Jacob Sullum parses the new drug arrest numbers. He reminds us:
As I noted in the January 2008 issue of Reason, there is no obvious relationship between marijuana arrests and marijuana use: Increases in arrests do not seem to be driven by increases in consumption, and busting more pot smokers does not seem to result in less pot smoking. While the odds of getting arrested have roughly doubled since the early 1990s, the overall of level of use is no lower now than it was then and may be somewhat higher (changes in survey methodology make it difficult to be sure). In the last several years, self-reported marijuana use has been essentially flat, while arrests continued to climb before dropping slightly last year, when survey data indicate that use rose slightly.
Mark Jacobson has an excellent piece on New York's schizophrenia on the subject here. The contradictions keep heightening.