There's nothing like getting kicked out of school to make a kid start jonesing for some weed.

That's the implication of a new study published last week in the American Journal of Public Health. The authors found that "students attending schools with suspension policies for illicit drug use were 1.6 times more likely than their peers at schools without such policies to use marijuana in the next year." That result held for the student body as a whole -- not just for kids who were suspended.

The study crunched numbers from the International Youth Development Survey, which surveyed representative samples of 7th and 9th grade students in Washington State and Victoria, Australia -- two places that are demographically similar, but where schools take drastically different approaches to drug use.

Those approaches are summarized below. "Washington school policies have been more oriented toward total abstinence and more frequently enforced with harsh punishment (such as expulsion or calling law enforcement)," the authors write, "whereas policies in Victoria schools have been more reflective of harm minimization principles."

For instance, 70 percent of Washington schools get the cops involved with instances of drug possession, compared to less than 30 percent of Australia schools. Washington kids caught with drugs are significantly more likely to face mandatory suspensions, and about 50 percent more likely to be expelled completely.

The survey asked 7th graders about their drug use, and then asked them again two years later as ninth graders. The researchers then looked at the different effects of schools' drug policies on their students' drug use. Most school policies didn't appear to have any significant effect on marijuana use -- expulsion, getting the police involved, enrolling kids in a substance abuse program, or mandatory counseling with a nurse. Even more surprisingly, suspending kids was actually associated with a 60 percent increase in the odds of drug use at that school -- even for kids who weren't suspended.

"That was surprising to us," said co-author Richard Catalano in a press release. "It means that suspensions are certainly not having a deterrent effect. It’s just the opposite." And according to Catalano and his colleagues, suspensions "related to unintended negative outcomes for the suspended student, such as disengagement from school, delinquency or antisocial behavior, smoking, and alcohol and drug use."

As with any study like this, a couple caveats are in order. Observational studies don't necessarily indicate causation. It's possible that other unseen factors are driving pot use at schools that suspend their kids for it. However, the authors did test for reverse causation, examining whether suspension policies are put in place in some schools because more kids in those school use drugs. But they didn't find any evidence for this hypothesis.

But they did find one factor that actually did seem to decrease the likelihood of drug use: student-teacher interactions. At some schools, students caught with drugs were sent to their teachers to discuss the consequences of drug use. This was associated with a 50 percent decrease in the odds of later marijuana use.

This is an important message for American policymakers, as "zero tolerance" policies are in effect at many American schools. They recently came under the spotlight after a Virgina school suspended an 11-year-old for a full year after finding a leaf in his backpack that turned out to not actually be marijuana. Just ask that student, who reportedly has been suffering from panic attacks and depression, how effective a suspension policy is.

Mandatory suspensions appear to be doing a lot more harm than good. Zero-tolerance policies for drug use haven't been effective in the criminal justice system, and this study finds that they don't work in schools either.