Marijuana and Driving: What Legalization Did to the Roads
Marijuana impaired driving is the part of legalization no state solved. THC is now the most common drug found in seriously injured drivers, ahead of alcohol. In Washington the share of drivers in fatal crashes who tested positive for THC doubled after retail sales began. In Colorado, deaths involving a THC-positive driver rose 240 percent. There is still no roadside test and no agreed impairment threshold. The record, every number sourced.
Drugs found in seriously or fatally injured road users
Share testing positive, 7,279 patients at seven trauma centers and four medical examiners, 2019 to 2021
55.8 percent tested positive for at least one drug; one in five for two or more.
Does marijuana cause crashes? The evidence, honestly
The industry cites one federal study to say marijuana does not raise crash risk. Here is that study alongside the rest of the literature.
Odds of a crash after marijuana use, by study
Odds ratio versus drug-free drivers (1.0 = no change)
The most conservative meta-analysis still finds a “statistically significant increase in motor vehicle crash risk.” Asbridge found the risk roughly doubled, “especially for fatal collisions.”
Asbridge et al., BMJ, 2012; Rogeberg and Elvik, Addiction, 2016; Hostiuc et al., Frontiers in Pharmacology, 2018; Chihuri, Li and Chen, Injury Epidemiology, 2017
The study the industry cites
NHTSA’s 2015 Virginia Beach case-control study found THC-positive drivers crashed 25 percent more often, an odds ratio of 1.25. After adjusting for age, sex and race the ratio fell to 1.05, which is the number advocates quote. What they leave out: the sample was two-thirds fender-benders with no injuries, and the adjustment removed the effect because, in the report’s words, “if the THC-positive drivers were predominantly young males, their apparent crash risk may have been related to age and gender rather than use of THC.” Young men who use marijuana still crash more. The study does not say THC is safe to drive on.
What a controlled trial shows
UC San Diego gave 191 regular users smoked marijuana or placebo and put them in a driving simulator. Performance was worse at 30 minutes and 90 minutes, borderline at three and a half hours, and back to normal at four and a half. Impairment was “unrelated to THC content, use history, or blood THC concentration.” At 90 minutes, 69 percent of participants said they were ready to drive “despite performance not improving.” The authors: this “may indicate a false sense of driving safety.” NHTSA puts it more simply: “Impaired drivers can’t accurately assess their own impairment.”
Marcotte et al., JAMA Psychiatry, 2022; NHTSA, drug-impaired driving
Before and after: the data from Colorado and Washington
The two states with the longest records keep the best data. Both tell the same story.
Colorado: traffic deaths involving a cannabinoid-positive driver
Among drivers who were drug-tested, the share positive for a cannabinoid rose from 16 percent in 2013 to 29 percent in 2022, and 19 percent of fatalities with a tested driver involved THC at or above the 5 ng/mL inference level, “the highest percentage since 2016.” The state’s caveat: only about half of drivers in fatal crashes are tested, and a positive test shows presence, not impairment.
Colorado Division of Criminal Justice, Impacts of Marijuana Legalization, 2025
Washington: drivers in fatal crashes who were THC-positive
The 2026 state fact sheet: 640 people have died in Washington since 2020 in crashes involving a THC-positive driver, “nearly one in every five traffic deaths.” Among drivers actually tested in 2024, 29 percent were positive for THC, and more than half of those were also positive for alcohol.
AAA Foundation for Traffic Safety, 2020; Washington Traffic Safety Commission, February 2026
Across every legal state
| Study | States | Finding |
|---|---|---|
| IIHS and HLDI, 2021 | CA, CO, NV, OR, WA | “A 6 percent increase in injury crash rates and a 4 percent increase in fatal crash rates” after legalization. IIHS president David Harkey: “legalizing marijuana for recreational use does increase overall crash rates.” |
| Farmer, Monfort and Woods, J Stud Alcohol Drugs, 2022 | Five legal states vs. controls | Legalization associated with “a 6.5% increase in injury crash rates and a 2.3% increase in fatal crash rates”; legalization plus retail sales, 5.8 percent and 4.1 percent. |
| Lane and Hall, Addiction, 2019 | CO, WA, OR and neighbors | “One additional traffic fatality per million residents” in legalizing states and in neighboring jurisdictions. |
| Windle et al., Am J Prev Med, 2022 | 64 studies, North America | “Recreational legalization was conversely associated with increases in fatal collisions.” |
| Myran et al., JAMA Network Open, 2023 | Ontario | Marijuana-involved traffic injury ER visits up 475 percent from 2010 to 2021; the jump came with commercialization, not with legalization alone. |
The pattern is the same one the Impact Report finds everywhere else: the harm tracks commercial sales, not the law change on its own. More stores, more potent products, more drivers who think they are fine.
Teens drive high more often than they drive drunk
of U.S. high school students who currently use marijuana drove after using it in the past month. Among all teen drivers, driving after marijuana (12.7 percent) was “more than double the prevalence of drinking and driving (5.0%).”
Li et al., JAMA Network Open, 2020, CDC Youth Risk Behavior Survey
Americans aged 16 to 25 drove under the influence of marijuana in 2022, compared with 2.4 million who drove under the influence of alcohol.
Colorado drivers aged 18 to 24 said they had driven within two hours of using marijuana, versus 7 percent of older drivers. In Washington, more than 40 percent of young adults who used in the past year drove within three hours of use.
Colorado Department of Transportation, 2024; University of Washington, Young Adult Health Survey
Young drivers are the group most likely to use, most likely to use the highest-potency products, and least likely to believe they are impaired. See youth marijuana use.
Why a legal limit for THC does not work like 0.08
Alcohol has a breathalyzer and a number. Marijuana has neither, and the federal government said so to Congress.
“THC level in blood (or oral fluid) does not appear to be an accurate and reliable predictor of impairment from THC.”
NHTSA, Marijuana-Impaired Driving: A Report to Congress, 2017
On the 5 nanogram limit several states adopted, the same report: “This per se limit appears to have been based on something other than scientific evidence.” THC leaves the blood within hours while impairment can outlast it; a frequent user can be impaired below 5 ng and an occasional user can test above it hours after the effect has faded. That is why SAM opposes non-zero per se limits and supports a permissible inference of impairment at any detectable level of active THC.
Where the states are
Twelve states have zero-tolerance laws for THC (Arizona, Delaware, Georgia, Indiana, Iowa, Michigan, Oklahoma, Pennsylvania, Rhode Island, South Dakota, Utah, Wisconsin). Illinois, Montana, Nevada, Ohio and Washington set per se limits of 2 to 5 ng/mL; Colorado uses 5 ng/mL as a permissible inference. Everywhere else, impairment has to be proven case by case, usually by an officer with drug-recognition training, of whom there are far too few.
Roadside oral-fluid screening is the tool that is coming. In Minnesota’s pilot, 87 percent of drivers screened tested positive for a drug and 62 percent for more than one.
National Conference of State Legislatures; Star Tribune, March 2025
Marijuana does not replace drinking and driving. It multiplies it.
Odds of a fatal crash
Drivers positive for both had higher odds than alcohol alone would predict, “a positive synergistic effect on fatal crash risk.” In Washington, 52 percent of THC-positive drivers in fatal crashes were also positive for alcohol.
Chihuri, Li and Chen, Injury Epidemiology, 2017; Washington Traffic Safety Commission, 2026
What users believe
In a 2025 AAA survey of marijuana consumers, 84.8 percent said they drive the same day they use, and 53 percent within an hour. Only 19 percent thought their driving was worse afterward; 34 percent thought it was better. Only 29 percent believed a police officer could tell. Among all drivers, 70 percent consider driving within an hour of using very or extremely dangerous, against the near-universal view of drunk driving, and only 26 percent think they would be caught.
AAA, Cannabis Users’ Perceptions and Self-Reported Behaviors, 2025; AAA Foundation, 2023 Traffic Safety Culture Index
This is the same pattern as the alcohol argument itself: legal marijuana stacks on top of alcohol rather than replacing it, on the road as everywhere else.
SAM’s drugged-driving agenda
Zero tolerance for active THC. Non-zero per se limits are scientifically invalid and should be avoided; states should instead allow a permissible inference of impairment at any detectable level of active THC in blood or oral fluid.
Roadside oral-fluid screening and toxicology after every serious crash. You cannot fix what you do not measure. About half of Colorado’s fatal-crash drivers are never tested.
More Drug Recognition Experts, and ARIDE training for every DUI officer.
Open-container rules, driver education on THC, and warning labels at the point of sale.
SAM marijuana-impaired driving one-pager (PDF); state policy resources; marijuana laws by state
Why it matters more every year
Weekend nighttime drivers testing positive for THC rose from 8.6 percent in 2007 to 12.6 percent in 2013-14, a 47 percent increase, and that was before most states had retail stores. Products are stronger, use is daily for more people, and the public does not treat high driving the way it treats drunk driving. Alcohol-impaired driving fell because the law, the test and the culture all moved together. Marijuana has none of the three yet.
Marijuana impaired driving: common questions
Does marijuana impair driving?
Yes. In a controlled trial of 191 regular users, driving performance was worse for about three and a half hours after smoking and back to normal at four and a half, regardless of THC dose or blood level, and most participants felt ready to drive long before they were. Meta-analyses find crash odds roughly 1.4 to 2 times higher after use, and combined with alcohol the odds of a fatal crash were 25 times higher than for sober drivers.
Did marijuana legalization increase car crashes?
The best evidence says yes. Across five legal states, injury crash rates rose about 6 percent and fatal crash rates 2 to 4 percent after legalization and retail sales (IIHS and Farmer 2022). In Washington the share of drivers in fatal crashes who were THC-positive doubled after sales began; in Colorado, deaths involving a cannabinoid-positive driver rose from 55 in 2013 to 187 in 2022.
Is there a breathalyzer for marijuana?
No. There is no roadside breath test for THC in use, and NHTSA told Congress that blood or oral-fluid THC levels are not reliable predictors of impairment. Some states are piloting oral-fluid screening devices that detect recent use; impairment is still established by officer observation and drug-recognition evaluation.
What is the legal THC limit for driving?
It depends on the state. Twelve states have zero tolerance for THC, five set per se limits of 2 to 5 ng/mL, Colorado uses 5 ng/mL as a permissible inference, and the rest require proof of impairment. NHTSA has said the 5 ng/mL figure “appears to have been based on something other than scientific evidence,” and SAM supports zero tolerance for active THC instead.
How long after using marijuana is it safe to drive?
The controlled trial evidence shows measurable impairment for at least three and a half hours after smoking, with full recovery at about four and a half hours, and users could not judge their own impairment. Edibles act later and longer. The only safe answer is not to drive the day you use.
Is driving high as dangerous as driving drunk?
Alcohol alone raises fatal crash odds far more than marijuana alone (about 16 times versus 1.5 times in one national analysis), but marijuana is now the most common drug found in seriously injured drivers, it is used before driving far more often than alcohol by teens, and the two together are worse than either: 25 times the odds of a fatal crash.
Related: Marijuana myths and facts · Marijuana potency · Marijuana laws by state · Legalization Impact Report · SAM FAQs
No test, no limit, no excuse.
States that legalized owe their drivers the tools alcohol has had for fifty years. SAM has the model language and the data for every legislature.
