Empirical Bayes Application in Highway Safety Research

• Activity Number: 234
• Type: Topic Contributed
• Date/Time: Monday, July 30, 2012 : 2:00 PM to 3:50 PM
• Sponsor: Section on Government Statistics
• Author(s): Roya Amjadi*+ and Kim Eccles
• Companies: Federal Highway Administration and VHB
• Keywords: Highway Safety ; The Evaluation of Low Cost Safety Improvements ; Crash Modification Factor (CMF) ; Empirical Bayes ; FHWA, Turner-Fairbank Highway Research Center Safety Evaluation ; FHWA Pooled Fund Study

Abstract

Improving safety of the Nation's transportation system is the highest priority of the U.S. Department of Transportation (DOT). State DOTs and other Transportation agencies are often reluctant to invest in safety improvements without scientific evidence that demonstrates improvements' effectiveness for reducing crash. In recent years, Federal Highway Administration has made effort for development and use of Crash Modification Factors to assist State DOTs in selecting proven countermeasures to reduce crashes effectively. In partnership with 29 State DOTs, FHWA, Turner Fairbank Highway Research Center is sponsoring the Evaluation of Low Cost Safety Improvements Pooled Fund Study that is a major on-going research for evaluating safety improvements with potential to reduce crash frequency and severity on national level. Data is collected from states that implement safety improvements (unproven) in U.S., and typically Empirical Bayes (EB) is used for evaluations. EB accounts for regression-to-the-mean and combining data from diverse jurisdictions, accommodates use of crash rates for normalizing data with traffic volume differences between study periods, and reduces uncertainty.