Citation

Zwahlen HT. J. Occup. Accid. 1976; 1(1): 21-38.

Copyright

(Copyright © 1976, Elsevier Publishing)

DOI

unavailable

PMID

unavailable

Abstract

Twelve subjects (20-37 years old) were tested in the laboratory and eleven out of these were also tested in a car in the field, first under a no alcohol condition and then under an alcohol condition (approximately 0.10% BAC). In the laboratory the subjects' simple and choice reaction times for two uncertainty modes were measured and their information processing rates (3 bits uncertainty) were determined. In the field the subjects' driving skill for driving through a gap (soft obstacles, 20 inches clearance) at 20 , their static visual gap (car and post) judgment capabilities and their gap risk acceptance decisions for a 46 feet viewing distance (using psychophysical experimental methods) were measured. Based upon the driving skill measure (standard deviation of centerline deviations in the gap), the mean of the psychometric visual gap perception function and the mean of the psychometric gap risk acceptance function, the "Safety Distance" and the "Driver Safety Index" (DSI), developed by Zwahlen (1973), were obtained.Based upon a statistical analysis of the laboratory and field data we may conclude: (1) that in all experiments the effect of alcohol varies widely from one subject to another; (2) only a few increases in the group averages of the reaction time means (1-5%) and the standard deviations (20-42%) due to alcohol are significant at [alpha] = 0.05, and (3) none of the changes in the group averages of the means and standard deviations of the driving skill measures due to alcohol are significant at [alpha] = 0.05 (although most changes are in the expected direction when considering the adverse effect of alcohol).A conceptual model illustrating a hypothetical localized system failure, developed by Blumenthal (1968), has been modified in such a way that the system demands on the driver-vehicle system are represented in terms of choice reaction times, while the driver performance at any point in time is represented by two choice reaction time probability distributions, one representing the sober condition, the other representing the alcohol condition. This modified system model is used to demonstrate that the combined time increases in mean performance and performance variability under the influence of alcohol provide a much better explanation for the higher accident involvement than the historically most frequently used rather small decrements in mean performance.