The Tiger That Isn’t by Michael Blastland and Andrew Dilnot, published by Profile Books
This is a book with a daft title but a very important message. It confesses that its structure is unusual but manages to stroll comfortably through its territory. Moreover, it brings an accessible and sometimes a colourful approach to that most taxing of subjects, the presentation of statistics.
The book’s central theme is that the presentation of statistics has become ubiquitous, but that we are frequently confronted with insufficient information, or by poorly presented, misinterpreted or inappropriately summarised reports that bring statistics into disrepute. Or simply mislead us.
In calling for a more mature and informed approach to using numbers, it brings to mind a comment by the philosopher AC Grayling that: “Science is the outcome of being prepared to live without certainty and therefore a mark of maturity. It embraces doubt and loose ends.” Statistics, being the science of numbers, can be looked at in the same way.
The sad truth is that we are so often ill-equipped to cope with these ‘loose ends’. We may recognise the complexity of the world around us, but it troubles us – in the debate over climate change , of the health risks or benefits of alcohol consumption, of the challenges in tackling inequalities between and within nations, the current crisis in the world’s financial institutions, or closer to home, the perilous state of pension provision, or the implications of the current flu pandemic. In the face of these ‘loose ends’, we yearn for certainty, for a simple answer to which we can cling. And we find as our allies journalists and headline seekers whose trade is to discard untidy evidence to produce a degree of diluted banality that would make a homeopath blush.
But it isn’t just the ‘general public’ that struggles with these things. You may have read recently about the trouble Harriet Harman has got into for the curious interpretation she (and that means her advisors and supporting civil servants one assumes) has put on the gender imbalance in income levels - see the story at http://www.bbc.co.uk/blogs/thereporters/markeaston/2009/08/statistical_gender_bending.html. As professionals, we need to guard against laziness or the temptation to over simplify.
This is the central thrust of ‘The Tiger that Isn’t’. In an age that is obsessed with risk, and which celebrates performance measures, targets, comparative tables of nations and quantitative assessments of policy impact, we need to do more to ensure that the numbers we use provide meaningful measures and draw people into the discussions about ‘performance’, ‘achievement’ and ‘progress’ rather than stimulating an unthinking reflex nod of approval.
Blastland was the creator, and Dilnot the first presenter, of the BBC R4 programme ‘More or Less’ (http://news.bbc.co.uk/1/hi/programmes/more_or_less/default.stm), and in their book they set out a delicious array of nonsense, examples of the misuse and misunderstanding of statistics. Most of the examples are not instances where there has been a deliberate intention to deceive, more often a dull headed, unquestioning approach to interpretation. The book takes the reader through the challenges of numbers and statistics viewed from different angles, starting simply with a review of how counting is not always a straightforward exercise. There is the example of a headline reporting that one in four young people ‘is a criminal’ – yes, a real example widely published in British papers in 2005. But once you explore the detail, it becomes clear that the definition of ‘criminal’ is so wide (including giving a sibling a shove during an argument, for example) that it becomes meaningless. Good headline though.
We frequently see stories about the risk introduced by certain behaviours. The book not only reminds us to be cautious, suggesting ways to reflect on these claims, but it debunks some urban myths too. Take mobile phone usage and ‘cancer’. A research paper from Sweden suggested an increase in the incidence of a certain brain tumour (acoustic neuromas) associated with greater mobile phone use. News coverage summarised this as a doubling of the risk of developing the tumours. Crikey, that sounds pretty bad! The British Radiological Protection Board warned that children should avoid using mobile phones. What the stories didn’t say was that the baseline risk was 0.001 per cent, or one person in 100,000 would naturally develop the tumours, but that the research suggested that prolonged exposure to mobile phones would increase this to two people in 100,000. Nor was prominence given to subsequent research with more substantial samples, which suggested that actually there may be no impact at all.
Another chapter reminds us that it is often valuable to personalise numbers – to re-calculate them using population numbers for example, or look at them in terms of numbers of people so we can see what they mean to us as individuals. Presented with an announcement of new public spending of £10 million to boost singing and music education in (English) primary schools, we might be struck by what a large sum it is, and have the impression that the impact will be significant. So it may be, but bear in mind that once we factor in the number of children in primary school (around 5 million), we might be in a better position to wonder how much the strategy is likely to deliver.
Having given a BBC radio programme a puff of publicity, it is only fair to give them a kick too. In a radio news discussion recently, a journalist was questioning a guest on the representation of women in sport and he rounded the piece off in an upbeat mood. “Well”, he pronounced (and I paraphrase from memory), “one in ten women play sport, and so the fact that one in five members of the boards of sports are women isn’t a bad proportion!”. Think about it – he is completely misunderstanding what each of those numbers tells us, and by throwing them together, is drawing a spurious conclusion.
So I would recommend this book as a thought provoking and constructive contribution for those of us who compile and work with statistics, whose role it is to summarise or present them for others, and for those that need to understand numbers in their every-day lives. Pretty much everyone, in fact.