Today there has been significant news coverage (at least in the UK) that the BMJ are withdrawing statements from two papers on the possible harmful side effects of statins. Here are the Guardian and BBC articles on it.
This led me to wonder how common such issues are in academia in general. In this case it appears the authors made an incorrect conclusion from a study which was not picked up in peer review. I can think of lots of similar possible situations where proofs may turn out to be incorrect or other data may be wrong/not properly dealt with.
In particular I'm thinking about less popular/controversial subjects where such things are unlikely to make the general news. Is it standard practice to issue a retraction if things are incorrect? If anyone has any numerical data on this that would be ideal.