The report considers accident frequencies on two-lane rural roads as affected by cross-section elements and roadside design. Relationships reported in the predominantly North American, English-language literature are reviewed. These indicate that the effect of roadside design on accident frequency is of a similar magnitude to the combined effects of lane and shoulder widths. The accident prediction relationship developed in the USA as part of the debate on geometric standards for 3R projects was identified as the most comprehensive in terms of the factors included and the underlying research. However, the relationship greatly underpredicts the accident savings from shoulder paving/sealing reported by other North American and Australian studies. Consideration of the different rural accident type mix in the USA and Australia, and limited Australian results on changes in accident frequency with changes in cross-section, were used to adjust the US relationship to better represent Australian experience, including the effects of shoulder seals. (A)
Abstract