Marijuana use and driving

Hindrik W.J. Robbe

Institute for Human Psychopharmacology, University of Limburg,
P.O. Box 616, 6200 MD Maastricht, The Netherlands

Robbe, Hindrik W.J. , 1994. Marijuana use and driving. Journal of the International Hemp Association 1: 44-48.

This article concerns the effects of marijuana smoking on actual driving performance. It presents the major results of one laboratory and three on-road driving studies. The latter were conducted on a closed section of a primary highway, on a highway in the presence of other traffic and in urban traffic, respectively. This program of research has shown that marijuana produces only a moderate degree of driving impairment which is related to the consumed THC dose. The impairment manifests itself mainly in the ability to maintain a steady lateral position on the road, but its magnitude is not exceptional in comparison with changes produced by many medicinal drugs and alcohol. Drivers under the influence of marijuana retain insight into their performance and will compensate where they can (e.g., by increasing distance between vehicles or increasing effort). As a consequence, THC's adverse effects on driving performance appeared relatively small in the tests employed in this program.


After alcohol, delta-9-tetrahydrocannabinol (THC), marijuana's major psychoactive constituent, is the drug which is found most often in the blood of drivers involved in road accidents. With some exceptions, epidemiological studies indicate the presence of THC in roughly 4-12% of drivers injured or killed in traffic accidents: for example, 10% in New York (Terhune 1982), 7% in New South Wales (Chesher and Starmer 1983), 8% in North Carolina (Mason and McBay 1984), 11% in DŸsseldorf (Daldrup et al. 1987), 10% in Tasmania (McLean 1987) and 11% in Ontario (Cimbura et al. 1990). The most recent data regarding the incidence of drugs in fatally injured drivers in the United States are available from a nationwide study conducted in 1990 and 1991 (Terhune et al. 1992): THC was found in only 4% of the drivers. This relatively low percentage may indicate a declining trend in the incidence rate of THC in fatally injured drivers in the United States, explainable by the declining prevalence of marijuana use throughout the 1980s (Johnston et al. 1992).

These numbers are, however, inconclusive regarding THC's contribution to accidents because alcohol has been a severe confounding factor in all surveys of injured or killed drivers: generally 60-80% of drivers who were found positive for THC also showed the presence of alcohol. Another problem of these surveys is the common lack of sound control groups. To determine whether drivers under the influence of THC are overrepresented in accidents, the THC incidence in accident victims should be compared to the THC incidence in randomly selected drivers passing the same accident site at the same times and days of the week. This has been done for alcohol (Borkenstein et al. 1974), but not for THC.

The lack of separate control groups has been circumvented by the use of a 'culpability index'. This index is the ratio of the percentage of drivers with detectable drug levels and deemed responsible for a traffic accident to that of drug-free drivers from the same sample who were likewise culpable. But results from this approach have not been consistent: three studies yielded a culpability index of about 1.7 (Warren et al. 1981, Terhune 1982, Donelson et al. 1985), two other studies failed to find a significantly elevated culpability index for marijuana users (Williams et al. 1985, Terhune et al. 1992). For this and other reasons given above, the independent contribution of THC to accidents remains exceedingly obscure. Several literature reviews, the most recent by Robbe (1994), have shown that the results from driving simulator and closed-course tests indicate that THC in inhaled doses up to 250 g/kg has relatively minor effects on driving performance, certainly less than blood alcohol concentrations (BACs) in the range 0.08-0.10 g%. In contrast to this, laboratory studies have repeatedly shown performance impairment occurring after inhaled doses as low as about 40 g/kg. These became large and persistent after 100-200 g/kg doses. Tracking, divided attention and vigilance test performance were particularly vulnerable to THC's effects. Assuming that both sets of results are valid for the particular circumstances under which they were obtained, they demonstrate that performance decrements obtained under the artificial and non-life threatening conditions in the laboratory do not automatically predict similar decrements in driving simulations that are closer to real-world driving. These conflicting results led, however, to opposing opinions regarding marijuana's effects on driving performance. Real-world driving studies were therefore warranted.

Only one study (Klonoff 1974) had been conducted in actual traffic before the present research program started. In a city driving study, Klonoff assessed the effects of two THC doses, 4.9 and 8.4 mg, which are equivalent to 70 and 120 g/kg for a 70 kg person. Aspects of subjects' driving performance were scored by a professional examiner. The results showed that subjects performed less competently when under the influence of the highest, but not the lowest dose. In particular, they scored lower on judgment and concentration scales. Several investigators, however, criticized the method used by Klonoff for measuring driving performance on the grounds that the examiners' reliability was never determined and that the scoring instrument had never been shown to provide measures related to driving safety (Moskowitz 1985, Smiley 1986). Furthermore, Klonoff administered relatively low THC doses to his subjects. The effects of high doses of THC on driving in real traffic still needed to be determined.

The studies reported in this article were conducted to escape these limitations. First, the effects of low, moderate and high THC doses on highway driving were determined, both in the absence and presence of other traffic. Second, Klonoff's city driving study was replicated, with some modifications with regards to the employed procedures and with the addition of another group of subjects who undertook the same driving test but then under the influence of a low dose of alcohol.

An important practical objective of the program was to determine whether degrees of driving impairment can be accurately predicted from either measured concentrations of THC in plasma or performance measured in potential roadside 'sobriety' tests of tracking ability or hand and posture stability. The results (Robbe 1994), like many reported before, indicate that none of these measures accurately predicts changes in actual driving performance under the influence of THC.

General procedures

Subjects in all studies were recreational users of marijuana or hashish, i.e., smoking the drug more than once a month, but not daily. They were all healthy, between 21 and 40 years of age, had normal weight and binocular acuity, and were licensed to drive an automobile. Furthermore, law enforcement authorities were contacted, with the volunteers' consent, to verify that they had no previous arrests or convictions for drunken driving or drug trafficking.

Each subject was required to submit a urine sample immediately upon arrival at the test site. Samples were assayed qualitatively for the following common 'street drugs' (or metabolites): cannabinoids, benzodiazepines, opiates, cocaine, amphetamines and barbiturates. In addition, a breath sample was analyzed for the presence of alcohol using a Lion-SD3 breath-analyzer.

Subjects were accompanied during every driving test by an licensed driving instructor. A redundant control system in the test vehicle was available for controlling the car, should emergency situations arise.

Marijuana and placebo marijuana cigarettes were supplied by the U.S. National Institute on Drug Abuse. The lowest and highest THC concentrations in the marijuana cigarettes used in the studies were 1.75% and 3.57%, respectively.

Pilot study


The pilot study was conducted in a hospital under strict medical supervision to identify THC doses that recreational marijuana users were likely to consume before driving. Twenty-four subjects, twelve males and twelve females, participated. They were allowed to smoke part or all of the THC content in three cigarettes until achieving the desired psychological effect. Cigarettes were smoked through a plastic holder in a manner determined by the subjects. The only requirement was to smoke for a period not exceeding 15 minutes. When subjects voluntarily stopped smoking, cigarettes were carefully extinguished and retained for subsequent gravimetric estimation of the amount of THC consumed.


Six subjects consumed one cigarette, thirteen smoked two and four smoked three. The average amount of THC consumed was 20.8 mg, after adjustment for body weight, 308 g/kg. It should be noted that the amounts of THC consumed represent both the inhaled dose and the portion that was lost through pyrolysis and side-stream smoke during the smoking process. There were no significant differences between males and females, nor between frequent and infrequent users, with respect to the weight adjusted preferred dose. It was decided that the maximum dose for subsequent driving studies would be 300 g/kg. This is considerably higher than doses that have usually been administered to subjects in experimental studies (typically, 100200 g/kg THC).

Study 1: Driving on a restricted highway


The first driving study was conducted on a highway closed to other traffic. One objective of this study was to determine whether it would be safe to repeat the study on a normal highway in the presence of other traffic. The second objective was to define the dose-effect relationship between inhaled THC dose and driving performance. The same twelve men and twelve women who participated in the laboratory study served again as the subjects. They were treated on separate occasions with marijuana cigarettes containing THC doses of 0 (placebo), 100, 200, and 300 g/kg. Treatments were administered double-blind and in a counterbalanced order. On each occasion, subjects performed a road-tracking test beginning 40 minutes after initiation of smoking and repeated one hour later. The test, developed and standardized by O'Hanlon et al. (1982, 1986), involved maintaining a constant speed at 90 km/h and a steady lateral position between the delineated boundaries of the traffic lane. Subjects drove 22 km on a primary highway and were accompanied by a licensed driving instructor. The primary dependent variable was the standard deviation of lateral position (SDLP), which has been shown to be both highly reliable and very sensitive to the influence of sedative drugs and alcohol. Other dependent variables were mean speed, and standard deviation of speed and steering wheel angle.


All subjects were willing and able to finish the driving tests without great difficulty. The study demonstrated that marijuana impairs driving performance as measured by an increase in SDLP; all three THC doses significantly affected SDLP relative to placebo. The driving performance decrement after smoking marijuana persisted almost undiminished for two hours after smoking. Marijuana's effects on SDLP were compared to those of alcohol obtained in a very similar study by Louwerens et al. (1985, 1987). It appeared that the effects of the various administered doses of THC (100300 g/kg) on SDLP were equivalent to those associated with BACs in the range of 0.03-0.07 g%. Other driving performance measures were not significantly affected by THC. Both the observed degree of driving impairment, and what subjects said and did, indicated that normal safeguards would be sufficient for ensuring safety in further testing. Hence, the final conclusion was to repeat this study on a normal highway in the presence of other traffic.

Study 2: Driving on a normal highway in traffic


The second driving study was conducted on a highway in the presence of other traffic and involved both a road-tracking and a car-following test. A new group of sixteen subjects, equally comprised of men and women, participated in this study. A conservative approach was chosen in designing the present study in order to satisfy the strictest safety requirements. That is, the study was conducted according to an ascending dose series design where both active drug and placebo conditions were administered, double-blind, at each of three THC dose levels. THC doses were the same as those used in the previous study, namely 100, 200, and 300 g/kg. Cigarettes appeared identical at each level of treatment conditions. If any subject had reacted in an unacceptable manner to a lower dose, he/she would not have been permitted to receive a higher dose.

The subjects began the car-following test 45 minutes after smoking. The test was performed on a 16km stretch of the highway and lasted about 15 minutes. After the conclusion of this test, subjects performed a 64-km road-tracking test on the same highway which lasted about 50 minutes. At the conclusion of this test, they participated again in the car-following test.

The road-tracking test was the same as in the previous study except for its duration and the presence of other traffic. Subjects were instructed to maintain a constant speed of 95km/h and a steady lateral position between lane boundaries in the right traffic lane. They were allowed to deviate from this only if it became necessary to pass a slower vehicle in the same lane. Data from the standard test were analyzed to yield the same performance measures as in the previous study.

The car-following test involved attempting to match velocity with, and maintain a constant distance from a preceding vehicle as it executed a series of deceleration/acceleration maneuvers. The preceding vehicle's speed would vary between 80 and 100 km/h and the subject was instructed to maintain a 50 m distance however the preceding vehicle's speed might vary. The duration of one deceleration and acceleration maneuver was approximately 50 seconds and six to eight of these maneuvers were executed during one test, depending upon traffic density. The subject's average reaction time to the movements of the preceding vehicle, mean distance and coefficient of variation of distance during maneuvers were taken as the dependent variables from this.


All subjects were able to complete the series without suffering any untoward reaction while driving. Road-tracking performance in the standard test was impaired in a dose-related manner by THC and confirmed the results obtained in the previous closed highway study. The 100 g/kg dose produced a slight elevation in mean SDLP, albeit not statistically significant. The 200 g/kg dose produced a significant elevation, of dubious practical relevance. The 300 g/kg dose produced a highly significant elevation which may be viewed as practically relevant but unexceptional in comparison with similarly measured effects of many medicinal drugs. After marijuana smoking, subjects drove with an average speed that was only slightly lower than after a placebo and very close to the prescribed level.

In the car-following test, subjects maintained a distance of 4550 m while driving in the successive placebo conditions. They lengthened mean distance by 8, 6 and 2 m in the corresponding THC conditions after 100, 200 and 300 g/kg, respectively. The initially large drug-placebo difference and its subsequent decline is a surprising result. Our explanation for this observation is that the subjects' caution was greatest the first time they undertook the test under the influence of THC and progressively less thereafter. The reaction time of the subjects to changes in the preceding vehicle's speed increased following THC treatment, relative to placebo. The administered THC dose was inversely related to the change in reaction time, as it was to distance. However, increased reaction times were partly due to longer distance. Statistical adjustment for this confounding variable resulted in smaller and non-significant increases in reaction time following marijuana treatment, the greatest impairment (0.32 s) being observed in the first test following the lowest THC dose. Distance variability followed a similar pattern as mean distance and reaction time; the greatest impairment was found following the lowest dose.

Study 3: Driving in urban traffic


The program proceeded to the third driving study, which involved tests conducted in high-density urban traffic. There were logical and safety reasons for restricting the THC dose to 100 g/kg. It was given to a group of regular marijuana (or hashish) users, along with a placebo. For comparative purposes, another group of regular alcohol users was treated with a modest dose of its members' preferred recreational drug, ethanol, or a placebo, before undertaking the same city driving test. Two groups participated, each composed of sixteen new subjects comprising equal numbers of men and women. Subjects in the alcohol group were regular users of alcohol, but not marijuana. Both groups were treated on separate occasions with the active drug and a placebo. Marijuana was administered to deliver 100ʵg/kg THC. The driving test commenced 30 minutes after smoking. The alcohol dose was chosen to yield a BAC approaching 0.05 g% when the driving test commenced 45 minutes after onset of drinking. Active drug and placebo conditions were administered double-blind and in a counterbalanced order in each group.

Driving tests were conducted in daylight over a constant 17.5 km route within the city limits of Maastricht. Subjects drove their placebo and active-drug rides through heavy, medium and low density traffic on the same day of the week, and at the same time of day. Two scoring methods were employed in the present study. The first, a 'molar' approach, required the driving instructor acting as the safety controller during the tests to rate the driver's performance retrospectively using a standard scale. The second, a more 'molecular' approach, involved the employment of a specially trained observer who applied simple and strict criteria for recording when the driver made or failed to make each in a series of observable responses at predetermined points along a chosen route.


The study showed that a modest dose of alcohol (BAC = 0.034 g%) produced a significant impairment in city driving, as measured by the molar approach, relative to a placebo. More specifically, alcohol impaired both vehicle handling and traffic maneuvers. Marijuana, administered in a dose of 100ʵg/kg THC, on the other hand, did not significantly change mean driving performance as measured by this approach. Neither alcohol nor marijuana significantly affected driving performance measures obtained by the molecular approach, indicating that it may be relatively insensitive to drug-induced changes.

Driving quality, as rated by the subjects, contrasted with observer ratings. Alcohol impaired driving performance according to the driving instructor, but subjects did not perceive it; marijuana did not impair driving performance, but the subjects themselves perceived their driving performance as such. Both groups reported about the same amount of effort in accomplishing the driving test following a placebo. Yet only subjects in the marijuana group reported significantly higher levels of invested effort following the active drug. Thus there is evidence that subjects in the marijuana group were not only aware of their intoxicated condition, but were also attempting to compensate for it. These seem to be important findings. They support both the common belief that drivers become overconfident after drinking alcohol and investigators' suspicions that they become more cautious and self-critical after consuming low doses of THC, as smoked marijuana.


The results of the studies corroborate those of previous driving simulator and closed-course tests by indicating that THC in inhaled doses up to 300 g/kg has significant, yet not dramatic, dose-related impairing effects on driving performance. Standard deviation of lateral position in the road-tracking test was the most sensitive measure for revealing THC's adverse effects. This is because road-tracking is primarily controlled by an automatic information processing system which operates outside of conscious control. The process is relatively impervious to environmental changes, but highly vulnerable to internal factors that retard the flow of information through the system. THC and many other drugs are among these factors. When they interfere with the process that restricts SDLP, there is little the afflicted individual can do by way of compensation to restore the situation. Car-following and, to a greater extent, city driving performance depend more on controlled information processing and are therefore more accessible for compensatory mechanisms that reduce the decrements or abolish them entirely.

THC's effects after doses up to 300 g/kg never exceeded alcohol's at BACs of 0.08 g% and were in no way unusual compared to many medicinal drugs (Robbe 1994). Yet THC's effects differ qualitatively from many other drugs, especially alcohol. Evidence from the present and previous studies strongly suggests that alcohol encourages risky driving whereas THC encourages greater caution, at least in experiments. Another way THC seems to differ qualitatively from many other drugs is that the former's users seem better able to compensate for its adverse effects while driving under the influence.

Although THC's adverse effects on driving performance appeared relatively small in the tests employed in this program, one can still easily imagine situations where the influence of marijuana smoking might have a dangerous effect; i.e., emergency situations which put high demands on the driver's information processing capacity, prolonged monotonous driving, and after THC has been taken with other drugs, especially alcohol. Because these possibilities are real, the results of the present studies should not be considered as the final word. They should, however, serve as the point of departure for subsequent studies that will ultimately complete the picture of THC's effects on driving performance.