METHODOLOGY FOR ESTIMATING TRAFFIC SAFETY IMPROVEMENTS AT INTERSECTIONS . PROCEEDINGS INTERNATIONAL CONFERENCE ON NEW WAYS AND MEANS FOR IMPROVED SAFETY, TEL AVIV, ISRAEL, FEBRUARY 20-23, 1989

Intersection accidents represent a significant proportion of overall motor vehicle accidents. With limited funds available for improving hazardous locations, transportation agencies require more accurate estimates of actual effectiveness of intersection safety improvements. This study develops and implements a methodology for post-implementation evaluation of safety improvement projects at intersections. Accidents are random, rarely occurring events. This produces in a given time period large random fluctuations in accident frequencies, which suggests that statistical analysis employing confidence intervals, rather than point estimates, is required. Two technical problems complicate this treatment of accident occurrence as a random variable, however, and preclude simple before-and-after analysis of unadjusted accident data. First, the identification of hazardous locations is generally based on above-average accident frequency during the most recent period(s) for which data are available. A sample of locations receiving treatment then will be characterized by selectivity bias (also known as the "regression-to-the-mean", effect), which implies that accident frequencies at the selected (hazardous) intersections may be expected to decrease in future periods even without the safety countermeasures. Some recent work in empirical highway safety analysis has accounted for this selectivity bias using empirical bayesian methodology. The primary contribution of this study is the development of a "combined" approach which addresses these technical issues. (Author/TRRL)