Theory, Explanation, and Prediction in Road Safety: Promising Directions

On November 20-21, 2008, the Federal Highway Administration and the Transportation Research Board co-sponsored a workshop titled, Future Directions in Highway Crash Data Modeling. The objectives of the workshop were (a) to explore promising future directions in highway crash data modeling and (b) to identify potential areas for advanced research to provide a theoretic foundation for explaining crash causation. The focus of the workshop was on the effect of highway infrastructure elements on safety. Goals of the research identified by workshop participants are the further development of science-based safety evaluation and the development of more stable, reliable, and transferrable highway safety predictive models. Twenty-five invited traffic safety researchers and practitioners attended the workshop. Experts in highway safety data analysis were brought together with experts from allied fields (such as epidemiology, public health, and human factors) to share their experiences and methods for quantifying treatment effects. Through breakout groups and plenary sessions, the participants were asked to describe critical issues and challenges; explore alternative modeling approaches and concepts; and identify promising new directions for explaining the contributing causes of crashes. This document summarizes the discussion and describes potential areas of advanced safety research identified by the participants. Four major research theme areas were identified: Structural Modeling; Surrogate Measures of Safety; Improving the Evaluation of the Intervention Effectiveness; and Multiscale Safety and Driver Behavior Modeling Approaches. Each theme area consists of several research components. These components identify specific research projects to address the issues associated with each theme area. These areas and their components are described in detail in this circular.