The overall rating of a review is different from a simple average of all individual ratings.
Share this review on
Lies, damned lies and ….
I was recently introduced to a slim little Penguin volume called “How to Lie with Statistics” by Darrell Huff. The book was first published in 1954, but has been updated a number of times since then. So how come I am enthused to write an opinion on a maths book?
Basically, I loved the book. Politicians love to quote figures, most of which are real, and which always seem to support that politician’s point of view. But then, have you noticed that the opposition take the same survey and quote a completely different set of figures, which now seem to support the opposite viewpoint. Clever eh? Or just confusing!
This book unwraps those sneaky practices. It gives the reader both useful hints and tips for when you might want to present some figures in the best light and a toolkit for examining and questioning statistics when faced with them.
The book takes you through some examples of ‘dodgy’ statistics, how to bias samples so the results come out your way and a choice of averages to show an average the way you want it. (Use the mean, median or mode when it suits you – they are all averages!)
The next chapters cover those little figures that seem to be missing from the story if you look closely at it and then how to make a mountain out of a molehill by getting bogged down in minutiae. My personal favourite is the ‘Gee Whizz’ graph – one of those graphs with sharp rises or drops that you see in newspapers. They make it look as if the whole stock market is crashing, until you notice that the whole scale of the graph is only 5 points. (I wish I could draw one for you to illustrate – they really are very dramatic!)
Statisculation – yes, that’s right – is the art/science of misinforming people by the use of statistical material. It really is quite fascinating.
Finally, the book suggests some questions to ask when faced with a statistic. These are really useful.
1. Who says so? - 8 out of 10 cats prefer Whiskas? A cat food manufacturer would say so. What about those actresses who advertise cosmetics “because I’m worth it…” Why are they saying that? Because they’re being paid to say it. Get the point?
2. How do they know? - Get Carol Vorderman or some other ‘brainy’ person to advertise your stuff and you believe everything they say because they’re clever. Right?
3. What’s missing? - So you can see what they ARE telling you. But what are they NOT telling you? 8 out of 10 cats prefer Whiskas? Remember they had to change it to “8 out of 10 cat owners who expressed an preference…” What about all those who didn’t express an interest? How many cat owners did they ask in the first place?
4. What’s changed? – Last year we only sold 500 pairs of Wellington boots, but this year we have sold 1000. Why? Was the national wellie whanging competition held in the area? Was it a particularly wet spring? Are wellies suddenly the coolest footwear? They are the same boots – or are they?
5. Does it make sense? Are the statistics being quoted to lots of unnecessary decimal places? Should the answer be a round figure or not. Does the answer look ridiculously high or low?
This book is fun and easy to read. It gives you some food for thought and a basis for that sneaking scepticism that you always had when someone quoted spurious figures at you. It also has some nice Mel Calman cartoons. At only £6.99rrp or £7.24 including postage from www.Countrybookstore.co.uk it is good value for money.
Why did you have to write this ? And why did they have to write this book ? 9 out of 10 statisticians asked said that they would rather the darker side of their art was kept secret !!
Seriously, great op on what seems a fascinating book (I must get a copy), but as one who often uses stats to 'prove' his case with figures at work (I'm an accountant) I don't want everyone knowing my secrets !!
I look forward to reading more of your ops.
offy 13.07.2001 22:22
Great op. Although I hate statistics and numbers, I might actually enjoy this book as it sounds like it shows you how the general public are manipulated. I was on a 'business analysis tools & techniques' course recently and we did some of this - change the scales on a graph to show whatever you wanted it to show!
Xelavie 12.07.2001 03:24
Come to think of it, there's great fun in playing with statistics. Something like there's only one customer last month and this month they had 3 so there's an increase of 200% or tripled. Numbers, numbers... Alex