By Ed Humpherson, head of the UK’s Office for Statistics Regulation¹. Ed is an EDI affiliate. This blog reflects his flash talk at the 2025 EDI Town Hall “Shaping the Future of Data Ethics” on 31 March 2025.

The regulation of official statistics may sound like a dry and technical subject. But in fact it is fundamentally an ethical endeavour.

Official statistics form part of the fabric of public life. They frame our understanding of the economy, prices, crime, health and education. They are a source for people who want to get insight into the world – whether they are campaigners, community groups, companies, charities or citizens (the five Cs, you might call them).

It follows from the prevalence of official statistics in public life that they are fundamentally social. They describe our world, and how we are represented as citizens and members of places and demographic communities.

It also means that two crucial social aspects of official statistics must be taken into account by my organisation, the Office for Statistics Regulation, when we consider the production and use of official statistics in the UK.

First, user engagement is crucial to the development of official statistics. It’s essential that professional statisticians engage with and ascertain both the questions that people are interested in (for example, what demographic breakdowns are important) and the ways in which they want to access statistics (for example, through summary tables, detailed data downloads, or APIs).

Second, how statistics are used in political discourse is also important. It is damaging to have statistics on a subject –  say migration – that are carefully produced by professional statisticians but, once they go out into the world, are used in a distorted and misleading way.

This brings me on to my Ethical Data Initiative: intelligent transparency.

Intelligent transparency arose in my work during the Covid-19 pandemic. We noted how prominent data about the pandemic was in public consciousness in the UK: infection rates, hospitalisation numbers, vaccine take-up percentages – these numbers were reported and scrutinised daily. But we also noted that there were times when Government sought to serve this appetite for data by announcing high level aggregates without making clear the source of the underlying numbers. This meant it was hard for citizens to understand or verify the numbers, even though the high level findings were transparent, in the sense of being quoted publicly. This led us to distinguish the availability of single high level numbers (transparency) from providing full, detailed, accessible underlying data (intelligent transparency).

From this starting point, we have built a set of principles of intelligent transparency. These principles argue that dumping individual factoids into the public domain (naked numbers, we sometimes call them) is inadequate. What is required is clarity on the sources and findability of the numbers; and recognition of what they mean and their limitations. In other words, intelligent transparency requires:

• Clear sources that are in the public domain
• Easy accessibility to the underlying data
• Respect for any professional caveats and limitations

We have had a lot of success in rolling out these principles across the UK government. For example during the UK General Election, we received several concerns regarding a claim made in the ITV election debate on 4 June 2024 by the Conservative Party that a Labour government would mean £2,000 of tax rises per working household. The claim was also used numerous times in party political broadcasts, campaign speeches, social media and the televised debate. We concluded that the claim, when taken in isolation, was missing important context, for example that £2,000 is an estimation summed together over four years.

To ensure our judgement reached a wide audience, we engaged with the media and published a news statement. This intervention received wide media coverage and also enabled journalists to call out the use of the statement during later interviews with candidates.

Another, non-election example relates to a claim made by the Prime Minister, Kier Stamer at the Labour Party Conference in September 2024. The claim in question was that there has been “a 23 per cent increase in returns of people who have no right to be here, compared with last summer”. At the time the Prime Minister made this claim there was no Home Office data or statistics available in the public domain for the relevant time period to support this statement.

Following the statement made by the Prime Minister, we engaged with the Home Office and requested that the supporting data were published via an ad-hoc statistical release. This ensured that this claim could be independently verified by the public. Whilst we didn’t write publicly on this issue, we did publish a related blog. 

In short, then, my Ethical Data Initiative is this: when putting evidence to the public, the underlying characteristics of the evidence should be respected – in line with our principles of intelligent transparency.

**

1. The Office for Statistics Regulation (OSR) provides independent regulation of all official statistics produced in the UK, and aim to enhance public confidence in the trustworthiness, quality and value of statistics produced by government. For more information, see https://osr.statisticsauthority.gov.uk ↩︎