Citation

Raff MS. Highw. Res. Board bull. 1953; 74: 18-45.

Copyright

(Copyright © 1953, National Research Council (U.S.A.), Highway Research Board)

DOI

unavailable

PMID

unavailable

Abstract

A study was conducted to find out how accident rates on main rural highways are affected by design features and use characteristics. Fifteen states provided information covering a year's accident experience on about 5,000 mi. Of highway. The basic technique involved dividing the study routes into a large number of short, homogeneous sections, which could then be combined so as to group the sections according to any factors whose effects were of interest. An accident rate, based on vehicle-miles or other suitable units, was computed for each group. The factors studied include number of lanes, average daily traffic volume, degree of curvature, pavement and shoulder widths, frequencies of curves and other sight-distance restrictions, percentage of intersection traffic on the minor road, and many others. Traffic volume was found to have strong effect on accident rates. On most types of highway sections the accident rate becomes higher with increasing traffic volume (except at extremely high volumes , where there is a slight reversal due to congestion). However, two- lane curves and intersections show a different trend, with the accident rates declining as volumes increase. Sharp curves have higher accident rates than flat curves for roads carrying the same amount of traffic. The volume effect described above is more pronounced on sharp curves than on flat ones. Wide pavements and shoulders encourage safety on two-lane curves. On two-lane tangents they do not have any consistent effect. At intersections, the percentage of the total traffic on the minor road is extremely important. It takes only about 15 percent cross traffic to make an intersection more than twice as hazardous as when the cross traffic is less that 10 percent. Another important intersection characteristic is the number of approaches, e.G., three-way intersections are considerably safer than four-way crossings. At bridges and underpasses there was found to be great value in having the roadway on the structure several feet wider than the approach pavement. A number of roadway features do not appear to have any consistent effect on accident rates. These include grade, frequency of curves, and the percentages of commercial and night traffic.