Ed Humpherson blogs on the key theme that came out of this year’s GSS Awards.
At their best, statisticians in Government do great things. They take complex data sets and bring out the meaningful insights for the public. They ensure that their organisations can communicate statistics in a trustworthy way. They have the courage to defend their statistics from inappropriate use by non-statistical colleagues. And they innovate and collaborate.
At OSR, we are concerned that these good things are not always more widely visible – even to statisticians in other parts of Government. Good innovations in methods or communications can sometimes take a long time to be spread more widely. That’s why the work of the Best Practice and Impact team is so important (find out more here).
And it’s also the role of the GSS awards, which formed one of the central moments in the recent GSS conference in Edinburgh. David Mais of the Best Practice and Impact team has done a really nice blog summarising the prize winners.
I want to build on David’s blog and bring out one key feature of all the winners: innovation, doing things differently, pushing the boundaries of what statistics in Government do. There wasn’t a prize for innovation – it just came up in all of the winners.
The collaboration award was won by the Reproducible Analytical Pipeline (RAP) champions network who are promoting the use of automated tools to prepare statistical outputs – taking out both the drudgery and error-proneness of a traditional cut-and-paste approach to compiling outputs. It’s an example of using automation for really tangible benefit.
Did I mention automation? This was the main feature of the winner for Communication: the ONS’s work the rise of automation in the economy – and including its own automated chatbot to answer users’ questions. Again, innovation to the fore.
The impact award came from the Office of Rail and Road, and their innovative work to build new insights into the experience of disabled people who travel by rail – and using that data to drive changes in rail policy and delivery.
And my favourite innovation came from the Methods award. It went to the Department for Transport (DfT) in collaboration with the ONS’s methodology experts for their work on road accidents statistics. It’s an age-old problem. How do you report statistics at a national level where they are produced from different local-level systems? The traditional answer is to add some text at the back of the report about the limitations of the administrative data- words like “users should interpret the data with caution” may appear. (As a colleague once said: “What does that mean? That I should be sitting in a cushioned room before I read the statistics?”)
The DfT team went far beyond this wordiness. They took the available data and knowledge about how different police recording systems operated; built a model; and made an adjustment to the figures to show what they might be if they were all produced by the same system. They publish the statistics as experimental statistics, indicating clearly to users that they are still in development; and their work aligns well with the Code of Practice’s expectation of innovation and improvement.
It’s worth saying that, as experimental statistics, these new estimates will be subject to further refinement and development, and DfT may well get user feedback that leads them to amend their approach. The key point for us, though, is not that these are perfect: it’s that there’s a willingness to experiment and develop a methodological approach to an age-old problem.
For me this is brilliant. It should be considered by lots of other statistics producers: taking data and using them to make a series of inferences and helping users make the best use of the data. It’s not easy; but it is ambitious, and it’s driven by a desire to make the statistics as useful as possible for users.
The Best Practice & Impact team has organised a sharing seminar where you can find out more about the work of the award winners. This will be held on 2nd December, and you can sign up here.
So, there it is: a celebration of the innovation within the GSS. If only we could make sure these innovations spread rapidly and fully…