We can all get to grips with statistics when they are presented properly
We all have to make judgements based on statistics. Take the claim that a glass of wine a day doubles the risk of breast cancer. Should women be teetotal? We mostly struggle to make sense of the numbers behind such judgements.
Many people assume our inability to deal with statistical evidence is just a special case of innumeracy. But Gigerenzer points out that - given the usual way statistical results are presented - experts are as likely to err as the rest of us.
But he also offers hope. Innumeracy is as much a problem of how statistical information is presented as about our ability to understand it. He tested real medics with information about breast screening presented in different ways. Using the usual way of dealing with conditional probabilities, two out of three medical experts would misinterpret the results (and give patients bad advice). Presented as natural frequencies (actual numbers that have not been normalised), the same statistics prompted eight out of 10 experts to give the right advice. When the right presentation is chosen for the facts, people reach the right conclusions.
Unfortunately not many of the experts who bombard us with statistics have taken this message to heart. But everyone who has to make decisions on the basis of statistics can inoculate themselves against many of the common errors by reading this; perhaps they should insist their expert advisers read it, too.
It is a well written and valuable book. It may be the most important - and the most readable - book about statistics you will ever find. Compulsory reading.
Reckoning With Risk, Gerd Gigerenzer, Penguin 2002
JOIN OUR MANAGEMENT REVIEWS
Are you up to speed with the latest management thinking? If you would like to review management journals or books for HSJ, email your suggestions to email@example.com