Before-After Crash Analysis: A Primer for Usign the Empirical Bayes Method

摘要

The Montana Department of Transportation spends a significant portion of its resources on reconstruction and pavement preservation projects. While the costs for these projects are well quantified, the resulting benefits are not, particularly with respect to traffic safety. In the past, the safety-related benefit attributable to reconstruction and pavement preservation activities have been evaluated by MDT's Safety Management Section using a before/after, case/control analysis for crash trends at various sites around the state. While these methods have been used for many years as a means of predicting safety improvements for roadway segments, recent studies have shown that the Empirical Bayes (EB) method of analysis is more accurate in the estimation of safety improvements for before-after studies. The EB method combines the strengths of before/after, case/control techniques with regression methods for estimating safety-related benefits. The Empirical Bayes method is now used in the USDOT's Interactive, Highway Safety Design Model (IHSDM) and will be used in the Safety Analyst software, making this method the standard in professional practice. This tutorial explains the applications of the Empirical Bayes method for both Interstate and non-Interstate roadways. Companion products include a report that describes the theory, application and limitations of the empirical Bayes method as applied to several Interstate and non-Interstate reconstruction and pavement preservation projects completed in 1997-1998 in the State of Montana, as well as an Excel spreadsheet that facilitates use of this method for MDT's Safety Management Section.