Contributions from Psychology to the Study of Drinking and Driving

The literature does not support the belief that alcohol produces uniform effects on behavior, but indicates that personality, skill and experience influence a person's response to alcohol. In addition, other characteristics (e.g., age and driving experience) influence the effect of alcohol on driving. The assumption of a linear dose—response relationship (i.e., a constant change in effect for a constant change in amount consumed) is naïve; while the linearity may hold at moderate to high blood alcohol levels (e.g., above 0.20%), at levels between zero and some critical concentration (e.g., 0.20%) the assumption has seriously destructive consequences for research. Nonlinear results (differences in degree of change of effect with a constant change in dosage) and even nonmonotonic results (change in direction of effect with change in dosage) have been observed. The pharmacological assumption that alcohol is a depressant is true only at a limited range of blood levels which are much higher than those produced by common social use. The variety of responses to blood alcohol levels obtained in social use makes the “depressant” hypothesis difficult to use as an explanatory principle and, except when supported by direct evidence, it should be avoided. None of the control methods, such as placebo, group interaction, equalvolume variable-concentration, can completely prevent the subject in experiments from knowing that he is getting alcohol and guessing the amount. A dose series administered in random order is probably most satisfactory because it gives the subject more information with which to confuse his guessing. The only near-adequate controls seem to be obtained by intravenous administration of alcohol. Even then, effectiveness is reduced by self-awareness of intoxication and, perhaps, by the information conveyed to the subject by the intravenous procedure itself. Complicating the drinking—driving problem are the unknown numbers of drinkers who also use other drugs. A first step in studying drug combinations is the classification of drugs that would lead to priorities of testing in combination with alcohol. Because of the number of drugs that can occur in combination, theories of joint action leading to mathematical models are essential. Critical combinations arrived at from a model known to be appropriate, “dry run” by computer and tested in subjects to determine their concurrence with expectation, could help make the problem of joint action manageable. Conclusions about drug interaction (e.g., as additive or synergistic) can be made only with precise knowledge of the dose—response relationships of each drug. Minimum experimental conditions must be satisfied before data can be used to evaluate combined action or build test models. These include the number of doses of each drug, kinds of drug combinations that can be used, some of the formal characteristics of responses, consideration of the concept of tolerance (thresholds) and the organism to be used. Unless the study is put together correctly, conclusions about joint action and behavior are pointless and invalid. Laboratory experiments are sufficient for developing models of joint action, but because of lack of knowledge about driving itself, it would be premature to attempt joint-action studies in driving. Traditional and more respected areas of psychology influence alcohol-and-driving research. The same things—reaction time, automobile driving courses, primitive automobile simulators, sensory thresholds, psychomotor tasks—are repeatedly studied; such studies add little or no new knowledge but preclude fresh ideas. Talented people are discouraged from attempting alcohol research by the ridicule of uninformed colleagues, by entrenched inexpert opinions on alcohol, and by the lack of intellectual support and respect of academicians, editorial boards, referees and funding agencies.