a public statement from us, or by representations directly with the statistical producer.
10. Second, we consider all complaints that come to us from the public, elected representatives or organisations. This does not mean we always provide a substantive response. In some cases, we will conclude that the question raised is not about the production or use of official statistics; or that it is more appropriate that the statistical producer responds directly. We also form a judgement, based on the materiality of the issue, as to whether to place any response we make on our website. Where we do conclude that our correspondence with a member of the public should go on our website,
we always seek their permission before doing so. If the complaint comes from an organisation (for example, a business or a political party) we would typically place our response on our website (especially where the letter raising the issue with us has already been published by the complainant).
Discussion with producers
11. We usually communicate with the organisation that produces the relevant official statistics or data to get their perspective on the issues raised in the case. This involves contacting the Head of Profession/Chief Statistician. Where the Head of Profession recognises that the producer has made a mistake or undertakes to clarify promptly an aspect of the statistics, we can decide not to write publicly, or to simply welcome their actions. Where the issue relates not to the publication and dissemination of statistics, but to how they are used by others in political discourse, contacting the Head of Profession for a briefing is not necessary. But in all cases we aim to share any letter shortly in advance of our publication – not for clearance but to make the producer or person who made the statement aware of our thinking and intention to publish.
Responsibility for Authority responses
12. Where the issue relates to the production of statistics and their dissemination by a Government department, this is most often dealt with by the Director General for Regulation, who heads the Office for Statistics Regulation.
13. Where the issue relates to broader political use, especially by Ministers and elected representatives, the Authority Chair will respond.
14. Any letter we wish to make public is placed in the “correspondence” part of our website. Very significant letters will also be highlighted on the front page of the website.
15. Where a response remains private, we will nevertheless record it on our published Issues log, which summarises anonymously all the issues brought to our attention.
16. Where our intervention involves recommending an action by a producer of statistics, the Office for Statistics Regulation will follow up to monitor implementation. Our presumption is that we will make our view or judgement clear in our initial published statement or private correspondence. As a result, we do not normally engage in an ongoing correspondence on a specific issue, other than to clarify an aspect of our thinking.
UK Statistics Authority
18 October 2017