Reckoning with risk: learning to live with uncertainty

Gerd Gigerenzer

Harmondsworth: Penguin, 2002
ISBN 0-713-99512-2 (pb) pp. 310, £14.99

Practical Philosophy  (Book Reviews) Spring 2003 Volume 6.1

Reviewed by: David Arnaud

Gigerenzer has written a book that succeeds in its aim to shock, disturb and educate. After reading this book, I felt that I wanted to shout from the rooftops - or at least a soapbox - many of its claims. A general failure on the parts of both professional and non-professionals to understand probability theory in particular, and to accept uncertainty in general, Gigerenzer argues, are responsible for a great deal of misinformation by professionals (part deliberate and part based on ignorance), and misunderstanding by us. This, as Girgerenzer illustrates with both statistics and personal stories, results in unnecessary suffering.

Here are some of his claims. Most AIDS counsellors fail to realise that AIDS tests can have false positives (discovered through pretending to be a patient), and treat any questioning over the reliability of a positive test as resistance to be broken down. The chance of being killed in a road traffic accident in Great Britain, given an average life of 75 years and the accident rates continuing as they are now, is about 1 in 220 (its about 1 in 90 in America). Women are told the benefits of breast screening, but the costs - false positives for 1 in 2 women who have 10 screenings, the identification of non-fatal cancers (old age gets you first), and the likelihood of increased risk (between 2 to 4 for every 10,000 screened) - are kept from them. This matters, Girgerenzer claims, as the individual needs to weigh the costs and benefits depending upon how much they value peace of mind, and a non-scarred body, and their willingness to take risks.

As well as pointing to the dangers, Girgerenzer has remedies. These vary from the pithy saying (remember Franklin’s law that nothing is certain but death and taxes) to the more sophisticated. The sophisticated include being aware of the reference class and to represent statistics as natural frequencies not Bayesian probabilities. I commend the book to the reader to find out more about these. As an example of being aware of the reference class, Girgerenzer asks us to consider the difference between the implications we draw when a DNA expert says either ‘The probability this match has occurred is about 1 in 100,000,’ or ‘Out of every 100,000 people 1 will show a match’. In the second, but not the first case we are invited to consider what the reference class is - if it is the total population of London then the DNA evidence is very weak as many other people will match the DNA (if the population is 6 million then 60 people will have the match).

This is a very welcome addition to the decision making and critical thinking literature, both for its analysis of how probabilistic thinking goes wrong, and positive contributions to how to represent statistics in ways that make them much easier to comprehend.

 

PRACTICAL PHILOSOPHY page last updated 01/07/2003 contact webmaster: David Arnaud http://www.practical-philosophy.org.uk