Traffic engineers must have an understanding of the factors associated with crashes in order to identify and implement effective countermeasures to reduce the occurrence of crashes. This paper presents the results of a study in which the speed, flow and geometric characteristics that significantly affect the crash rate on four-lane highways were identified and models developed to describe the effect of these characteristics on the crash rate. Multiple Linear Regression, Robust Regression and Multivariate Ratio of Polynomials were used to develop the models. Data were obtained from speed monitoring stations and police accident reports. The paper develops three equations of the multivariate ratio of polynomial form using mean speed, standard deviation of the speed and flow per lane as independent variables and crash rate as the dependent variable. (A)
Abstract