I've seen conflicting articles claiming that DUI fatalities either increased or decreased after states legalized marijuana. Here are two samples, though there are many articles out there;

"Since marijuana legalization, highway fatalities in Colorado are at near-historic lows", The Washington Post


My question, do studies suggest legalization of marijuana results in a net increased or decrease in DUI fatalities?

For this question DUI includes deaths due to alcohol, marijuana, or any other substance that can cause a DUI.

I am looking at all states which legalized marijuana, not any one state. In fact I would accept studies from locations outside of the US if they were relevant in predicting the affect on DUI fatalities that would occur from a country legalizing marijuana.

I'd be interested in the affect of non-fatal accidents as well if found, though I only require information on fatalities for this question.

  • 3
    Let's be very clear: are you asking about correlation or causation? The question of how fatalities changed after legalization is a question of fact, which we could answer by simply looking at data. The question of whether this was caused by legalization is entirely different, would require careful analysis for confounding factors, and is probably impossible to establish given that we have a sample size of n=4. Commented Sep 30, 2016 at 1:04
  • 1
    The Washington Post article is careful to avoid making any claims about causation. The USA Today article says "causing" in its headline but nothing in the text of the article seems to support that conclusion. Commented Sep 30, 2016 at 1:07
  • 1
    @NateEldredge I'm asking to have my cake and eat it too! Which is to say if someone can find a study which corrects for enough variables to give a reliable evidence of causation I would love that. Since I doubt we have sufficient data and analysis to prove causation yet I'll settle for correlation, but I phrased it as i did to indicate I'd prefer studies that didn't just compare pre & post legalization fatalities but also attempted to control for any other variables that may have caused a change to get a better idea of the affect of legalization, even if not enough to prove causation
    – dsollen
    Commented Sep 30, 2016 at 1:17
  • The Washingtom Post article only cites data for Colorado; the Today article, only for Washington state. I'm not sure if either of those (or both together) constitute a notable claim for, say, the U.S. at large. Commented Sep 30, 2016 at 14:09
  • Please rephrase the question to remove the notion that marijuana was legalized in the U.S., otherwise we're talking about a ficticious reality, and we can't answer this for you. Example) change "Did DUI fatalities increase or decrease due to legalization of marijuana?" to "Did DUI fatalities increase or decrease due passage of null marijuana laws?"
    – user34490
    Commented Aug 11, 2017 at 4:24

1 Answer 1


We don't know.

The study cited for all the articles about the statistics for the state of Washington was one produced by the AAA Foundation for Traffic Safety.

Prevalence of Marijuana Involvement in Fatal Crashes: Washington, 2010-2014

The paper, itself, is quick to point out that it is not attributing cause, just looking at the incidence levels.

Some facts in the paper -

10% of all drivers in that period (2010-2014, legalization was in December 2012) had THC in their system.

  • 34% of those with THC did NOT have alcohol or other drugs in their system

  • 39% had alcohol, only, in addition to THC

  • 16.5% had other drugs in addition to THC but no alcohol

  • 10.5% had other drugs and alcohol in addition to THC

You can see that determining cause would be very difficult to do. There is also no cut and dried measurement for how much THC impairs you, as there is with alcohol.

There are also factors to consider - alcohol impairs you and lowers inhibitions, which would be an especially bad combination for driving. The noted marijuana side-effect of paranoia might cause someone to be more cautious when operating, which might mitigate the risks of impairment. I'm not stating this as fact, but pointing out how complicated teasing out cause can be for something like this.

This complication is noted in the study, as well. They recommend a two-tiered approach to safeguarding public safety, where a positive THC test at certain levels, in addition to a separate impairment test, as the best route moving forward.

  • Hand typed the study url, let me know if the link isn't working, and I'll re-check my syntax. Commented Sep 30, 2016 at 17:33
  • Also, THC is detectable at levels far lower than what is needed to produce any effect.
    – jamesqf
    Commented Aug 9, 2017 at 18:20
  • 1
    Unfortunately, like Drunk Driving, these studies count ANY ONE with THC as a Marijuana related accident even if the Marijuana had nothing to do with the accident (i.e. if a driver drank a beer then was rear-ended at a stop light by a non-DUI driver, it would still count as an alcohol-related accident). These discrepancies help over-inflate the numbers. Commented Aug 9, 2017 at 18:55
  • 1
    @HannoverFist - This study makes a point of not claiming that the marijuana had anything to do with cause, it's just recording incidence (was present) for accidents where people were tested. Commented Aug 9, 2017 at 19:10

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .