The problem of developing a satisfactory criterion scale for evaluating driving performance has been bothering experimenters in this field for many years. It has been found that accidents used as a criterion are quite unreliable, especially among lay drivers. This is partly due to variance traceable to exposure factors. Sometimes a spurious reliability may be obtained by correlating periods of time in which there is no control made of the risk factor. Studies made in Sweden (1) and elsewhere have shown that a serious error is introduced when correlating segments of driving record due to the spuriosity introduced by exposure which is very difficult to control even when using bus drivers operating in similar sections of a city.
Proceedings of the Iowa Academy of Science
©1957 Iowa Academy of Science, Inc.
Lauer, A. R.
"A Criterion Scale for Classification of Automobile Drivers,"
Proceedings of the Iowa Academy of Science, 64(1), 542-545.
Available at: https://scholarworks.uni.edu/pias/vol64/iss1/64