Statistical correlation between road surface roughness and traffic accidents

Road surface condition is widely accepted as one of the most important factors that affect the safety of a land transport system. Although various researchers examined different types of surface condition, only a few focused on how road roughness, particularly quantified by International Roughness Index (IRI), can contribute to the occurrences of Road Traffic Accidents (RTA). Previous studies have found no consistent correlation and trend between both variables due to different regression techniques adopted, and the inclusion and exclusion of different contributory factors. This paper presents an analysis of the influence of IRI and various types of road properties on the observed number of accidents (μ-value) using a statistical regression analysis. The study road, where the accident and road properties were examined, was a high-traffic volume and high-speed highway section in Brunei Darussalam. The outcome of the analysis has indicated that the IRI values have a weak correlation to the observed number of accidents given that other explanatory variables are included into the regression. Limitations of the statistical regression acquired are discussed and recommendations to refine the selected variables and analysis methods are provided.