The main purpose of this study was to evaluate the effect that the analysis method has on the assessed impact of an intersection treatment. Accident data for three years before and three years after the treatment were collected for 20 improved intersections. The after/before accident ratio at each intersection was computed by four different methods: the direct comparison method, the Empirical Bayes approach, the regression technique, and the statistical accident- prediction models. The results of these four methods did not produce the same conclusions about the impact of the treatment on traffic safety. Analysis of the results indicated that the impact of intersection treatment on traffic safety is significantly affected by the evaluation method used. In general, the regression technique was found to underestimate the impact of intersection treatment. The Empirical Bayes approach considered the effect of factors that had been ignored in other methods, and is believed to provide a better estimate of the impact of an intersection treatment. (a).
Abstract