Actually, that is precisely why you should form a hypothesis based on correlation. A properly conducted experiment should elucidate any causative link between

Actually, that is precisely why you should form a hypothesis based on correlation. A properly conducted experiment should elucidate any causative link between the elements and the direction of those elements. Hypotheses are cheap. Create lots of them. Being 'wrong' about a hypothesis is awesome. It's the path of Reason. (Yes, I woke up this morning wanting to capitalize terms for definitional emphasis. Sorry.)