Restlessness for the true cause is a constructive habit, an insurance against gullibility. And though correlation does not prove causation, it is often a good hint, but a hint to start asking questions, not to settle for easy answers.
There is one caveat. Here and there you will come across a tendency to dismiss almost all statistical findings as correlation– causation fallacy, a rhetorical cudgel, as one careful critic put it, to avoid believing any evidence. But we need to distinguish between causal associations often made for political ends and proper statistical studies. The latter come to their conclusions by trying to eliminate all the other possible causes through careful control of any trial, sample or experiment, making sure if they can that there is no bias, that samples are random when possible. The proper response is not to rubbish every statistical relationship but to distinguish between those that have taken some thought and those that were a knee-jerk.
So what, finally, about the correlation at the beginning of this chapter, of being overweight and longevity? It is true that the data from America shows overweight people living a little longer than thin people. So what is the third factor that makes unreliable a causal link between putting on weight and adding years? One possibility is illness. When people become very ill, they often tend to become very thin. It is not, in their case, thinness that causes death, but impending death that causes thinness. Put their fates into the mix and it has a marked effect on the results.
This argument is far from settled. A technical quarrel? Yes, but more importantly an imaginative and a human one. You do not need a course in statistics to be struck by the realisation that seriously ill people often lose weight. We can all make causation errors, and we are all capable of detecting them. Think twice.
Finally…
This book has aspired to simplicity. Here, almost seriously, is simplicity simplified: a guide to the basic principles for seeing through a world of numbers, on one page.
Size is for sharing
Numbers are neat. Life isn't
People count (with backache)
Chance lurks
Stripes aren't tigers
Up and down happens
'Average = middle' = muddle
You can't see wholes through keyholes
Risk = People
Most counting isn't
No data, no story
They don't know either
Easy shocks are easily wrong
Thou art not a summer's day, sorry
This causes that – maybe
Acknowledgements
This book began over a pizza as an idea for a radio programme that few took seriously. 'Numbers? On the radio?!' Over the years, with the encouragement and imagination of those few, it found a place on Radio 4 as the programme More or Less. The listeners' response was overwhelming. It acquired a growing, often devoted and delightfully interfering audience of up to a million, a web site, imitators in the press, the enthusiastic support of the Open University, the interest of publishers, a place in BBC journalist training, and finally, though we hope this is not the last of its manifestations, it became this book.
Helen Boaden was the far-sighted Radio 4 Controller who took the plunge; Nicola Meyrick the programme's incisive editor since its inception. Mark Damazer was Helen's successor and became, to our delight, More or Less cheerleader in chief.
We have been lucky to work with some talented journalists who, one by one, came through the programme as if it were a revolving door, bringing energy and a smile and leaving it, smile intact, to spread the word. Some of their reports were the basis for examples used here. Thanks to Jo Glanville, Anna Raphael, Ben Crighton, Adam Rosser, Ingrid Hassler, Sam McAlister, Mayo Ogunlabi, Jim Frank, Ruth Alexander, Paul O'Keeffe, Richard Vadon, Zillah Watson, our PAs Bernie Jeffers and Pecia Woods, and especially to Innes Bowen, whose tireless intelligence has been invaluable. Many others in Radio Current Affairs and Radio 4, studio managers, internet and unit support staff, provided creativity and quiet professionalism that enabled us to spend our time as we should, pulling out our hair over the content. Thanks too, to Gwyn Williams, Andrew Caspari, Hugh Levinson and the many colleagues and kind reviewers, the hundreds who have written to us, the hundreds of thousands who have listened, all of whom, in one way or another, have egged us on.
There have now also been hundreds of interviewees and other direct contributors to the programme and hence to our thoughts in this book. To single out any one from the ranks of wise and willing would be unfair. All deserve our sincere thanks.
We would particularly like to acknowledge the extensive help, comment and advice of Kevin McConway, the best kind of scrupulous and generous critic, and others at the Open University, and also Helen Joyce, Michael Ranney, Rob Eastaway, Rachel Thomas, Gwyn Bevan, Richard Hamblin, and the assistance of Catherine Barton.
Andrew Franklin at Profile Books, razor sharp as ever, and all his skilful colleagues, Ruth, Penny, Trevor and others, somehow make publishing fun and humane, even while labouring against our awkwardness. Thanks, again.
Finally, thanks to Catherine, Katey, Cait, Rosie and Julia for their love, thoughts, support and, what's most important, for being there.
We have done our best to avoid mistakes, but we know that we have failed to catch them all. We are very grateful to readers who spotted mistakes in the first edition and very much look forward to hearing from readers who spot more. Thank you in advance.
Further Reading
There's a growing library of popular science books on all that's fascinating about numbers – from stories about enduring mathematical theories to histories of zero and much more – and they're often intriguing and entertaining. This list, by contrast, is of books about how to make sense of the sort of numbers that find their way into the news or are likely to confront us in everyday life. Books, in other words, to further the aim of this one.
The best of those on bad statistics is still Darrell Huff's classic, now more than fifty years old, How to Lie with Statistics (W. W. Norton, 1993). It is short, lively, timeless, and gives numerous amusing examples. If you were inclined to think people must have grown out of this kind of rudimentary mischief by now, you'd be wrong.
Joel Best comes at similar material as a sociologist. His concern is with why people say the daft things they do, as much as with what's daft about them, and he shows how the answer – interestingly, given the title of his books – is not only to do with duplicity. Damned Lies and Statistics, and now More Damned Lies and Statistics (both University of California Press, 2001 and 2004), have hatfuls of contemporary examples of rogue numbers, including one now famously described as the worst social statistic ever. These books are clear, thoughtful in their analysis of the way the number industry works (particularly among advocacy groups) and entertaining, and offer a good guide to more critical thinking.
Innumeracy by John Allen Paulos (Penguin, 2000) is equally replete with examples, often funny, occasionally complaining, but sometimes brilliantly imaginative, about all manner of numerical garbage. It tends towards the psychology of these errors, asking why people are so susceptible to them and offers forthright answers. The innumeracy he has in his sights is partly an attitude of mind, which he strives to talk readers out of. The book also gives a useful reminder of how simple classroom maths can be used to represent the everyday world.
Gerd Gigerenzer annoys one or two proper statisticians by preferring to be broadly intelligible than always technically correct. Reckoning with Risk (Penguin, 2003) does two valuable things well: it weans the reader off an attachment to certainty, and shows how to talk about risk in a way that makes more intuitive sense, even if it does cut one or two corners. We use the same method here.
Dicing with Death by Stephen Senn, is full of wry humour and also deals with risk and chance, particularly in health. It is more technically difficult in places, and adds chunks of historical colour, but is well worth the effort for those who want to begin to develop an academic interest.
Risk
by John Adams (UCL Press, 1995) is mostly accessible, is skilfully provocative on subjects the reader is surprised to find are not straightforward, and makes a sustained case about the nature of behaviour around risk. It also includes some social theory, which won't be to everyone's taste, but the power is in the numbers.
Simon Briscoe's Britain in Numbers (Politico's, 2005), is an extremely useful survey of the strengths and weaknesses of a wide range of economic and social indicators. He has a sharp eye for a flaw and a sardonic line in political comment.
The Tyranny of Numbers by David Boyle (Flamingo, 2001) is, as the title suggests, a polemic against the fashion for measuring everything. It overstates its case, makes hay out of the woes of certain historical figures, and altogether has a great time railing at the world's reductive excesses, just as a polemic should. For fun and provocation rather than measured argument.
Two more specialist but excellent analyses are David Hand's Information Generation: How Data Rule Our World (Oneworld, 2007) and Michael Power's The Audit Society (OUP, 1997).
For a more formal, elementary introduction to statistics, Statistics without Tears by Derek Rowntree (Penguin, 1988) is a good place to start, particularly for non-mathematicians.
It's worth making one exception to our rule about books on numbers in the news, with three that are worth reading as entertaining introductions to a more numerical way of thinking, all by Rob Eastaway. Why Do Buses Come in Threes, How Long is a Piece of String and How to Take a Penalty (all from Robson Books, 2005, 2007, 2003 respectively).
If you want to find out some up-to-date numbers, the website of the Statistics Authority (which also leads to the site for the Office for National Statistics), www.statistics.gov.uk, is good for the UK, with Social Trends, www.statistics.gov.uk/socialtrends37, also a useful place to start. For an outstanding presentation of numerical information about the world try www.gapminder.org. For all manner of nation-by-nation data generally, searchable in umpteen different ways, there's www.nationmaster.com/index.php or the OECD factbook http://oberon.sourceoecd.org/vl=18803964/cl=12/nw=1/rpsv/factbook/ or you could try the United Nations Statistics division http://unstats.un.org/unsd/default.htm. Most government departments, health, the Home Office for crime, and so on, have statistics or research divisions with much useful material available online, and are generally easy to find.
There are a number of good web-based commentators. Chance News regularly steps into the fray on statistically related news stories: http://chance.dartmouth.edu/ chancewiki/index.php/Main_Page. STATS http://www.stats.org/, attempts something similar, though feels less academically detached and seems to regard its main objective as challenging food and environmental health scares. John Allen Paulos has a column which is always entertaining on the ABC News website: http://abcnews.go.com/Technology/Whos-Counting/. Channel 4 Fact Check is usually sharp, relevant and thorough http://www.channel4.com/news/factcheck/. There are a great many other blogs and the like that regularly find their way onto statistical stories, and of varying political persuasions, but apart from http://blogs.wsj.com/numbersguy at the Wall Street Journal, which is consistently level headed, and Marcus Zillman's useful collection of statistics resources http://statisticsresources.blogspot.com/ readers are probably best advised to find their own favourites.
The Tiger That Isn't Page 20
