Ed Humpherson reflects on the last year, and how OSR will continue to uphold the philosophy of trustworthiness, quality and value going forward.

Statistics are not just for official decision makers like Ministers and the Bank of England. They are a public asset.

And public confidence in statistics is more important than it’s ever been. We are increasingly in a world of data abundance, where the very widespread presence of data in society and our lives threatens to overwhelm us, leading to a loss of confidence in all statistics. It’s our job to ensure that people who produce statistics, and people who use them, act in ways that uphold that confidence.

This overarching vision of public confidence comes through two key OSR publications in the last week.

We published our Regulatory Vision as an exposure draft. The background is that, frankly, we’ve been better at promoting trustworthiness, quality and value (TQV) in others than we have been at explaining ourselves. This has led a range of organisations to adopt the TQV framework on a voluntary basis. Last week, for example, the Financial Conduct Authority adopted it for a major survey. Our relative weakness in explaining our role struck me when I was giving evidence in April to the Public Administration and Constitutional Affairs Select Committee. I found that I was explaining our purpose in ways that were consistent with everything we do, but which we hadn’t captured into a single, simple statement.

So, the exposure draft aims to explain what we do and how. It starts by emphasising that we are not ONS: we have a clearly separate identity, and we do not produce any statistics. The exposure draft then describes how we deliver our vision of public confidence in an age of data abundance:

  • Upholding TQV through a range of regulatory activities, including the National Statistics designation
  • Protecting the role of statistics in public debate, through our interventions when we see materially misleading or incorrect uses of statistics
  • Collaborating to develop a much better understanding of the public good role of statistics, so that there’s a rich evidence base demonstrating the role of statistics as a public asset

The exposure draft is designed to get comment and suggestions, so please let us know what you think.

We also published today a standalone Annual Review of 2018/19, which sits alongside the UK Statistics Authority’s Annual Report and Accounts. I was keen that we had a more clearly separate statement than we’ve had in the past. I wanted to highlight our role, and also bring out systemic findings – themes which crop up across our work.

Looking across our work we see progress in a lot of areas: ONS’s development of better regional statistics; the beginnings of a more coherent approach to health statistics in England; and an appetite for innovation across much of the Government statistical landscape – including strong evidence of data linkage from producers in Scotland, Northern Ireland and Wales. But there is much, much more for all producers to do, including further improvements in statistics on health and care; housing; and earnings. We highlight a missed opportunity in the ONS’s take up of the new Code of Practice.

And we emphasise a need to focus on public value. For example, our November 2018 report on accident and emergency waiting times statistics argued that the value of the statistics was undermined because the public purpose of the statistics was unclear; and in February 2019 we published a summary of social care statistics which highlighted how these statistics are not meeting all user needs in any of the four countries of the UK. This theme of public value reflects a need for many statistics producers to move on from a mechanical “collect and count” approach and think about the questions that really matter to users.

One final point about the Annual Review. The Review includes some basic figures on OSR – including our budget and the number of outputs we produced. That means that we are putting new figures in the public domain, but which are not official statistics. We want to reassure users about the trustworthiness, quality and value of these figures. So, we have set out in a short table how we have achieved TQV in relation to the Review: another example of adopting the Code even where the outputs are not official statistics. It shows that we challenge ourselves to practice what we preach.


Please help us improve our website by answering a few quick questions.