10 October 2018

Lies, damn lies and statistics – how dodgy figures feed fake news


As soldiers in a war of words go, the head of the UK Statistics Authority might seem an unlikely candidate.

But earlier this week Sir David Norgrove came out swinging over the cavalier use of figures by Education Secretary Damian Hinds.

Sir David wrote:  “I am writing to raise with you serious concerns about the Department for Education’s presentation and use of statistics.

“The UK Statistics Authority has had cause to publicly write to the Department with concerns on four occasions in the past year. I regret that the Department does not yet appear to have resolved issues with its use of statistics.”

That language is pretty much as damning as it gets from a civil servant and statistician. Indeed, the only time I can recall such harsh language coming from Sir David previously was when he raised concerns about Vote Leave’s infamous ‘£350m a week for the NHS’ claim back in 2016. (He repeated his telling off in September 2017, writing to Boris Johnson after he continued to push the line.)

Yesterday’s intervention centres on Hinds’ claim, repeated at the recent Tory conference, that 1.9 million more pupils are now studying in good or outstanding schools. That prompted a complaint from his Labour shadow Angela Rayner and a ticking off from Sir David. In his letter to Hinds, Sir David says that while the number Hinds and his ministerial colleagues use is “accurate as far as it goes, this figure does not give a full picture”.

Without getting too deep into the weeds, there is important context around an increase in student numbers and gaps between Ofsted inspections of schools that helps in understanding the 1.9 million figure.

Another complaint put to Hinds was about his boasting of Britain’s rise in the reading ability charts, which the Secretary of State clearly wants to attribute to Conservative policies. The problem is his department did not give the proper timeframe for Britain’s ascent up the league tables. Indeed, in a reply yesterday, Hinds had to concede: “We agree that we could have been clearer that the improvement from 19th to 8th was between 2006 and 2016.” Rather crucial information that.

This is just the latest example of data being misused in public life, typical of the post-truth era in which we live. The likes of Channel 4’s Full Fact, Philip Tetlock and, most famously, Nate Silver, have all tried to improve how we understand probability and statistics.

However, this task becomes harder for those who try to explain the data, and those of us who try to understand it, when politicians and campaigners are so ready to stretch the credibility of the statistics they use to breaking point – just look at one of the Lib Dems’ outsized local campaign graphs for several fine examples of the genre.

Iain Duncan Smith was another who frequently clashed with the UKSA when he was Work and Pensions Secretary. In May 2013, Sir Andrew Dilnot, the organisation’s chair at the time, castigated IDS for publishing a press release that claimed 8000 people had got back into work as a result of the benefit cap he had introduced. Sir Andrew noted pithily that the figure was “unsupported by the official statistics published by the department”. Ouch.

IDS’s response was telling. Speaking on the Today programme at the time, he said: “I have a belief I am right”. You could not hope to find a more perfect post-truth answer – using emotion instead of empirical evidence to push a political agenda.

Such an approach to statistics and data from senior politicians is seriously damaging to public trust, providing an environment ripe for fake news. It makes the whole debate around the statistics far more confusing than it needs to be, reducing it to a typical he-said-she-said political spat, rather than an analysis of the available evidence.

There are also key matters of principle at stake. If the data being pushed by the Government is inaccurate, why should people believe it? Why shouldn’t they believe a random number they see on the internet instead? More broadly, why should the public believe politicians at all if they cannot be trusted in these circumstances? Again, this confusion makes for a perfect environment for fake news to gain prominence.

In the introduction to his book The Tiger that Isn’t, written with Michael Blastland, Sir Andrew laments: “Too many find it is easier to distrust numbers wholesale, affecting disdain, than to get to grips with them.” This is hardly surprising given their regular misuse in public discourse.

As Sir Andrew also points out in his book, numbers are not perfect, but when used properly they are crucial to understanding public policy and its effects. They can be a vital tool in helping us navigate post-truth era. But to have any hope of this, we must demand our politicians stop abusing them. Sadly, I remain pessimistic this will happen anytime soon.

Charlotte Henry is a writer and commentator, and author of the forthcoming book 'Not Buying It'.