Ben Goldacre is a medical doctor, writer, broadcaster and journalist and has been writing in The Guardian and on his website about pseudo-science, evidence based medicine and media reporting of health and science stories.
Bad Science: The Book
Bad Science is brilliantly written, persuasive, clear, well argued and well referenced.
It would still be entertaining if it simply presented a collection of columns from The Guardian , but it does much more than that. It equips the reader with a "bunkum detector" and presents a very accessible account of how science, and particularly a medical trial should be conducted.
Goldacre teaches how to read a media report of a science finding, particularly a health related one and how to debate with proponents of the more outrageous health claims.
How to Spot Bad Science
There are several elements that should feature in all methodologically respectable trials. If the study lacks those, alarm bells should ring.
- Did it it have control groups?
- How were the participants selected? Were they volunteers? Did they belonged to special groups?
- Were they asigned to trail groups on random?
- Was the trial blind or double-blind trials? Neither participants nor doctors should be aware which participants receive the new treatment and which receive the old one or placebo.
- What was the hypothesis? Was it a genuine hypothesis formulated before the study began, or was it created post-hoc to fit the data?
- What was the statistical significance of results? Were they due to chance or real?
- What was the size of the observed effect? Even statistically significant effects can be small and of little clinical importance.
- Were there other factors that could explain the effect?
- How was the size of the effect reported? "Double the risk" sounds rather bad, but an increase form 1% to 2% is less immediatelly frightening, while using natural numbers (1 extra heart attack per hundred people) is the most accessible to all.
Bad Science: The Examples
The specific targets of Bad Science come mostly from various areas of current mumbo-jumbo-verse.
Particular attention is paid to homoeopathy, with Goldacre presenting convincing evidence of its lack of effectiveness. He also deals with several other bizzare schemes including "Brain Gym" programme inflicted on UK children in many state schools and preposterous claims of makers of skin moisturisers.
A lot of space is devoted to the claims made by practitioners of pseudo-science calling themselves nutritionists with a particularly entertaining but relatively restrained dissection of Gillian McKeith.
It's not just the so-called alternative-medicine that Goldacre targets. He devotes a chapter to the misdeeds of the Big Pharma (supression of clinical trials and research papers, manipulations of study designs, marketing techniques used to push drugs) and extensively covers media treatments of the MRSA furore and the MMR scare.
Good Science: The Wonders of Placebo
The chapter dealing with placebo and other possible explanations for genuine positive effects of fundamentally ineffective treatments is particularly fascinating. It should be a compulsory reading for anybody attempting to form a judgement on what and why might seem to work despite the experimental evidence to the contrary.
*****
Overall, Bad Science is truly indispensable book which makes its points in a light, wry, entertaining manner while maintaining a clarity of presentation throughout. It focuses mostly on UK cases, though the general material is universally applicable.
Use it to answer your own questions, detect sensationalist tactics, educate the confused and question the preachers of bunkum.
Highly recommended, and clear five out of five.
Paperback, 352 pages
Fourth Estate, September 2008
ISBN: 978-0007240197
