This report evaluates the effect on accident experience of providing three-foot paved shoulders. Fifteen project locations were selected from five of the department's nine highway district areas. Control sites selected for comparison had similar operational, geometric, and geographic characteristics. These rural roadway segments were high speed, two-lane, two-way roadways. Approximately 111 miles of three-foot paved shoulder projects and 97 miles of non-paved shoulder control sites were evaluated. The average daily traffic ranged from about 1,300 to 8,300 vehicles for the project locations and control sites. The accident experience for two years "before" and "after" was reviewed for each project and for the corresponding control site. Only accident information for mid-block locations was used for analytical purposes with intersection-related accidents excluded. The selection of both project locations and control sites were non random samples. The total number of accidents, property damage accidents and injuries for the project locations increased. The number of injury accidents was unchanged while the number of fatal accidents and fatalities decreased. The total number of accidents, property damage accidents, injury and fatal accidents increased for the selected control sites. Accident types (i.e. run-off-road, head-on, sideswipe, rear-end) increased in some categories and decreased in others for both project and control sites. Statistical tests to determine the significance of the changes were utilized and are discussed within the text. It is concluded that this shoulder treatment had a statistically significant impact on reducing the frequency of run-off-roadway, vehicle overturn, sideswipe-opposite and fixed-object (sign) accidents. Although total accidents at the test sites increased 6 percent, this was 19 percent less than the expected value for total accidents predicted using the 29 percent increase at the control sites. Volume changes at the control and test sites were similar.
Abstract