IDENTIFYING BLACK SPOTS ALONG HIGHWAY SS107 IN SOUTHERN ITALY USING TWO MODELS

Black spots (BS) are highway locations where the potential for accidents is unacceptably high when compared to the established risk tolerance criteria. It is argued in this paper that to effectively utilize available funds for safety improvements, one must first designate high priority BS locations. This paper presents the results of two model applications for establishing the potential for accidents and designating safety BS along a highway. The two models are based on multivariate Poisson regression and empirical Bayesian (EB) methods. The application involves a 25-km stretch of highway in southern Italy, for which accident and exposure data are available for the period 1993-1999. The pattern of BS suggested by each model is compared. The EB model was found to yield fewer BS locations than the Poisson regression model. If, as argued in this paper, safety countermeasures are best applied at BS, use of the EB model could result in significant cost savings.