In philosophy we've been studying the varrying types of logical falacies. one, called "modus tollens", seems to be the fallable method used for 99% of the anti-drug "studies" I've seen.

To explain modus tollens:

If A, then B.
therfor, if B, then A.

expanded to a situation

if it's raining (A), then I get wet (B).
therfor, if I'm wet (B), then it must be raining (A).
obviously there are numerous other causes that cause you to get wet, so your being wet does not by default mean it's raining.

and to cite an example of a drug study:

If a person is depressed (A), they smoke weed (B).
therfor, if a person smokes weed (B), then weed is why they're depressed (B).

doesn't work, yet most people buy that unforunately.