top | item 17239080

(no title)

alien_at_work | 7 years ago

Except that's not right. It may be. In fact persuing correlations is how we do science. At some point we gain enough confidence in the correlations that we call it causation.

discuss

order

systoll|7 years ago

'correlation does not imply causation' is just the negation of 'if correlation, then causation'.

What do you think the relevant difference between the terms 'imply' and 'conclusively prove' is?

alien_at_work|7 years ago

I view 'imply' as 'suggest' and 'prove' is 'this is a fact'. Correlation does in fact imply causation (which is why correlations are investigated). It just doesn't prove the causation. It may be that the two things you're looking at are both related to some other cause (e.g. ice cream sales and murders increase in summer, supposedly. This does not mean ice cream sales has a direct relation to murder but they probably both have a relation to outside tempature), it may even be pure coincidence. But if you see a correlation you see a place to begin your investigations.