Are 20% of automobile drivers under the influence of marijuana?

Question

An article by Quinton James and Johnkeria Kinglocke titled "The Truth About Marijuana" published on the website of the Hanley Foundation (an anti-drug abuse organization in Florida) claims (my emphasis):

Would you allow someone to drive you while they are high? According to the Department of Children and Families (DCF), marijuana-impaired driving fatalities have more than doubled since marijuana use has become legalized. Did you know one in five drivers are under the influence of marijuana?

It is unclear if the authors intended to source their claim that 20% of drivers are under the influence of cannabis on research from DCF (or rather was sourcing DCF only to support the doubled fatalities claim), but, in any event, no DCF resources are cited in the article's bibliography.

Are 20% of automobile drivers under the influence of marijuana?

Discussion:

I would consider a claim that 20% of drivers have some trace of cannabis metabolites or other evidence of past drug exposure in their system that could be detected with a sufficiently sophisticated medical laboratory to be plausible, but that is not what the claim is. Rather, I find the idea that 20% of drivers one encounters on the road on any random day, time, or place are literally so whacked out on weed that they cannot exercise ordinary safe driving practices to be absurdly high (pun intended).

I did find a 2019 article "Driving Under the Influence of Marijuana and Illicit Drugs Among Persons Aged ≥16 Years — United States, 2018" from the US Centers for Disease Control (CDC) that reports that it found that 4.7% of those 16 or above reported driving under the influence of marijuana in the past year. 4.7% is much lower than 20%, and still even isn't the same claim. The CDC is reporting on any instance of stoned driving in the past year even if the person drove sober (or not at all) on the other 364 days of the year, while the Hanley Foundation appears to claim that 20% of drivers on the road right now (or on average) are high.

I assume the 20% figure is specifically for the US, not globally? — F1Krazy, Nov 21 '22 at 21:08
Does that article actually say how many marijuana related fatalities there are? — Joe W, Nov 21 '22 at 21:09
I suspect that any article that starts with 'The Truth About ...' is likely to be deliberately false. As for being able to detect drugs - there was a famous episode of Blue Peter (a British Children's programme) where one of the presenters ate a breadroll with poppy seeds on, and they could detect opium in her urine afterwards. I doubt anybody would claim she was under the influence. — j4nd3r53n, Nov 22 '22 at 10:24
@j4nd3r53n that's exactly my point. There is typically a social, medical, and legal distinction between having any trace of a substance in your body and being under the influence of it. If I smoked a blunt yesterday morning and get in the car to drive to Memphis tonight, I probably still have detectable THC levels in my body, but I'm no longer "high", "stoned", or "under the influence" enough for it to materially affect my driving. — Robert Columbia, Nov 22 '22 at 11:39
You probably need a lot less than "so whacked out on weed" to have your driving impaired, like you don't need to be completely drunk to suffer the same problem. The actual degree of impairment and its correlation with quantities and when they were consumed is a complex problem, with a lot of debate and uncertainties. See https://en.wikipedia.org/wiki/Cannabis_and_impaired_driving — jcaron, Nov 22 '22 at 12:12
@jcaron true, I was exaggerating a little bit to emphasize my point that there is a difference between having detectable amounts of THC in your system and being impaired by it. — Robert Columbia, Nov 22 '22 at 12:21
I wonder whether it's 1 in 5 drivers, or 1 in 5 drivers who are stopped by a cop. — Kyralessa, Nov 22 '22 at 15:15
Separate fallacy: A doubling of a tiny number is still tiny. If there were three marijuana-related fatalities a year before (number pulled out of thin air), and six the year after, that's a doubling - but compared to 20% of drivers, it's so tiny as to be a rounding error. If they don't cite a specific number, then a relative change is meaningless. — Bobson, Nov 22 '22 at 15:52
@F1Krazy I assume that the 20% figure is specifically for Florida and not the US. Why are you scoping the source larger than it appears to be. OP, I would question you as well. The Hanley Foundation is based in Florida and the DCF is a Floridian agency. Why are you, or are you, expanding the scope of the source beyond its bounds? The last sentence of the first paragraph in the source also bounds it to Florida, not the US as a whole. — CGCampbell, Nov 22 '22 at 18:08
Or maybe the intended quote was supposed to be "1 in 5 *has driven* under the influence". The claim as quoted is superbly unbelievable, since 1 in 5 drivers are bad enough already that I don't think the weed would hurt. — , Nov 22 '22 at 19:29
@CGCampbell If 20% of Florida drivers were under the influence of Marijuana, that would really be alarming -- after all, this would mean that 80% of Florida drivers are under the influence of meth ... :) — Hagen von Eitzen, Nov 22 '22 at 22:23
I actually ran into the same problem after listening to a presentation by Mothers Against Drunk Driving, which made the claim "1 in 3 drivers on the road with you when you drive to work in the morning are under the influence of alcohol." The source for this data? I think you could be generous and call it a misunderstanding of statistics. But more accurately you could call it manipulation to try and win a point. — CaptainSkyfish, Nov 22 '22 at 22:46
@j4nd3r53n the limit was increased in 1998 because of that reason, so it's fairly irrelevant now, though I guess it depends on location and what the limit is in that area. — Aequitas, Nov 22 '22 at 23:53
"marijuana-impaired driving fatalities have more than doubled since marijuana use has become legalized" - that's the claim. Forget "the truth about"... it has *"Did you know"* ffs. - When did we start testing participants of an accident for THC? I should be on the list of "reported driving under the influence" but that's not something I'll ever admit to. Doubtful that it's us 40-somethings getting high AF and in accidents, because we've been at this for two decades, *on the DL*. Presumably it's the 16yo's who are also dumb enough to admit using drugs to anyone who asks. — Mazura, Nov 23 '22 at 01:49
@RobertColumbia - my guess is they mean 1 in 5 **in accidents** and it's just badly written. Simple as that. — Fattie, Nov 23 '22 at 12:39
"The Florida Department of Children and Families (DCF)... provides social services to children, adults, refugees, domestic violence victims, human trafficking victims, the homeless community, child care providers, disabled people, and the elderly" — https://en.wikipedia.org/wiki/Florida_Department_of_Children_and_Families Difficult to imagine why they'd be compiling statistics about driving fatalities in the first place. It's _almost_ as if whoever wrote this just wanted to lie about statistics with the words "children" and "families" in the mix. — Paul D. Waite, Nov 24 '22 at 10:09

score 47 · Answer 1 · answered Nov 21 '22 at 23:11

47

In short: No

I first tried to find the article's own source, however the Florida Department of Families and Children did not have any such information that I could find. There are other Departments of Families and Children, but the article mentions Florida specifically several times so that seems like the most relevant one.

So I went looking for other sources. The best data I can find for driving under the influence of marijuana in the US comes from the NHTSA's National Roadside Survey. The most recent NRS was conducted across 2013-2014, and the results can be accessed here (the relevant portions are under Volume III: Drug Results). That survey found national numbers for drivers with a positive test for THC at 8.7% during the day and 12.6% at night. Certainly not "one in five", but probably a lot higher than people would expect. It may be worth noting that the category of "Nighttime drivers in the Midwest" did reach 20.5% according to that survey, but that seems like an extremely cherry-picked statistic for an article written about Florida if that's their basis.

However, the NRS doesn't tell the whole story. A 2017 report to Congress on the issue of Marijuana-impaired driving mentions several key factors complicating the collection of data on the subject including references to that same NRS report. In particular, THC remains detectable in blood tests long after it stops impairing their driving; a study they cite even showed a chance for detection 30 days after use. Additionally, further studies cited in the report show that THC concentration in the blood is not closely related to the level of impairment reported subjectively or measured in tests.

Finally, taking a broader look the Substance Abuse and Mental Health Services Administration puts out reports about drug use in general. The 2019 report shows that responses indicating "used marijuana at least once in the last year" were at 17.5%. Considering that number, "20% of drivers" seems at least nearly impossible.

In conclusion, according to the national data I could find, it's beyond unlikely that 20% of drivers are under the influence of marijuana. Using some blatant cherry-picking and extremely loose definitions (e.g. "anybody who uses marijuana regularly is 'influenced' by it even if they aren't currently impaired by it") they might be able to come up with a "justification" for the claim, but that wouldn't make it true (and they haven't done so in the article).

It's possible that the authors are referencing some Florida-specific data which I haven't found, but all their actual citations are from national sources which do not support this specific claim.

answered Nov 21 '22 at 23:11

Kamil Drakari

2,401
1
15
21

2

Given that 17.5% "used marijuana at least once in last year" number, yeah, I wonder if these folks are inventing narratives like "they get high, get the munchies, and drive to the store for snacks" to claim that drug impairment is higher in samples of people currently behind a wheel than in samples of the general population. But fortunately your NHTSA data rules that out (at least for getting to 20%) by sampling actual drivers behind the wheel. – Peter Cordes Nov 22 '22 at 05:38
23

Background question on the NHTSA survey: Do they only test people for drugs and alcohol whose driving already looked suspicious enough to pull them over (meaning most people tested will test positive on either alcohol or some drug) or is that a sample over all drivers so that the vast majority of drivers should test as clean and fit to drive? – quarague Nov 22 '22 at 08:03
8

note that such numbers exclusive reflect the results of tests done in highly specific locations, at specific times. And those times and locations are selected to coincide with areas and times when high rates of abuse are known to happen. E.g. on the exit roads of festivals and bar/pub areas towards residential and hotel areas. So the percentage of impaired drivers will be significantly higher than that of the general population. – jwenting Nov 22 '22 at 08:33
@quarague see my previous comment. It reflects targeted traffic checks (whether blanket checks or point checks of suspicious drivers) in areas and times with a high impairment rate. – jwenting Nov 22 '22 at 08:34
2

Regarding the NRS survey in 2013-14, it's worth noting too that numbers in that survey still don't support the claim "*since marijuana use has become legalized*". Whilst 2012-14 was when legalisation started rolling (ahem!), the majority of states still did not allow recreational marijuana use by the end of 2014. – Graham Nov 22 '22 at 11:55
9

@quarague The methodology section of the report indicates that they set up survey stations at randomly selected locations, and pulled drivers out of traffic randomly. I haven't read everything to check whether they have additional mechanisms setup to avoid biased data collection, but one thing that could impact results is that surveys were only performed in locations where local law enforcement was willing to cooperate. I doubt it's perfect data, but they at least claim that the sampling is random. – Kamil Drakari Nov 22 '22 at 16:24
"THC remains detectable in blood tests long after it stops impairing their driving" - that in itself needs a citation – Stewart Nov 22 '22 at 16:38
6

@Stewart My source is the one linked in that paragraph, specifically this sentence from the bottom of page 4: "Thus, while THC can be detected in the blood long after ingestion, the acute psychoactive effects of marijuana ingestion last for mere hours, not days or weeks" – Kamil Drakari Nov 22 '22 at 16:43
2

@PeterCordes: I wouldn't be surprised if nighttime drivers were somewhat more likely to be marijuana users; young people are more likely to use marijuana, and [young people also tend to be chronically sleep-deprived](https://startsleeping.org/statistics/) (read: there are more likely to be up and about at night). The correlation between "people who use drugs/alcohol" and "people who are still on the roads well after dark" seems likely to be significant; the 20.5% figure could be correct simply due to sampling bias (as noted, the checks were targeted for times/locales with high rates of abuse). – ShadowRanger Nov 22 '22 at 17:22
This doesn't mean the positive tests correspond to actual impairment of course; as repeatedly noted, THC metabolites last a *long* time, and there's no strong correlation between the detected levels and actual impairment. That's really been one of the arguments against marijuana legalization all along. Tests for alcohol impairment are easy and fairly specific (they don't test positive when you were drunk last night, and they're correlated with level of impairment). Marijuana tests provide neither feature, so it's harder to detect/punish impairment in safety-critical scenarios. – ShadowRanger Nov 22 '22 at 17:28
Mind you, [it does appear that the need for such detection/punishment may be fairly limited](https://skeptics.stackexchange.com/q/17407/38242); the impairment doesn't seem to last as long (metabolizes exponentially, based on a half-life, where alcohol metabolizes linearly), and even while impaired, they recognize the impairment and tend to drive more slowly/cautiously to compensate (where drunk drivers frequently do the opposite). Just wanted to make that clear, since the testing issue is one of the strongest objective arguments against legalization, but in practice it may not matter much. – ShadowRanger Nov 22 '22 at 17:43
@Graham "In 2014, the Florida Legislature passed the Compassionate Use Act which was the first legal medical cannabis program in the state's history. The original Compassionate Use Act only allowed for low-THC cannabis (Charlotte's Web strain) to be dispensed and purchased by patients suffering from cancer and epilepsy." I wonder if many here are turning this into a US National scope, instead of what appears to have meant to be a Florida-specific one. Note that I don't know if Florida has a bigger or smaller percentage of THC users than other states. – CGCampbell Nov 22 '22 at 18:02
7

@quarague There's a huge built in bias. 1) The daytime samples were taken when a lot of people would be at work (and thus not driving). 2) Also, it looks like the samples were done as "paid voluntary survey" with roadside signs advertising (so you wouldn't get anyone who's actually driving as part of their work). So you have a bias towards people driving around in the middle of the day who have time/inclination to pull over to do a survey. I'm not sure how that impacts the results, but I'm pretty sure that's not an average group of drivers. – user3067860 Nov 22 '22 at 18:36
17.5% is what people answered on a survey, the actual percentage is likely to be significantly higher. Surveys are valuable but they're not an accurate method of determining prevalence. – President James K. Polk Nov 22 '22 at 22:36
Note that the number with detectable metabolites can be higher than the number who used it. I wouldn't be surprised if I had detectable metabolites all the time through high school even though I don't smoke, period--I had to pass close enough to the crowd of smokers that I would smell it. If I can smell it some must be ending up in my system. – Loren Pechtel Nov 23 '22 at 04:15
2

@jwenting RE selection bias.: True enough. If you test samples from the 7.30 a.m. mid-week morning commute I'm sure the positive rate would be close to zero ;-). (But you might find an elevated cocaine rate in cities with a large financial district.) – Peter - Reinstate Monica Nov 23 '22 at 07:36
@KamilDrakari Thank you. I found the text which is the citation. I have to admit, I don't understand it. It says, "THC is stored in fatty tissues in the body and can be released back into the blood sometimes long after ingestion." To me, this says psychoactive effects are totally possible 30 days later. The missing datum is how much THC does it take to cause "psychoactive effects"? (assuming said effects have a clear-cut definition) Continuing the comparison with alcohol made in the study, it's said that ANY alcohol in your blood affects driving. Is the same not true for THC? – Stewart Nov 24 '22 at 20:50

Are 20% of automobile drivers under the influence of marijuana?

1 Answers1

In short: No