Is legalized marijuana a gateway drug to other forms of substance abuse? Or would safer access encourage people to ease back on other vices? This has been one of the major questions bedeviling the legalization debate. Now that four American states have legalized recreational use of cannabis, we ought to be getting some answers. Sadly, a review of the studies done so far suggests the picture isn't getting much clearer.

Tracking the use of illegal drugs is hard, but it should be easier to discover whether the use of alcohol and tobacco goes up or down when the legal attitude to pot changes. Dr. Katarína Guttmannová of the University of Washington has investigated the alcohol side of this equation. "We chose to focus on alcohol because even relatively small changes in alcohol consumption could have profound implications for public health, safety and related costs," Guttmannová said in a statement.

The costs of alcohol consumption are estimated to be greater than any illicit drug, being implicated in well known health effects and a third of driving fatalities.

The association between marijuana and alcohol is difficult to predict. Some people like to combine the drugs, so increased use of one might lead to greater consumption of the other. Others are more likely to substitute a few joints for a night out drinking.

After checking 750 peer-reviewed studies on the topic, Guttmannová found 15 that tested the effects of policy changes on alcohol consumption, be they decriminalization of personal use, medical marijuana regulations or legalization of recreational marijuana.

Unfortunately, the results failed to produce a clear pattern. The paper starts by quoting an article from the Washington Post, saying: “Marijuana policy is nothing if not complicated,” and concludes the same is true of research in the area. Some studies found lower rates of drinking among high school seniors, where personal use has been decriminalized, but others report that college students' legal use of cannabis is associated with higher alcohol consumption.

One interpretation of the data Guttmannová explored is that there are multiple complex effects, and the variation is explained by differences in the demographics and forms of alcohol use examined. For example, a picture could be drawn where easier access to pot increases binge drinking but reduces overall underage consumption. However, many of the findings may be distorted by confounding factors, for example changes to the way liquor is sold in Washington state (where many of the studies were done) the year before recreational marijuana use was legalized there.

There are also questions of what the most important measures might be. Drinking and puffing at the same time increases blood levels of THC relative to the same amount of smoking done on its own, and the two drugs used in combination have been found to double the risk of drunk driving and violence. Nevertheless, establishing the effect of legal changes on total alcohol consumption would be a good start, if only we could do it.