Best. Comic. Ever.
Now, here's the real fun. I posted almost this exact same thing over on Scatterplot. How long do you think it will take for someone to point out archly that since we have information on the ordering of events (i.e. first he thought correlation=causation, then he took a statistics class, then he decided correlation didn't equal causation) we may be able to infer causation? I bet we get it in five comments or less.
Special thanks to xkcd for the awesome.