Assisting CalTrans identify high crash locations

One of the first steps in effectively managing a roadway network is to identify the locations in need of more formal safety evaluation. The UNC Highway Safety Research Center will be assisting the California Department of Transportation to develop methods for identifying high crash roadway segments and intersections on California roads where potential safety improvements may be implemented. HSRC will evaluate various methods using crash, roadway, and traffic data from California state highway locations and identify the methods that are optimal for determining the locations best suited for cost-effective safety improvements.

"Conventional methods that only use crash counts or rates are problematic and can trigger 'false positives'," said Raghavan Srinivasan, Ph.D., principal investigator of the project. "These methods do not effectively account for the potential bias due to regression-to-the mean phenomenon in which sites with a randomly high crash count could be incorrectly identified as being hazardous and vice versa."

Another issue associated with conventional methods that makes use of crash rates problematic is the assumption that crash frequency and traffic volume are linearly related, i.e., that a 20% increase in volume will result in a 20% increase in crashes. Recent studies have shown that the relationship between crashes and volume is non-linear, and this relationship depends on a number of factors, including the type of facility.

The review and evaluation provided by HSRC will provide fresh and independent insights on the most promising methods that should be applied using the data from California state highways. The projected completion date for the project is summer of 2010.