This research provides a demonstration of a statistical model of accident frequency that can eventually be used as part of a proactive program to allocate safety-related highway improvement funds. Negative binomial regressions of annual accident frequency on sections of principal arterials in Washington State were estimated using data from two years (1992 and 1993). In all, 31.306 observations were used in model estimation (annual accident frequencies on specific sections of highway). The authors' estimation results isolated the effects of various highway geometric and traffic characteristics on annual accident frequency. Subsequent elasticity computations identified the relative importance of the variables included in their specifications. The findings show that the negative binomial regression used in this paper is a powerful predictive tool and one that should be increasingly applied in future accident frequency studies. (A)
Abstract