A number of our readers wrote in this weekend to point out that drug czar John Walters was stumping the "marijuana causes mental illness" bandwagon. It was probably inevitable. After all, a year ago we reported, "Reefer Madness Strikes a Leading British Newspaper," and this and other spurious claims have continued to emanate from various outlets and agencies ever since. Still, propaganda is no less irritating for having anticipated it. So I could only sigh when I received a copy of a New York Times story that a member had forwarded, with his note "Walters is at it again." The article did quote people on the other side, which is good. But there's no way around the headline, which is what most people will ever read and which did not reflect any controversy or disagreement over the drug czar's claims. Master stats and criminology expert Matthew Robinson (author of the famed "Lies, Damn Lies, and Drug War Statistics" picked a similar title for his detailed critique of Walters, "Here We Go Again: White House Makes Scary Claims About Marijuana." I'll leave it to readers to follow the link for the bulk of Robinson's analysis, but the major thing to keep in mind is that Walters has not met the three-level burden of proof to back up his claims. Those levels are the following:

1. One must show a correlation. Marijuana use and mental illness have to show up in many of the same people. That might not be so hard to demonstrate, but the reason for the correlation may be as simple as the fact that lots of people use marijuana, so most physicial or psychological issues may be represented among its users. Which leads to the second needed level:
2. One must show a temporal order. That is, it is necessary to prove that marijuana use preceded the onset of mental illness. If marijuana use began later, there obviously is no causation. Even if they start at about the same time, there may be no causation.
3. And then there is a third, very crucial intellectual requirement for drawing the conclusion that marijuana use causes mental illness. That is the need to demonstrate a "lack of spuriousness" -- which means eliminating the possibility that other factors could have led to both the marijuana use and the mental illness. For example, physical or other life issues may have led an individual to become depressed, and that person may have then begun using marijuana because of being depressed. Or there could be biological or personality factors that make both depression and drug use more likely. Or there could be other things going on.

And now you know more about statistics than the drug czar does. :)