Since the legalization of limited marijuana sales in Colorado launched in January 2014, there's been lots of debate among experts about the risks of stoned driving, as well as dueling statistics aplenty. Note that one report that claimed legalization hadn't caused an increase in highway fatalities, while another maintained that pot-impaired driving fatalities were up 100 percent over the past five years.
Now come new studies from the National Highway Traffic Safety Administration, which argue that while drivers with marijuana in their system are more unsafe than those who remain entirely sober, they're far less dangerous than those who drive while drunk.
We've included the entire report below, but here's an excerpt:
The results in Table 6 show that alcohol (?0.05 BrAC),together with no drug presence, has the greatest effect on crash risk (shaded), raising the risk 6.75 times over that for drivers with no alcohol and no drugs. The adjusted odds ratios for alcohol levels ? 0.05 BrAC with drugs, and for alcohol levels ? 0.05 BrAC without drugs, are both significantly increased (more than 5 times higher). The relatively small difference between the odds ratio associated with alcohol levels at or above 0.05 BrAC, with and without drugs, was not statistically significant.
And here's a look at the aforementioned graphic:
How to interpret these numbers? 9News puts them in laymen's terms: "Drivers with a blood alcohol level of 0.08 are 400 percent more likely to get into a car accident than a sober driver" —while "drivers testing positive for THC are about 25 percent more likely to crash." However, "when you factor in demographic information like age and gender, that number drops to about five percent."
Those digits appear to suggest that drunk drivers are between sixteen and eighty times more likely to crash than stoned drivers. But you can bet the findings will be disputed soon, despite them having been assembled by a federal agency.
Look below to see a KIRO-TV investigation into stoned driving, followed by the NHTSA document.