Experimental, epidemiological, and real-case studies have different advantages and limitations when used to study the effect of substance use on the risk for involvement in a road traffic crash. It is easier to perform well-controlled experimental studies than well-controlled epidemiological studies due to difficulties related to selection bias, information bias, and confounding. On the other hand, it is difficult or impossible to perform experimental studies using single and repeated substance doses similar to those used by drivers and problematic drugs users. Real-case studies indicate which substances may cause observed impairment and involvement in road traffic crashes and at which concentrations; however, those studies cannot be used t...