The Portland Police Bureau and other local law-enforcement agencies are pledging to step up DUII patrols on New Year's Eve to catch intoxicated drivers.

But a new national study says the Oregon roads are getting less drunk and dangerous—and it cautiously credits legal weed.

The frequency of traffic fatalities in Oregon has dropped significantly since the state legalized medical marijuana, according to a new study from Columbia University's Mailman School of Public Health.

The researchers analyzed car crash data from 1985 to 2014 for the U.S. states that have enacted medical marijuana laws. For seven of these states, including Oregon, the main effect model showed that legalizing weed was significantly associated with a reduction in traffic fatality rates, primarily among people aged 25 to 44.

The Columbia University study says the rate of traffic fatalities per 100,000 people in Oregon was 20.6 in 1985 and only 8.5 in 2014.

Oregon voters approved medical marijuana in 1998.

The Washington Post first reported on the study last week, but did not single out the Oregon findings.

The effect of marijuana legalization on traffic crashes is one of the most significant questions surrounding the end of cannabis prohibition.

Many had thought that legal access to marijuana would increase the number of fatal traffic accidents, as it would lead to more impaired drivers on the road. This study tells the opposite story, but says the results aren't simple.

"We are uncertain what the causal chain may be," the researchers admit.

A theory called the "substitution hypothesis" proposes that people in states like Oregon are swapping out alcohol use for marijuana use, which would track with the overall decrease in alcohol-related traffic fatalities.

But why people are less likely to drive high than drunk—or at least get into serious accidents while doing so—is something the Columbia researchers say requires further study.

"Examining the role of operational dispensaries," the study concludes, "would provide additional information on whether increases in marijuana availability via dispensaries lead to changes in fatality rates."