A man smokes a marijuana joint at a party celebrating weed April 20, 2016, in Seattle. (Elaine Thompson/AP)

Last week, a pair of studies came to seemingly opposite conclusions on whether rising marijuana use is causing an increase in car crashes in states that have legalized the drug.

The first, conducted by the Insurance Institute for Highway Safety, analyzed insurance claims for vehicle collisions filed between January 2012 and October 2016. The IIHS researchers compared claims in states that had recently legalized marijuana (Colorado, Washington and Oregon) with claims in similar neighboring states that hadn't.

They found that over that time period, collisions claim frequencies in the states that had legalized marijuana were about 3 percent higher than would have been anticipated without legalization. The researchers characterized that number as small, but significant. Collision claim frequency refers to the number of claims filed divided by the number of insured vehicle years.

“The combined-state analysis shows that the first three states to legalize recreational marijuana have experienced more crashes,” said Matt Moore, senior vice president of  the IIHS's Highway Loss Data Institute, in a statement.

But hot on the heels of that analysis came a second study, published in the American Journal of Public Health (AJPH), that found no increase in vehicle crash fatalities in Colorado and Washington, relative to similar states, after legalization.

The authors of that study analyzed federal data on fatal car crashes from 2009 to 2015. “We found no significant association between recreational marijuana legalization in Washington and Colorado and subsequent changes in motor vehicle crash fatality rates in the first 3 years after recreational marijuana legalization,” they concluded.

On the one hand, a finding that legalization led to a small but significant increase in crashes. On the other, a study concluding that legalization had no effect on fatal crashes at all. Do the two contradict each other?

Not necessarily. The studies measured slightly different things: IIHS looked at claims for motor vehicle collisions, while the AJPH report focused more specifically on fatal crashes. It seems plausible that legalization could lead to a slight increase in minor accidents that don't prove fatal.

Indeed, federal research has shown that while smoking weed before driving does indeed elevate your risk of crash, it's nonetheless far less impairing than alcohol, which dramatically increases the likelihood of a crash even at small doses.

Beyond that, the two studies used different baselines for comparison. The IIHS report compared legalization states to their neighbors. The AJPH study, on the other hand, chose comparison states not based on geography but on other shared characteristics: traffic patterns, makeup of roadways and population.

Both approaches have their strengths. But in the end, it's hardly surprising that choosing different comparison states will yield slightly different results on similar measures.

Any increase in car crashes is cause for concern. But on balance, public health experts worry more about fatal crashes than nonfatal ones, for obvious reasons. The AJPH study is heartening news on that front, and it comes on the heels of previous research showing that medical marijuana laws are also not associated with increased vehicle fatalities, and may in fact lead to fewer traffic deaths.

Voters in California, Maine, Massachusetts and Nevada just approved recreational marijuana use. Here's what they can learn from Washington, Colorado and Oregon, states where marijuana use has already been legalized. (Daron Taylor,Danielle Kunitz/The Washington Post)