Yesterday we heard that a new study had been published in the Archives of Neurology reporting that people who had (ever) smoked were 1.4 times more likely than people who had never smoked to develop MND. The full details of the research paper are Wang et al Archives of Neurology 2011, 68(2): 207-213.
I’ve been struck by the way that the figures about this study have been reported in the newspapers. So, I decided it was time to get on my soapbox and try and explain the numbers involved.
This study falls into an area of research called epidemiology. The standard way of reporting results is by quoting ‘relative risks’. When two things are equally likely to happen, a relative risk of 1 is assigned. Factors that may make something less likely to happen would reduce the relative risk to less than 1 (eg 0.7). Factors that may something more likely to happen would increase the relative risk to more than 1 (eg 1.4).
A relative risk of 1.4 means that there is a 40% increase in something happening, however, the really important thing to remember is how big the actual number is to start with.
Here is an example, earlier in the month there was a news story about plans to reduce the size of a bar of chocolate, from 140g bar to 120g bar. Which would cause you greater outrage: the fact that the new bar is 14% smaller or that there is 20g less chocolate? Of course, they are both the same, it is just a different way of explaining the reduction in size.
Rather than give you more examples, if you’re interested in this subject, I strongly recommend that you read the Making Sense of Statistics guide produced by Sense about Science. If you have access to a copy of New Scientist magazine, there is an excellent article on statistics in the 12 February 2011 magazine.