| hsrob said: It's also apparently used in the social sciences, but as you said this is not the primary application. If it can in fact be applied here it's still not going to give you a definitive answer, nothing can when studying society and behavior. Even if it's not strictly applicable here it's generality should allow it to provide a framework within which to look at this problem which has to be better than the blanket, "correlation does not equal causality" line. It was just a thought. |
It can be used, but you want it in situations that can be more homogeonized. The way different cultures view black markets for goods can cause wildly different results from banning said goods. I stick to the hard-line correlation does not prove causation because it is true for 90% of cases. If researchers were to set-out to prove somethign that can only be shown correlationally you can sometimes get a good result, but most often you just have a correlation.








