Strong evidence of a weak effect is nothing to panic aboutThanks to the World Health Organisation,we can now be certain that processed meats get some cancers more likely. How much should we worry? It is stretching things a bit to say that bacon causes cancer in the same strong sense that fossil fuels cause global warming. Meat consumption is neither a necessary nor a sufficient cause of colorectal cancer: some people can eat processed meat and die of other things; others can contract the cancer without ever touching bacon or even beef. It is in the nature of statistical correlations that we can know that processed meats cause some cases of cancer, but we can’t know which ones in specific.
This kind of linkage seems particularly difficult to think about, or the baits of the media don’t get thinking easier. Strong evidence of a weak effect is too easily read as strong evidence of a large effect,and stories are written to get this exciting confusion easier to tumble into. Conversely, whether the effect is small, or we tend to suppose that the evidence for it must be weak,too. That inference is just as inaccurate.
Continue reading...
Source: theguardian.com