I have just finished reading Georgina Sturge’s book, Bad Data.
The author is a senior statistician who works at the House of Commons Library and provides MPs with answers to their questions and requests for information. But her researches have led her to believe that much of the data we rely on in public life is what she calls ‘bad’. We put our faith in it, but it is often inadequate or misleading or just wrong. She subtitles her book: How governments, politicians and the rest of us get misled by numbers.
I saw an example of it last week.
The BBC for our region posted on their website an article with the headline: Knife crime rises 140% in Yorkshire and northern Lincolnshire. This made me sit up as it was a statistic I did not recognise and, on the face of it, is alarming. However, it is in fact what Georgina Sturge would recognise as bad data – for three main reasons.
First, it is hard to know where the information on which the headline is based comes from. Crime is recorded by police forces. There are four police force areas in Yorkshire and Humberside: North Yorkshire, West Yorkshire, South Yorkshire and Humberside. There is no force called ‘northern Lincolnshire’, though there is a force for Lincolnshire. So it is hard to be sure which crime figures are being considered. The BBC should – as my old maths teacher used to say – ‘show your workings’. But I will assume the BBC has used the statistics of the four forces of Yorkshire and Humberside.
Second, what I think the BBC researchers did was to compare figures for knife crime in the year ending June 2022 with the figures for 2012. If we look at the Yorkshire and Humberside figures for the twelve months April 2021 to March 2022 (the financial year), the increase for knife crime is not 140% but 80% – still a big increase but not the enormous rise of the headline. The headline is almost certainly wrong.
But third, and more seriously, it misleads.
It is rarely a good idea to compare crime figures for a current year with a single past year. The past year might be a ‘normal’ year as far as crime goes, or it might have been a year in which crime was unusually high or unusually low. For example, the year of the Covid lock-downs and restrictions, 2019, produced dramatic falls in many crimes, so it would make little sense comparing that unusual year with any other. And it would make little sense comparing any year with a single past year without a great deal of explanation and contextualising.
A better way of looking at crime figures is not to compare the current year with a single past year, but to look back over the past few years to see if there is a trend. Figures for a particular year might be up or down compared to any other year; but as you look at the past few years, you can see whether there is a trend: is crime going up or down or staying about the same over that period of time? This is the only way we can know whether particular interventions to tackle crime are working or not.
When I looked at the figures for South Yorkshire over the past five years, knife crime has been falling. From a high in the 12 months April 2017 to March 2018 of 1772 crimes, it has fallen by 11.5% to 1567 crimes in the 12 months April 2021 to March 2022. (I have already noted that 2020 was very unusual because of Covid.) Noticing this falling trend produces a very different headline: Knife crime falls in South Yorkshire by 11.5%.
So what did the BBC do by writing their headline: Knife crime rises by 140% in a decade in Yorkshire and northern Lincolnshire? They not only gave a wrong figure, they also implied that knife crime is going up. This is how some immediately understood what they were being told. One community activist said the police were ‘fighting a losing battle’ – yet knife crime has been falling for the past five years. So it was a misleading statistic and one that would cause alarm. That was hardly responsible – bad data, bad outcomes.
This headline is now in the public domain on a website where no one can challenge it or correct it. And I fear that BBC journalists and others will go on quoting it for a long time to come. But it’s bad data.
(Since writing the above, the BBC has changed the headline to read 90% and not 140%, though it has not explained the change or apologised for the original figure. Meanwhile, the original, false and misleading headline figure has been quoted by other media – where it remains on several websites.)
Each week I am out and about meeting groups and individuals somewhere in South Yorkshire. The meetings are sometimes formal, sometimes not. Last week, for example, I met, formally, Bawtry Town councillors to discuss policing in their town, and then, informally, I had conversations with the Earl of Scarbrough and his gamekeeper at Sandbeck Park, about rural crime.
I am often accompanied by police officers, especially where the meetings are with community groups or elected members. They will have local or specialist knowledge which can help inform the conversations. I value all this and together we can usually answer most of the questions.
I have also noticed something. Quite often, although people ask for information – and it is important that they receive it – what they are also asking for is re-assurance. This, however, is rarely stated as such and is not something tangible. If I use the Bawtry councillors as an example, they needed some factual information: how often did the police come into Bawtry? What patrolling did they do? What were the issues here? And so on. The local sergeant, who came with me, patiently answered all the questions. But underlying the whole conversation was the matter of reassurance.
People want the relevant statistics. They also want to feel – especially in some of our smaller towns and villages – that they are not a police after thought. That is not something that information alone can convey. A great deal depends on the police officer who is giving it. Sergeant Hollie Stenton at Bawtry was reassuring and it made me think about why this was. (It is not always so!)
There is a particular quality that police who engage with the public need to have. I call it ‘psychological awareness’ or simply ‘awareness’. What awareness consists of is the ability not just to hear what is said but to listen – to think about it, turn it over in the mind, take it seriously. The officer who does that becomes tuned in to what people are saying and asking at different levels, and as he or she responds, they deal with what lies behind the questions but is unspoken – the need for re-assurance.
I don’t know whether you can teach ‘awareness’, though it is, after all, something we all have to some extent as human beings interacting with others day to day, especially family and friends. Perhaps that’s how you do it – treating members of the public as if they were part of your extended family. At any rate, if police officers do not have it, all the statistics in the world, however favourable, will never be enough.
Degrees of policing
A couple of years ago, the means of entering the police force changed. On joining the force, recruits either had to have a degree already, or they had to work for a professional degree as part of their initial training. The first batch of those who joined this way graduated from Sheffield Hallam University on Friday. Unfortunately, I was not at the ceremony, which was held at Ponds Forge, but if you were in town on Friday you might have seen them in their gowns and hoods. Studying and being a police officer at the same time has involved a lot of hard work. Congratulations to them all.