There have been several consequential reviews of the role and performance of the UK statistics system over the last few years, each having great impact. To my mind, there are four key ones:
- Densie Lievesley’s review of the UK Statistics Authority (UKSA)
- Sir Robert Devereux’s review of the Office for National Statistics (ONS)’s performance and culture
- Professor Alice Sullivan’s review of sex and gender data
- The Public Administration and Constitutional Affairs Committee (PACAC)’s report on the UK’s evidence base
These reports all make statements early on about the importance of statistics. Can you match the reviews with the statements below?
A – Lievesley Review
B – Devereux Review
C – Sullivan Review
D – PACAC 2024 review ‘Transforming the UK’s evidence base’
“Without accurate and trusted statistics, we move closer to a post-truth world of ‘alternative facts’… The UK has a long tradition of systematic social investigation, originating in the desire for information regarding population trends following the industrial revolution, and spurred on in part by the aim of understanding the effects of the great plague in London. This produced the empirical social research tradition of ‘political arithmetic’ which seeks to provide reliable data to inform public policy and debate.”
“I am honoured to lead this Review because of my long held belief in the critical importance of official statistics which are fundamental for evidence-based policy and decision making. Statistics we can trust are essential for a healthy society as they help to ensure well-informed decisions by putting the best available evidence at the heart of policy development and implementation. They also enlighten through making explicit what is known through scientific evidence and, importantly, what is not known.”
“Good evidence empowers us all. In government, good evidence can help Ministers in delivering their commitments to the electorate, in strengthening critical public services, and in growing the UK economy. Citizens rely on robust official evidence when they cast their votes, and when they make decisions about the schools their children attend or where they choose to live.”
“…statistics guide critical decisions of Government, independent institutions like the Bank of England and Office for Budget Responsibility, private sector companies and individuals.”
(The answers are: 1. C, 2. A, 3. D, 4. B)
At one level, this is just a bit of fun. It shows that people who do reviews say more or less the same things in their opening remarks; they clear their throats, so to speak, in a consistent way.
But at another, it reflects a long-standing hunch of mine – that these opening statements about the importance or value of statistics, while superficially similar, give nuanced, but vital, insight about each report that conditions everything to follow. I’ve been thinking about this because, as we at the Office for Statistics Regulation (OSR) implement our new strategy for 2026–2029, it’s important to be clear about the value of official statistics.
Reading between the lines
Take the fourth statement above. This comes from the summary of Sir Robert Devereux’s review of economic statistics, published in June 2025.
The Devereux Review has been hugely consequential. Alongside OSR’s own review of economic statistics, which we published a couple months prior to this report, the Devereux Review was a key driver of the ONS’s significant reset in the second half of 2025. The review highlighted the need to change the ONS’s organisational culture, prioritisation and leadership, and has underpinned the changes led by Darren Tierney, the ONS’s new Permanent Secretary.
The Devereux Review focuses on the role that economic statistics play – and the ordering of the actors in the quote is instructive: it starts with critical decisions of government and independent institutions like the Bank of England and the Office for Budget Responsibility. Other actors – private sector companies and individuals – are not exactly sidelined, but they do come at the end of this list of weighty and named institutions. Economic decisions matter to the Devereux Review.
By contrast, the 2024 PACAC report on the UK’s evidence base (the third quote) is more explicitly universal. It says “Good evidence empowers us all”. It then gives separate sentences to government use and citizen use, and enumerates the types of decision that are influenced by evidence in each case. Having reliable evidence for use by a wide range of actors matters to the 2024 PACAC review.
The Sullivan Review has a very specific focus – the collection and publication of data on sex. But the opening remarks are anything but narrow. They take an expansive historical sweep, highlighting the social and political role of statistics as far back as the great plague in London and the industrial revolution. These are not merely interesting historical observations. They are used by the review as the basis for its analysis of the long-standing prominence of recording the sex of the population in demographic and social analysis. The historical sweep matters to the Sullivan Review.
The Lievesley Review’s introduction, again, subtly but crucially differs from the others. Its focus is not just on the role that statistics play in supporting “a healthy society”, but in specifying the role of understanding of statistical uncertainty: statistics “enlighten through making explicit what is known through scientific evidence and, importantly, what is not known.” In this sense, the Lievesley Review conveys most strongly that we need statisticians as embodied actors who bring their expertise in data and their uncertainties to benefit society. The professional role of statisticians matters to the Lievesley Review.
First words and the value of statistics
I think it is easy to disregard these opening statements, to want to skip over them to the meat of the reviews, that is, the findings and recommendations.
But in my view that would be a mistake. These convey important information – and in what follows I’m grateful to Ken Roy, an independent researcher into official statistics, who has provided a lot of these insights to me:
- The introductions reveal what the review is going to focus on. For example, the role of statistics for economic decision makers (Devereux) is a very different lens to that of the long-running role of social science measurement of the population (Sullivan). Indeed, the reviews fall into two categories: those which are about a perceived problem (Devereux, Sullivan) and those that are more holistic (PACAC, Lievesley).
- They are also revealing of the breadth of the role of official statistics. As Ken has pointed out, there is no set definition of the purpose and role of official statistics, and debates about statistics tend to cleave into two distinct purposes: a policy-making, decision-supporting purpose on the one hand; and a purpose to support democratic insight/public understanding on the other. This breadth means that the introductions to these reviews need to start by setting out which of those perspectives they are going to focus on. They can’t draw on a commonly agreed basis. So they need to set it out.
- The PACAC and Lievesley reviews tell us about the importance of statistics and statisticians to society as a whole: whether that is as components of a healthy society, which is honest about the limits of knowledge (Lievesley), or as evidence to support choice by a wide range of actors (PACAC). But even then, they do not explicitly recognise the full range of potential non-government users. Ken set these out in a blog he wrote for OSR a few years ago. He characterised the range of potential outcomes for official statistics (as expressed in visions, missions and so on) as:
-
- knowledgeable, represented, empowered citizens
- healthy public debate
- economic success
- effective government (demonstrating good governance)
This breadth of the purpose is reflected in part in the UK’s Statistics and Registration Service Act (2007). It focuses on statistics that serve the public good, and gives a two-pronged summary of what that means: both informing the public about social and economic matters, and assisting in the development and evaluation of public policy – and in that order, too: Informing the public comes first.
But this section in the Act is brief, and leaves a lot to be developed. To explore this idea further, at OSR we have developed a body of work on understanding the public good, from both the conceptual perspective and also through applied research into what people actually want and use statistics for. The overall conclusion of our work exploring how statistics serve the public good is that we should consider the widest possible public as the audience of statistics – and not to confine the focus to policy makers.
Final words
This all leads me to two conclusions:
Firstly, the next time you see a big review, don’t glaze over the first paragraphs. They say something important about what will follow.
Secondly, the question of the purpose of official statistics is far from settled. And perhaps it never will be – maybe these different uses will never be put into an established hierarchy. Which means that those of us who believe in the public value of statistics need to continue to advocate our case whenever the purpose or performance of the official statistics system is questioned.
And as we implement our new strategy, we will continue to emphasise the public purpose and public value of official statistics.