The Office for National Statistics does surveys “because it gives people the chance to have a voice”. Oh no it isn’t, it’s about collecting and publishing statistics about the nation. This was just one noteworthy comment in a weird interview with the National Statistician yesterday. That said, the “giving a voice” comment might help to explain why there’s been a mess up with the sex question in the Census, where a judge stepped in last week to get the ONS to correct its ways. Diamond also got into forecasting a third wave of Covid-19 in the autumn: I really think that “statistics” and the ONS should be about the past, not forecasts (especially since the ONS does not do epidemiological models). The National Statistician might also have quoted data* that’s not in the public domain – breaking his own code of practice! And all before breakfast on a Sunday!
Just occasionally you get an insight into what is happening in bureaucracies and this chink of light from Sir Ian Diamond, National Statistician, on The Andrew Marr Show yesterday was one such moment. I am sure that the ridiculous idea that surveys are about “giving people the chance to have a voice” is linked to cultural change in the ONS. This in turn has led to the organisation’s policy capture by Stonewall and others. It’s just a small step then to mess up the Census.
Last week the ONS was told by a judge that it had to withdraw its advice on how to complete the Census form. This happened in the month of the Census after years of planning, and after the guidance had been published – and many forms filled in early! The trans lobby had captured the ONS to the extent that the Census guidance was encouraging the female/male sex question to be answered with respect to self identification rather than according to biology. The sad story can be read about in many places: Daily Mail, Guardian and BBC.
In the interview Marr got little clarification. Diamond could not accept that anything was wrong, only adding that ONS “will accede to the judges request”. I can find nothing on the ONS website to clarify this. The top search return on the UKSA site was the paper explaining the now discredited policy.
Marr started the interview with a question about the underreporting of crimes against women. Diamond was keen to point out that there was “a very large underreporting both of rape and assault in which includes penetration. Indeed …… only about a third are reported to the police and then very many few of those go to prosecution.” Adding: “One of the reasons why we think is incredibly important to do these crime surveys because it gives people a chance to have a voice.”
The “have a voice” line sounded very scripted and is not one I can recall hearing from ONS or other similar bodies before. The public was presumably meant to conclude that the nation’s statisticians really do care! If the ONS aims now to “give a voice” rather than collect decent data, we are all doomed.
It’s a shame Marr didn’t ask about data quality as we’re probably not getting half decent numbers. You would not realise from the interview that the ONS, Police and Home Office had been accused of multiple failings with regard to crime statistics in many reports over many years. (I note especially one that I worked on as an adviser to PASC, in 2014.) One core problem is that “Statistics on Crime in England and Wales” are still de-designated from the National Statistics status, in other words they are deemed not to be reliable. This includes a publication called “Focus on Violent Crime and Sexual Offences (Violent Crime)”. The right question to Diamond might have been about the actions that have taken place since de-designation seven years ago. And, specifically, what numbers on “Violent Crime and Sexual Offences” can we trust.
And what has OSR, the statistics regulator said on the topic? Not much. (I typed “crime” into their search bar.) There was a blog in 2018 and a wrist slap letter in 2017. I can see no comment on the impact of reduced sample sizes and declining response rates which might affect the data quality. Not to mention whether any non-expert who happened to be interested in the sort of issues that have been debated in the last week would gain any real insight by going on the ONS web pages. Easy to understand data about women as victims is very hard to find on the ONS site.
Next there’s the question as to what data Diamond was referring to when he was discussing the pick up in economic activity after the weakness of January as shown in the official data. He said: “…. that from our business impact of coronavirus survey and also from data we have on lorry flows some of the data from the end of January and beginning of February are starting to see a pick up.” I can’t see the data that lets me easily draw this conclusion from the “Business insights and impact on the UK economy” data set. In contrast, the Vehicle flows around ports data seems only to be monthly and indeed the Jansury seems to be higher than November and December which is all a bit counterintuitive.
I have written to OSR to ask where I can see that data*. (* I got a reply – see below.) It’s odd as OSR wote to the Cabinet Office on 5 March querying the use of data in a press release. I do hope that Diamond was not referring to data the use of which had been frowned upon by his UKSA colleagues.
Finally Diamond dabbled in a bit of “Covid-19 wave and peak ” forecasting. We all do it but I really don’t think the National Statistician should be making such announcements in public. The quote prompted stories in plenty of media outlets. For example, The Metro had an article with the opening line: “The UK is ‘no doubt’ facing another wave of coronavirus in the autumn, the country’s national national statistician has warned.” Diamond might be right – things tend to go up after they have gone down! – but it’s not his job, or that of ONS, to forecast.
All in all the interview wasn’t a high point. The ONS needs to stick to its core work – getting better statistics. The ONS website declares: “Our main responsibilities are collecting, analysing and disseminating statistics about the UK’s economy, society and population.” The business plan says more, notably: “In 2020 the UK Statistics Authority launched its new strategy: Statistics for the public good. It sets out a radical vision for the UK official statistics system, sets the overarching mission and describes the core principles that underpin the mission. The strategy will seek to deliver high quality data to inform the UK, improve lives and build for the future.”
That is the correct and laudable aim. I’d encourage Ian Diamond and the ONS to stop “giving a voice”, stop fringe groups from corrupting surveys, stop wasting time and money on court cases, end the practice of making forecasts and (possibly) stop quoting unpublished data*. Instead, I’d like a refocus on getting decent crime statistics and improving the guidance so that people can understand the pitfalls.
* Update 18 March:
I approached OSR who kindly got data from ONS – and pointed to the exports sentiment data, to HGV and ships data and roro freight data. It is clear than February was more active than January but most of the data sources do not have data for the same period last year, or earlier years, so while Diamond’s comments might be right, the evidence is not solid.