Previous research into relationships between road accident frequencies and traffic volumes is reviewed. Methods used are assessed and the main conclusions of the soundest studies drawn together. Some researchers have deduced theoretical relationships between accident numbers and traffic volumes and checked these against data, while others have derived relationships more empirically. In order to deduce an empirical relationship, it is necessary either to study different roads carrying different volumes of traffic, or to examine time-series data. Research of the former type has often shown that the number of single-vehicle accidents per unit of vehicle distance travelled tends to decrease with increasing traffic, while the number of collisions between two or more vehicles tends to increase. Knowledge about relationships between accidents and traffic volumes at intersections is less extensive but includes useful results. Time-series studies have usually shown that accident numbers have grown more slowly in the long term than have volumes of traffic, but, because of the many other time-dependent factors which affect the incidence of accidents, care must be taken in drawing conclusions from such studies about the underlying relationship between accidents and traffic volumes. Comments are made concerning the data and methods likely to be useful in further research. (Author/publisher)
Abstract