The Accident Risk Measuring Model for Urban Arterials

In Taiwan, the fatal accident rate was 146 per million populations in 2008, roughly 1.8 times the Organization for Economic Co-operation and Development average rate. Many studies have focused on traffic safety and attempted to identify methods for preventing traffic accidents. However, the analytical models used most frequently, including Statement Statistic, Regression Analysis, Factor Analysis (FA), Cluster Analysis (CA), and Data Mining, assess the correlations or relationships between one dependent variable and independent variables. Those methods cannot indentify overall causal relationships and cannot combine different accident indexes to define accident risk. This study uses Structural Equation Modeling (SEM) to construct a theoretical framework for traffic accident risk analysis for urban arterials. Different accident indexes, such as accident number, fatality rate, injury rate, and casualty rate, are combined to construct an accident risk evaluation model, to explore factor significance, and to identify relationships between factors. These factors include driver characteristics, vehicle characteristics and road characteristics. This study segmented urban arterials according their geometry, and determined accident risk for each section. Analytical results show that accident risk on urban arterials can be calculated by combining accident number, injury rate, and casualty rate. Driver characteristics and road geometry, particularly the road division pattern affect accident risk significantly; however, vehicle characteristics did not, indicating that one can focus on driver behavior and road design when attempting to decrease road accident risk.