Paper Details

TRB 91st Annual Meeting (January 22-26, 2012)

Event Number:	498
Event Title:	Improved Safety Through Better Design for All Users
Event Date:	Jan 24 2012 8:30AM- 10:15AM
Event Location:	Shoreham, Blue Room Foyer
Event Description:
Event Agenda:

Analysis of Injury Severity on State Routes Incorporating Pavement Quality Using Bayesian Ordered Probit Model (12-3764)

The injury severity of traffic-related accidents has been studied by many researchers in the last few decades. A growing body of literature has focused on the impacts of various factors on the outcome of accidents using statistical analyses based on Maximization Likelihood (ML) estimation methods. However, the evaluation of some factors is still in dispute and until this point, no studies have taken into account pavement quality as a factor of interest. This study attempts to estimate the influences of some key factors including pavement quality on injury severity. Four types of accidents including single-vehicle accident, two-vehicle rear-end crash, two-vehicle sideswipe crash, and two-vehicle angle crash occurring on Tennessee State Route highways from 2004 to 2008 were analyzed. The Ordered Probit (OP) models were fitted on each accident type to identify the variables that contribute most significantly to the outcome. The Bayesian Ordered Probit (BOP) models were simulated afterwards with the selected variables from the OP models. The results indicate that pavement quality is significant in all four collision types. In addition, AADT, median width, lane width, vertical alignment, number of lanes, speed limit, rural or urban, type of terrain, peak hour, light condition, weather condition, driver age, driver sex and vehicle type, as well as the interaction between Pavement Serviceability Index (PSI) and light condition were found to be significant in all or some of these collision types. Differential effects of light condition, driver sex, driver age and vehicle type were found between single-vehicle accident and two-vehicle accident, and the influence of some factors varied among the three types of two-vehicle accident as well.

Authors
     Jiang, Ximiao , University of Tennessee, Knoxville
     Huang, Baoshan , University of Tennessee, Knoxville
     Richards, Stephen H., University of Tennessee, Knoxville
     Yan, Xuedong , Beijing Jiaotong University, China
     Zaretzki, Russell L., The University of Tennessee, Knoxville
     Shu, Xiang , University of Tennessee, Knoxville
     Dong, Qiao , University of Tennessee, Knoxville