This paper describes a study which investigated the effects of alcohol and marijuana, alone and in combination, on driver performance and behavior in a fully interactive driving simulator. The simulator provided the driver a complex visual scene similar to a rural nighttime drive, and allowed the driver full control of steering and speed maneuvers. Performance and behavior data were collected during a 10-12 mile drive requiring about 15 minutes to complete. A variety of events were encountered during the drive, including curves, obstacles in the roadway, and winding roads. Accidents, tickets, and speed were recorded as traffic safety measures during the overall drive. Driver behavior, speed control, and steering performance were collected during each event to provide insight into the impairment mechanisms of alcohol and/or marijuana on the driver.

A full placebo experimental design was employed which included all combinations of 3 marijuana (0, 100, and 200 mg THC/kg body weight) and 2 alcohol (0 and 0.10 percent BAC) levels. Based on a large number of driver performance and behavior variables, alcohol was found to have a pervasive and significant impairing effect. Simulator accidents increased reliably under alcohol, which was accounted for by increased steering and speed control variability. Marijuana effects were minimal, the primary one being speed reduction. This speed reduction, while statistically reliable, was minimal in terms of actual driving behavior and is probably of no practical significance. A significant drug interaction effect was observed in simulator accidents;​ however, the data do not allow us to identify the impairment mechanism leading to this result.