You have to give Uncommon Descent poster DaveScot credit. He's not one of life's overly specialized intellects. He's a good, old fashioned generalist, able to talk about absolutely any area of science with exactly the same degree of spectacular incompetence. Today, he's turned his attention to the intersection of mental health and substance abuse.
DaveScot's uninformed ire was sparked, in this particular case, by a news report discussing a paper that recently appeared in the Journal of Clinical Psychiatry. According to the report, the researchers found a strong association between cannabis use and an earlier age for the onset of psychosis:
The results showed a significant gradual reduction in the age at which psychosis began that correlated with an increased dependence on cannabis. Compared with nonusers, age at onset was reduced by 7, 8.5, and 12 years among users, abusers and dependents, respectively, the researchers report.
In further analysis, the effect of cannabis on age at onset "was not explained by the use of other drugs or by gender," they also note. The finding was similar in the youngest patients, suggesting that this effect was not due to chance.
These results "point to cannabis as a dangerous drug in young people at risk of developing psychosis," Gonzalez-Pinto and colleagues conclude.
DaveScot, in his infinite wisdom, read this news report and immediately concluded that the research article is, "a wonderful demonstration of how crap science that supports something politically correct is used and abused all the time". I'm really not sure why Dave thinks that it's "politically correct" to believe that marijuana is potentially a dangerous drug, but that particular question is best left for another occasion. Instead, let's see why he thinks that this is "crap science":
Immediately obvious to me is the possibility that voluntary marijuana use is a symptom of an underlying problem that has nothing to do with marijuana use. People often resort to recreational drug use to escape and/or ameliorate some underlying problem. Alcohol abuse is a classic case of being symptomatic of some other problem. These researchers had no control group to rule out the very likely possibility that people who tend toward psychosis are unconsciously or consciously attempting to self-medicate. The medication isn't the cause, in other words, its a symptom. . . .
But no, the researchers in fact did nothing at all to discriminate between cause and symptom and it's obvious in seconds to even a casual observer such as myself that the study and its conclusions are flawed. Where was the peer review that should have prevented this junk science from reaching the pages of the Journal of Clinical Psychology without correction of obvious flaws?
If you actually take the time to read the research article itself, and not just the Reuters piece, you'll find that the authors of the paper wrote the following:
One explanation for this possible link is that the illness is precipitated by substance use, although it remains uncertain whether this effect is limited to people with a predisposition to psychosis. Another possible explanation is that the early onset of symptoms is a risk factor for substance use. The experience of symptoms could make patients vulnerable to substance use, perhaps in an attempt to cope with the illness or to self-medicate.
Would anyone like to guess how much of the actual paper DaveScot would have had to read to learn that the researchers were in fact aware of the possibility of self-medication as an explanation for the relationship? The answer is right below the fold
If you guessed either "one paragraph" or "seven sentences", congratulations. You're a winner.
Seriously, it's no surprise that the researchers were clever enough to spot the possibility that people might be self-medicating. It is, as Dave's post clearly demonstrates, a possibility that's instantly obvious to anyone with more than two functioning brain cells. The researchers were aware of the possibility, and they explicitly acknowledge that it can be difficult to demonstrate a cause and effect relationship with their methods. In this particular case, they feel that they have at least demonstrated that it's unlikely that the marijuana use is an effect of the psychosis because they were able to show that there is a clear inverse relationship between the age of onset for psychosis and the amount of pot consumed.
Of course, DaveScot has his own suggestion for the right way to figure out if marijuana is the cause or the effect:
What you need to discriminate between cause and symptom is take a randomly selected group of people who aren't marijuana users and administer marijuana to half the group and monitor all of them for onset of clinical psychosis. If marijuana is a cause then the marijuana user group will have a higher percentage of psychotics or if the same rate then earlier onset. If there is no difference in percentage or age of onset then marijuana use is simply symptomatic. If they'd done that they might even find that the non-user group has the worse problem with subsequent psychosis and the self-medication is actually effective to some degree.
I will say one thing for that suggestion: it's efficient. Dave managed to produce a suggestion that's illegal, unethical, and logistically impossible, and he did it all in a single sentence. He then followed up on that by completely failing to recognize any of the many similarities between his suggested prospective study and the retrospective analysis he was criticizing.
The researchers were not able to start from scratch and administer illegal drugs to enormous numbers of people in an attempt to inflict mental illnesses on human beings. Instead, they examined a population of patients with clinically diagnosed psychosis. Like all scientists, they did the best that they could with the data that they could get.