4.1 Major Findings

This review of the literature provides strong evidence that impairment of some driving-related skills begins with any departure from zero BAC. By 0.050 g/dl, the majority of studies have reported impairment by alcohol. By BACs of 0.080 g/dl, 94% of the studies reviewed reported impairment. These results include behavioral response areas which are on the one hand insensitive to the effects of alcohol and on the other hand scarcely representative of the demands of driving, such as critical flicker fusion and simple reaction time.

There is evidence that behavioral areas differ in their relative sensitivity to the impairing effects of alcohol. This is in agreement with Moskowitz and Robinson (1988), Kruger (1990), Holloway (1994), and other investigators. Table 2 reports the lowest BACs at which different behavioral areas exhibit impairment.


By Lowest BAC at Which
Impairment Was Found

By First BAC at Which
50% or More of Behavioral Tests
Indicated Consistent Impairment

0.100 Critical Flicker Fusion Simple Reaction Time, Critical Flicker Fusion
 0.060-0.069   Cognitive Tasks, Psychomotor Skills, Choice Reaction Time
0.050-0.059   Tracking
0.040-0.049 Simple Reaction Time Perception, Visual Functions
0.030-0.039 Vigilance, Perception Vigilance
0.020-0.029 Choice Reaction Time, Visual Functions  
0.010-0.019 Drowsiness, Psychomotor Skills, Cognitive Tasks, Tracking Drowsiness
0.001-0.009 Driving, Flying, Divided Attention Driving, Flying, Divided Attention

The first column lists behavioral areas by the lowest BAC at which impairment was found. The second column lists behavioral areas by the first BAC at which 50% of the behavioral tests indicated impairment. That is, the point at which the majority of behavioral tests showed impairment. Note that, with the exceptions of simple reaction time and critical flicker fusion, all driving-related skills exhibited impairment by 0.070 g/dl in more than 50% of tests.

This review supports the suggestion by Ferrara et al. (1993) that discrepancies in test results reflect a lack of standardization in testing methods and that failures to find alcohol impairment at low BACs may be attributable to the use of tasks which are not sensitive to behavioral changes caused by alcohol. If studies only involving driving (in simulators and on the road), simulated piloting, divided attention, and vigilance are examined, 73% of the test results in those areas exhibited impairment by 0.039 g/dl. Including tracking and drowsiness, 65% of the tests performed by 0.039 g/dl showed impairment. Decisions with regard to BAC limit should not be determined on the basis of behavioral areas that are relatively insensitive to alcohol. Crash risk is determined by impairments of those behavioral areas which are important determinants of driving and which are the most sensitive to alcohol.

Virtually all subjects tested in the studies reviewed here exhibited impairment on some critical driving measure by the time they reached 0.080 g/dl.

4.2. Methodological Issues

It is impossible for a reviewer who is not physically present at the execution of the study to stipulate beyond the authors designation how to classify some of these studies. Unfortunately, the variability in results between studies, even within a category, limit the ability to provide advice on the use of response measures to investigate alcohol effects on driving. Researchers in psychometrics, who develop new behavioral tests, are obligated to provide adequate evidence of the validity and reliability of these tests before they are used in measuring behavioral functions in patients. It would appear incumbent on experimental investigators of alcohol and other drug effects to at least provide some defense as to the adequacy of their response measures. Hopefully, this review will contribute to putting to rest the utilization of critical flicker fusion and simple reaction time as measures for examining alcohol effects.

In discussing the imposition of BAC limits, the issue has been raised that not every individual is necessarily impaired at that specific limit. It should be noted that the BAC at which every single individual is impaired has not been an issue in any of the above studies. The requirements of experimental design precludes doing such an analysis. None of the behavioral variables examined, except perhaps for drowsiness, is so over-learned that there are no order or practice effects during an experiment. The majority of studies reported are within-subjects designs where each subject acts as his/her control and where the order of treatment, alcohol or placebo, is counterbalanced. Some subjects receive the alcohol treatment on the second day after their performance has improved from the first placebo treatment day. This makes it difficult to demonstrate impairment in all individuals, since the practice or order effect and the alcohol treatment effect are confounded. It is possible to make the assertion that subjects were impaired by 0.08 g/dl because a substantial majority of subjects would have to exhibit impairment in order for the study to report statistical significance. Thus, even for studies which have reported impairment at 0.010 g/dl, nearly all subjects would have had to demonstrate impairment.

4.3. Future Research

A valuable future area of research would be to examine the interaction of alcohol with sleep deprivation and circadian rhythms. There is strong evidence, produced by the studies on drowsiness, that ability to remain alert and functioning is impaired by alcohol. Nearly all the experiments included in this review involving drowsiness were performed during the day. Noting that the majority of alcohol-related crashes occur at night, it is clear that additional research on time of day is called for.

An additional area which should be examined further is the effects of alcohol on subjective responses. This would include effects on emotion, motivation and judgment. Many theories of driver behavior emphasize motivational and attitudinal factors as important determinants of safety. There is experimental literature demonstrating alcohol effects on aggression and other subjective behaviors. If the importance of the topic is granted, it would appear advisable to review the literature on the role of emotional/motivational factors in driving and the effects of alcohol on such factors. At this point in time, the literature appears incomplete. Note that both Kruger (1990), and Holloway (1994) indicated that subjective reactions were among the most likely to demonstrate impairment at low BACs.