The Tiger That Isn't

This is more of an informal book review than anything, but I feel I have to share my delight at finding a book that helps people to make sense of numbers.  The full detail of the book is "The Tiger That Isn't" with a sub-title of Seeing Through a World of Numbers, by Michael Blastland and Andrew Dilnot, published by Profile Books, in 2007.  It is available through all good book sellers it seems and they did an RSA lecture on it in September.

For anyone who  uses numbers, this is essential reading, particularly (and this is a personal whinge I think) PR people who are writing press releases on anything to do with numbers.  Any journalist should be locked in a room with it until they can recite the whole thing verbatim, and when questioned, can actually interpret what it means.

The concepts are introduced clearly with examples that are easily pictured and examples of abuses that are amusing and thought provoking.  They look at issues of size and magnitude, how things are counted (a particular issue in the world of skills and education), chance, averages, targets and a common topic on this blog site - risk.

One of their examples looks at the risks posed to the health of children by mobile phone use - invisible waves have been a talking point in this blog, so I go back to this unashamedly.  The key being that while research showed a doubling of risk of acoustic neuroma, the question not posed and the information not provided was what was the baseline risk?  In this case, about 1 in 100,000.

Statisticians will generally put caveats on their numbers - I know if I am reporting something that looks a bit dramatic or is based on a small sample, I like to put a confidence interval around it or put a level of certainty on it.  Press releases will often include this information in the technical bits at the bottom that don't get reported (always see if you can find that original press release if you are worried about a report), or even in the technical notes of the report that the press release is about.  It is worth doing that bit of digging to make sure you have understood the numbers being talked about.

The Tiger That Isn't makes this point again and again, by getting the reader to question every number they are presented with - in many ways, this is what real numeracy is about.  Good work chaps and thanks for a great read.

Comments

[johntheexpat]

'In this case, about 1 in 100,000.'

But haven't you yourself fallen into one of the traps that statistics always throws up?
Is that 1 in 100,000 uses, or 1 in 100,000 average users per year or 1 in 100,000 average users over the course of an average users life? Statistics are nothing without a framing reference within which I can interpret just how often something occurs, surely. Sorry to be pedantic, but 9 out of 10 pedants will agree with me on this, under normal circumstances.

[Jane Holland]

Wow, John the Expat - even more pedantic than me! Mind you I have yet to challenge Private Eye - have you done that?

Anyway, and rather more politely, good point. The facts in the book are these - the baseline was that 1 in 100,000 people would have an accoustic neuroma if they didn't use a mobile phone. The risk doubles with 10 years of regular phone use (I know we could get into a whole arguement about what regular use is). Anyway, the coda to this is that a larger and more robust study later found that doubling of risk to be a statistical fluke.

I hope that puts your mind at rest.

[johntheexpat]

So I guess I will be hard pushed to blame the Mobile Phone Co's for my stinking headaches then. Oh well, perhaps I shall have more luck suing big Pharma or the Oil Giants. They must be guilty of something that affects me.