CORRELATING ACCIDENTS AND VOLUMES AT INTERSECTIONS AND ON URBAN ARTERIAL STREET SEGMENTS

A summary is given of the results of a study to investigate simple correlations between accident frequency and average daily traffic volumes. The first part of the study developed a linear regression equation based on before-and-after accident rates at junctions in Phoenix that had been fitted with traffic signals in the period 1976 to 1979. Good correlations were obtained, but the break-even accident rate was found to differ from that of a similar study in Utah, probably because of differences in traffic characteristics at the two sites. Other types of intersections are studied, both with and without traffic signals. Good correlation was found between the accident frequency and the daily average intersection volume at intersections with signals. Results show that the accident rate increases rapidly with traffic volume at intersections without channelisation on approach. At volumes above 40000 vehicles per day, the accident frequencies rise more rapidly. It was not possible to obtain a linear correlation between accidents and volumes at an unsignalled intersection where a number of variables affect results. Fairly good non-linear correlations are obtained between volumes and non-junction accidents on mile-long segments of major urban arterial streets. (TRRL)