The paper is concerned with the analysis of road accident frequencies before and after similar changes in road conditions are made at each of a number of sites. At any one site it is assumed that the total number of accidents recorded is binomially distributed between the before period and the after period, the parameter of the distribution depending on the relative lengths of the periods as well as on the effect of the change. A method of estimating the average effect of the changes is proposed. It is shown how the accuracy of this estimate depends not only on the chance variations arising from the smallness of the accident frequencies but also on any real differences that may exist between the effects of the changes at different places. The methods proposed are illustrated by a numerical example.
Samenvatting