In the lead-up to this year’s general election, volunteers and staff at the independent fact-checking organisation Full Fact were putting in 18-hour days, seven days a week. From the 6am news broadcasts onwards, the team worked to provide a rapid verification of the figures that peppered every speech and debate.
And this work continues. Via its Facebook page – which welcomes fans with a banner image that reads “Tired of semi-skimmed answers?” – Full Fact posts its latest assignments and projects.
A recent post, for example, tested the conflicting migration figures in the media: “The Daily Mail claims that only one in five asylum seekers in the EU are Syrian. The Guardian says it’s wrong. We’re on the case.”
Full Fact’s director, Will Moy, has an ambition to turn every person in the UK into a citizen statistician, and he has a three-point plan for how to do it. “We have three questions we tell people to ask,” he said. “The first is: ‘Where does the data come from and do you have any reason to trust the source?’ The second is: ‘What are they actually measuring?’ and the third is: ‘What have they done with the data?’”
But it is clear that these questions are not being asked nearly often enough. When the Royal Statistical Society (RSS) and King’s College London conducted research into the public’s view of numbers versus actual figures the distance between perception and reality was stark.
Two examples are particularly revealing: how much of the UK’s population is Muslim and how many teenage girls fall pregnant. According to the survey, the UK public estimates that nearly 24% of the population is Muslim. The official statistics put the proportion far lower; Muslims make up just 5% of the population in England and Wales. When asked to put a figure on rates of teenage pregnancy, the gap between the official verdict and public understanding was similarly vast. Respondents thought that 15% of teenage girls fall pregnant – 25 times higher than official statistics, which place the figure at around 0.6%.
This perception gap poses risks not only to the way policy is created, but also to the economy. The UK’s lack of data skills means it is missing out on thousands of jobs being created in the “big data marketplace”, according to a recent report from the British Academy.
It’s not just ordinary people and businesses that are struggling to understand data. Even the prime minister has had his knuckles rapped by the UK Statistics Authority (UKSA), the watchdog set up to oversee use of official data.
In a Daily Telegraph article last July, since corrected, David Cameron claimed Tory policy had led to an increase in the number of new jobs going to British workers instead of foreign workers. However, the data he used actually showed the flow of people in and out of employment, as Sir Andrew Dilnot, chairman of the UKSA, pointed out in a letter (pdf)that was copied to Cameron.: “I want to be quite clear that these official statistics do not show the number of ‘new jobs’.... The official statistics provide estimates of net change in the number of people in employment. This is not the same as the number of people who move into employment; rather, it is the difference between flows of people into employment (eg from unemployment, education, ill health, care for dependents, retirement, or from living in another country) and the flows of people out of employment”
Clearly, the quantitative skills gap presents politicians and the media with an open goal. These kinds of misrepresentations of data can have a profound effect on public understanding of immigration levels. The RSS research also showed that the public believes 31% of the population are immigrants while he official figure is less than half, at 13%. Even taking into account illegal immigration, the statistic is closer to 15%, according to the RSS.
Back at Full Fact, Moy offers another example to show how problematic the current use of statistics is in political debate: “One day at PMQs the opposition stood up and complained that a record number of people are waiting beyond four hours in A&E. Then the government stood up and said that, actually, a record number of people are waiting less than four hours in A&E. And they’re both right, because a record number of people are going to A&E.”
It’s such an entrenched issue a 19th-century prime minister, Benjamin Disraeli, coined the famous quote: “There are three kinds of lies: Lies, damn lies and statistics.” So, how can we move beyond the problem?
First we must understand it’s not just the numbers that are to blame. Trust is also an issue, but not in the way you might assume.
The Office for National Statistics (ONS) found, in an independent survey (pdf) carried out in February this year, that the level of public trust in official statistics is actually rather high, at 81%.
This does plummet when it’s the government using the numbers, however. Only 28% of respondents agreed with the statement: “The government presents official figures honestly when talking about its policies”. And it falls even further when it comes to the press: just 19% agreed that “newspapers present official figures honestly”.
Part of the solution, according to experts, including the national statistician John Pullinger, is to have a less tolerant approach to those found guilty of messing with statistics. Moy goes further, and says there should be legal repercussions for anyone breaking the the code of practice (pdf) for official statistics.
Heten Shah, executive director of the RSS, is a little more positive. He’s keen to highlight that politicians are making efforts to use data well. He explains how they often try to get beyond the limiting headline figures, with select committees taking time to “build up a more detailed evidence-based picture of an area”. He also points to government efforts such as What Works centres, where research teams based in areas such as healthcare and policing try to improve the evidence base for policy. Elsewhere the RSS is using education to address quantitative skills gap in public life. It has signed up 55 MPs for training in statistics on World Statistics Day, 20 October, after running a campaign ahead of the general election asking them to pledge time for training.
These efforts are part of a wider range of schemes and projects attempting to boost public understanding of statistics. An immigration quiz, devised by the ONS, was developed to provide a reality check for people by asking them to guess the level of migration in their area and then provide the official figures. It is also trying to build its presence on social mediaThe ONS has more than 160,000 followers on Twitter and is steadily building a bank of Facebook-friendly infographics.
Charities, too, are playing a significant role. Shah notes the crucial part played by the Nuffield Foundation in funding fact-checking in the lead up to the general election. It donated to a range of independent organisations, including the Institute for Fiscal Studies, Full Fact and the Conversation in order to help make sense of the numbers flying around.
And, on all sides, journalists are being taken in hand: 1,500 were trained in basic statistics by the RSS during the past four years. Full Fact has likewise run training at national media organisations, and Moy notes that its finder tool is “in use in various newsrooms”.
Shah says he has found it “fascinating to see the rise of data journalism. Over time we hope that there won’t be a separate discipline called ‘data journalism’; good use of data will just be what is expected.”
Moy holds high hopes for the future of data in the UK: “We’ve seen individual claims getting corrected by politicians or the media … you can improve this. You can make data an opportunity, rather than just a problem.”
The big data debate at the political party conferences
Liberal Democrats party conference, Monday 21 September
The NHS in public debate: a case of data deficiency?
Our panel assesses the impact of quantitative politics on the most hotly contested campaign issue: the NHS.
Labour party conference, Monday 28 September , 12.45pm – 2pm, GB2 in the Brighton Grand Hotel
Migration into the UK: Can the public count on official statistics?
We discuss the impact of the increasingly quantitative nature of political debate on migration policy.
Convservative party conference, Tuesday 6 October , 5.45pm to 7pm, Exchange 4-5 in the Manchester Central
Can wellbeing usurp economic indications of success?
This session will bring party representatives and industry experts to explore the use of wellbeing statistics in political debate and decision-making.
