This paper presents a new theory of speed zoning and the methods are results of tests designed to validate the theory. The basic concept of the theory is that a relationship exists between the rate of occurrence of accidents and the distributive on before and after zoning. The results of the tests substantiate the theory presented. The coefficient of skewness, which is a well defined, easy to calculate the parameter of the speed distribution, is a good indicatator of the potential accident reduction as a site. Changing the distribution from non-normal to normal, as determined by this coefficient, is an effective method of reducing accidents. A second measure of normality, the coefficient of kurtosis, is not a reliable indicator of the expected accident rate reduction on rural state highways. It is felt that this parameter, which is based on the fourth power of the deviation, is too sensitive to individual observations to be reliable. All tests used in this analysis were run at the 95% level, so conclusions drawn from the results of the statistical tests are reasonable. The warrants for speed zoning presently being used are not basted on the speed distributions would increase the effectiveness of speed zones considerably.
Abstract