Mary Gregory blogs about the new techniques and technologies being used by OSR to enhance its reviews, starting with the media analysis of policing statistics published today.


Today we have published our media analysis on the value of policing statistics to public debate. We hope this small step is the first in a series of steps that can support bigger change, both in the way we operate in OSR, and the impact we can have on supporting statistics that serve the public good.

Today’s report represents the first time we have used web-scraping and text analysis to enhance our understanding of the use of statistics in the media. This has given us an insight into the most prevalent topics in media coverage on policing – for example mental health – as well as an understanding of where statistics are available to inform these topics and where they are lacking. We have also learnt about where official statistics publications appear to be driving the topic of debate, for example when the Crime Survey for England and Wales is published it appears to generate articles around stop and search and knife crime. We undertook the media analysis after our phase one report on Use of statistics in public discourse: the example of policing statistics was complete. Phase one was based on quantitative analysis which included talking to a range of users and influencers and reviewing statistical outputs. We found that the two approaches complemented each other, and that the media analysis undertaken in phase two generally reinforced what we had found in phase one of our work.

The new findings published today are just a small part of a wider evidence base which we hope will be valuable in underlining the need for more meaningful information on modern policing and the importance of official statistics to create informed public dialogue. We will continue to champion the production of official statistics which allow citizens and policy makers to understand the world around them, informing their decisions and enabling citizens to hold policy makers to account.

Similarly, we hope that this piece of work sows the seeds for significant developments in how we approach our regulatory work in OSR. We plan to introduce greater automation and analysis to allow us to be more proactive and draw on a broader evidence base to inform our judgements and drive our work programme. We see huge potential in what we can do with greater use of technology, for example:

  • automating some of our information gathering so that we have more consistent access to a range of sources and are better informed;
  • understanding more about how statistics are used and what individuals need from official statistics; and
  • gaining greater insight into the most important areas of public debate, including where statistics are available to inform the debate and where there are significant gaps in the information available from official statistics.

None of this will be a substitute for the more traditional approaches we take to our work such as speaking to users of statistics and key influencers to understand current issues, and undertaking detailed reviews through our systemic review and assessment programmes – which remain extremely valuable – but we are really excited to be moving forward with these new approaches which can only enhance our work and help us achieve what we really care about: statistics that serve the public good.

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