This article is a stub. Please feel free to expand it and make it more encyclopaedic.
One of the complications when trying to establish whether a causes b is that:
- "Association does not equal causation"
One of the reasons why mistakes are made is due to the existence of "counfounders".
Confounding occurs when something is associated with an outcome, but does not cause it.
Take, for instance, this recent headline:
- "People who eat organic 25 per cent less likely to get cancer"
The effect size is very small: an absolute risk reduction of 0.6%.
People who buy organic also tend to be to richer; eat more fruit and vegetables and less meat, and drink less alcohol; so it is likely that "eating organic" is actually a confounding factor; a factor which is not the cause of the risk reduction, but likely a marker for other factors which are the true cause.