Citation

Binder A. Accid. Anal. Prev. 1971; 3(4): 237-256.

Copyright

(Copyright © 1971, Elsevier Publishing)

DOI

unavailable

PMID

unavailable

Abstract

In establishing the existence and methods of operation of factors that raise the probability of an automobile accident, the obvious dangers in experimentation under real-life conditions make it necessary (with rare exceptions) to proceed by epidemiological or simulation pro- cedures. In the research on the effects of alcohol upon driving, for example, we have (a) the well-known epidemiological studies of Haddon and Bradess (1959),McCarroll and Haddon (1962), and Borkenstein et al. (1964), and (b) the experimental approaches of Lewis and Sarlanis (1969), Loomis and West (1958), and Drew, Colquhoun and Long (1958). The epidemiological studies involved analyzing the blood-alcohol level (BAL) of accident- involved and/or fatally injured drivers on various highways and led to inferential statements like “about 50 per cent of those killed in single vehicle accidents had BAL’s of O-05 per cent or more”. On the other hand, the laboratory experiments involved administering controlled amounts of alcohol to subjects and then testing their performance on tracking and similar tasks. Such measures as latency, tracking error, and visual acuity were recorded. This research led to such inferences as “driving performance begins to deteriorate at blood-alcohol levels of 20 or 30 mg/lOO ml”.

One disadvantage of the epidemiological method is the lack of control over the variables under investigation; one can only look at the phenomena as they actually occurred and record frequencies of association. There are many obvious difficulties with the experimental approach in the simulation laboratory. But only three of these are of immediate concern. First, in experiments involving social drugs like ethanol and marihuana there is a distinct artificiality in asking subjects to drink screwdrivers or smoke ‘joints” in a waiting room prior to experimentation. It is well-known that many of the effects of alcohol and marihuana are dependent upon the cultural and interpersonal context in which they are consumed; MacAndrew and Edgerton (1969), in fact, present persuasive evidence that drunken comportment results from an interaction between learning and social permissiveness. Second, the subjects commonly chosen for laboratory experiments may not be truly representative of the population of interest, It is clear that the experienced drinking or smoking driver compensates for any drug effect quite differently from the novice, who is more typically involved in laboratory experimentation ...


Keywords: Cannabis impaired driving; Ethanol impaired driving