Da** Statistics!
Craig Dalzell - 14th October 2021
Back in 2019, I wrote an article to try to draw attention to an impact of Brexit that I hadn’t seen anyone talk about and I was worried that it would pass unnoticed. The way the exit agreement had been structured, the UK faced the possibility of dropping out of the Eurostat statistics database at the end of the post-Brexit Transition Period.
This was important because it’s not enough for a single country to measure its own statistics, those statistics must be comparable to other countries or we can’t judge ourselves as doing better or worse than them in areas of interest and can’t look to others for ideas on how to fix our shortcomings.
It’s not just exclusion from the database itself that worried me but also the agreement between EU countries to measure statistics based on a common framework. This is also critical to that kind of cross-comparison work. Imagine, for example, one country defined “unemployment” as “anyone who doesn’t do paid work for at least one hour per week” but another defined it as “Only those who actively sign on to the Government Unemployment Programme and stay on it for six months”. The two countries could end up with very different measurements of “unemployment” despite having similar numbers of people not working.
I am very aware of how niche an issue this kind of worry might be outwith statistics geek circles but I ended up striking a nerve within those circles and was inundated with emails from folk (including quite senior stats researchers) who thought they, too, were the only one who spotted this coming problem. The article led to a series of questions being put to the UK Government resulting in the Department for Exiting the EU promising that continued participation in Eurostat would be negotiated after the Transition Period ended in December 2019 and assurances from the UK Statistics Authority that UK stats would continue to adhere to international best practice no matter what happened after Brexit.
We’re now almost two years past Brexit Day and the promise has yet to materialise. As far as I can tell, negotiations to re-enter Eurostat haven’t even begun, the Government Department that dealt with the question has been closed and the Junior Minister who made the promise has been reshuffled into the Transport Ministry. With all of the focus on the UK trying to deflect blame for its own mishandling of responsibility with regards to the Northern Ireland Protocol, I can’t see any hint that anyone in the UK Government considers the UK’s growing Data Desert to be a problem, much less anyone who has the will to fix it. Since January 2021, Eurostat has ceased publication of most UK statistics. It is now much harder for statisticians to work out how policy outcomes compare between here and neighbouring countries. The UK has been relegated to being a featureless grey box between Ireland and the North Sea.
(Source: Eurostat)
There could have been a solution in Scotland at least. In 2019, SNP conference delegates overwhelmingly supported a motion calling for the creation of a Scottish Statistics Agency which would aim to fill our own gaps in statistics provision. This agency could also act to ensure continued cross-comparison with Eurostat and other international statistical databases. Not necessarily by feeding Scottish data into these databases but by doing the opposite. By acting as a portal where those data can be shown alongside comparable Scottish numbers, we could see at a glance how we’re measuring up compared to our peers and start discussing how to correct shortfalls. Unfortunately, despite that overwhelming support, the Scottish Government has yet to pull this policy from their party library and bring it into the Programme for Government, nor to my knowledge has it even acknowledged that the policy was passed at all. Evidently, more lobbying has to be done here and if you think this would be valuable please write to your MSP and tell them that you want this to happen.
Of course, once we have the statistics in place we also have a duty to use them properly. The UK Statistics Authority recently called out both the UK and Scottish Governments for the misleading use of statistics or for the mishandling of statistics in public statements. In their paper on lessons learned from the Covid crisis, they stressed that such misuse can damage public trust in statistics and therefore damage the ability to use said statistics to counter misinformation or even to make sound policy that the public can believe in.
Even when politicians aren’t saying anything strictly untrue when commenting on statistics, there is a duty of care around how their statements should be framed. This is, of course, somewhat more subjective and every spokesperson has a natural tendency to try to spin as much gold out of a bad story as they can. Take, for example, the Scottish Justice Secretary’s recent statement about reoffending rates in Scotland. “One of the lowest rates in 20 years” is a perfectly true statement – for a given value of “one of” – but when we look at the actual data, it’s not necessarily the most obviously line to draw. “Largest single year increase in reoffending rates in 20 years” is another perfectly true statement to draw from the data as would be “a sudden reversal of a 20 year downwards trend” or “the second highest rate since Nicola Sturgeon became First Minister”.
I’m well aware that complaining about statistics is about as niche an area as it gets but without data it is impossible to make policy that accurately reflects the real world and, perhaps more importantly, it is impossible to measure policies to see if they’re actually having the intended effect.
And we also need our politicians to be ambassadors of the truth when they are talking about the stories that statistics are telling them – both when it is good news and when it is bad. As the UK Stats Authority said in their report, statistics work best when they are produced and communicated in a way that is open, transparent, collaborative and communicative. We could do with a few more folk in power trying a bit harder to raise statistics up rather than allowing them to be misused while the lies and damned lies that are all too common in the world are far too readily believed.