Of the roughly 3,000 Washington drivers involved in fatal crashes from 2010-2014, almost two-thirds of them tested positive for drugs or alcohol, but a large portion of those crashes seemed to have involved some combination of alcohol, marijuana or other drugs, obscuring how often marijuana was a factor.

The Washington Traffic Safety Commission looked at information filed through the state’s fatal crash reporting system and analyzed it against toxicology results for dead and surviving drivers in fatal crashes.

The study did not track fatal crashes against changing laws for recreational marijuana use.

The researchers also found 84 percent of drivers who tested positive for any cannabis in 2014 were positive for one of its more intoxicating ingredients, compared to 44 percent in 2010.

During that period in Clark County, of the 98 drivers tested following fatal crashes, 56 tested positive for driving having used a drug.

Sixteen tested positive for only alcohol; three for one of the active ingredients in marijuana; one for residual traces of the drug, which is more of an indicator of past use; and nine for a combination of the two or other drugs. The commission released the data Thursday.

There are some limitations to the state’s report, namely that it’s not a very representative sample of drivers, said Staci Hoff, research director at the commission.

“This describes an already high-risk group,” she said. They’re the kind of people who will use and drive, and are likely more reckless than other drivers.

Other studies the commission cites, ones that found the causal relationship between crashes and pot use to be minimal, didn’t distinguish different THC compounds in testing, often didn’t collect test samples soon enough and have problematic sample sizes.

The number of drivers in fatal crashes who tested positive for only delta-9-THC — a psychoactive chemical that enters the blood and brain soon after consuming marijuana — was very small, as it often is in such studies. Small sample sizes make it difficult to draw conclusions.

Whittling out all the drivers who had used alcohol or another drug leaves only 56 people who tested positive for the quickly intoxicating compound of THC.

“Because of small sample sizes, they look at drivers who not only had THC; they had alcohol, they had other drugs,” she said. “Not only can we not make those inferences, no one else really can at this point.”

Many more used alcohol or some sort of combination of drugs.

“Really, I think that the message that comes out of this report is the combination of THC and alcohol — not only is this increasing, but that driver group revealed itself as the most high risk,” Hoff said.

Hoff expects opponents and supporters of marijuana legalization will try to find something they like in this or any marijuana research, which is part of why the commission’s study has its limitations listed so close to the front cover.

“I just think there are so many limitations to what’s already been put out there, and even the conclusions based on those studies are just insane,” she said. “As a researcher, I’m just appalled daily.”

The best studies she’s seen, ones with the most well-done research, indicate that active THC ingredient noted by the commission does increase risk behind the wheel.

What that means will become clearer with time and more research.

“It could be a long time before we determine the level of impairment due to marijuana,” she said.