Increasing freeway crashes in China has put lots of pressure on freeway safety improvement. Freeway should be the safest type of highway because ofits highest design criteria. However, the unique design and traffic operation in China have resulted in somewhat opposite impression by general public that the freeway is the most dangerous highway. This paper presents the results of a study conducted by Beijing University of Technology on the development of predictive safety performance function for freeways. By carefully examining the significance of each proposed variables, four variables were used in the final multivariate safety prediction model. Traffic volume and environmental variable are the most influential factors in crashes occurrences based on the modeling results. Using well accepted statistical method, the model validation and verification are also presented in this paper. The results of this study will serve as a tool in freeway safety evaluation and crash countermeasure selection in China.
Abstract