Population studies of schizophrenia over time are not as useful for several reasons.
1. The diagnostic criteria of schizophrenia have changed significantly over time. You can try to back out with one set of criteria but obviously that gets a little hazy, which is why you see some studies that say it's gone way up and others that say it's pretty much the same, and even a few that say it's gone down recently.
2. Marijuana use on the population level hasn't changed much in the past
30 years. If marijuana use does cause schizophrenia at 5 per 1000 and the base rate of schizophrenia is 10 per 1000, then we would expect schizophrenia rates to go from...
.33 * 15 + .67 * 10 = 11.65 per 1000 in 1985
...to...
.38 * 15 + .62 * 10 = 11.90 per 1000 in 2013
...and that .25 difference is about one-tenth the error bar in population studies even though (for America) that .25 difference means 75,000 schizophrenics. "How can that be statistically insignificant?", you might ask. Science is hard, that's why.
Longitudinal studies are more useful because you can control for a lot of the things population studies can't, and you can find cause and effect. Even if you believed the population studies that say schizophrenia has increased, there's no way of knowing from that whether the schizophrenia was caused by marijuana use, vice versa, or whether the two are totally unrelated. With a longitudinal study, you follow your people from time A to time B to time C, and you keep track of which person is which. Consider...
Person A and person B are born.
Fifteen years later, person A smokes marijuana.
Five years after that, person B develops schizophrenia.
Two years after that, person B begins smoking marijuana in a (mistaken) attempt to mitigate their symptoms.
Three years after that, both participate in a population study, which finds that in this sample there is 100% marijuana use and 50% schizophrenia incidence.
Had they each been participating in a
longitudinal study, the authors would have instead recorded person A as user -> no schizophrenia and person B as non-user -> schizophrenia -> user.
You can hope these distinctions average out over a large enough sample in your population study, but I don't see why you'd bother when the longitudinal studies have already been done, and continue to be done many times over by many authors.
.
Thread: Age of Marijuana Use and Psychosis
You're a joke.
At least it doesn't take me a month to get a punchline. :D