The purpose of this study was to determine the effect on accidents of lane widening, shoulder widening, and shoulder surfacing. Detailed traffic, accident, roadway, and roadside data were collected on 4,951 miles of two-lane roadway in seven states. Statistical testing was used along with an accident prediction model to determine the expected accident reductions related to various geometric improvements. Accident types found to be most related to cross-section features included run-off-road, head-on, and sides wipe (same direction and opposite direction) accidents. The roadway variables found to be associated with a reduced incidence of these related accident types (and included in the predictive model) are wider lanes, wider shoulders (paved slightly safer than unpaved), better roadside condition, flatter terrain, and lower traffic volume. Lane widening was shown to reduce related accidents by 12 percent for 1 foot of widening (for example, 10-foot lanes to 11-foot lanes), 23 percent for 2 feet of widening, 32 percent for 3 feet of widening, and 40 percent for 4 feet of widening. The effects of shoulder widening on related accidents was determined for paved and unpaved shoulders. For shoulder widths between zero and 12 feet, the percent reduction in related accidents due to adding paved shoulders is 16 percent for 2 feet of widening, 29 percent for 4 feet of widening, and 40 percent for 6 feet of widening. Accident reductions due to adding unpaved shoulders were slightly less than for paved shoulders.
Samenvatting