Citation

Cleveland DE, Kostyniuk LP, Ting KL. Transp. Res. Rec. 1984; 960: 1-13.

Copyright

(Copyright © 1984, Transportation Research Board, National Research Council, National Academy of Sciences USA, Publisher SAGE Publishing)

DOI

unavailable

PMID

unavailable

Abstract

The complex relationships among several geometric design elements and accidents on two-lane rural roads were studied. Two data sets were used in modeling effects of traffic volume greater than 2,000 vehicles per day, driveway and intersectional conflict frequency, roadside obstacle characteristics and geometric design elements on total accident occurrence for a national data set, and off-road accident frequency and severity for a Michigan route data set. Geometric design elements were aggregated into bundles or groups that are actually found in the field as a result of design policies. Advanced multivariate techniques were used to study these interactions. It was found that accidents interact in such a complex way with traffic volume that use of the conventional vehicle mile exposure rate in modeling is less fruitful than treating average daily traffic (ADT) as an independent variable. For the prediction of accidents, the effects of ADT were found to be most important followed by driveway and intersection density and the geometric elements. The interactive effect of access point density with volume was also important, as was the interactive effect of access point density and geometric characteristics. Longitudinal alignment elements were found to dominate in off-road accident prediction for rural two-lane roads at ADT values of 4,000 vehicles per day or less, whereas roadside elements were of more importance at higher ADT values. No significant independent effects of cross-sectional elements were found in the total accident prediction. Off-road injury accident prediction on rural two-lane highways was more sensitive to roadside obstacle location and characteristics than total accident prediction. Simple categorical models developed from this research explained 63 percent of the total accident variance in the national data set and 54 percent of the off-road accident variance in the other.

Language: en

Keywords

MATHEMATICAL MODELS; ROADS AND STREETS; HIGHWAY ACCIDENTS - Analysis; HIGHWAY ENGINEERING - Accident Prevention