In fact, I believe the frequency with which correlation implies causation is a primary tool of diagnostic medicine. However, what you say next is equally true: You may not check for a broken thumb if your patient has a sore knee, but even if you strongly suspect arthritis you check it out rather than just assuming that's the cause.

My old friends "evidence" and "proof" again: The former is not synonymous with the latter.

Definitely no argument there, but then, I've of the school that says, "even if a desire can be definitively proven to be wrong we should seek to overcome rather than eliminate the desire. " Trying to make people want what the majority deems acceptable makes me very nervous; with a few exceptions (mainly where those incapable of legally granting consent are involved) I'd rather have them want to do things from which they refrain due to popular censure.