Analysis of Accident Severity on Chinese Two-lane Rural Highways using Multinomial Logit Model

摘要

In order to identify the possible contributing factors on two-lane rural highways to accident severity, the multinomial logit model was applied to traffic accident data collected from police-reports. A total of 253 accidents, occurred on two-lane rural highways in 2007, were considered in this paper. The dependent variable is divided into three categories, namely fatal accident, injury accident, and property damage only accident, and 13 independent variables were selected in this paper. Of 13 independent variables obtained from police-reported accidents, five were significantly associated with accident severity, namely location, weather, clients' sex, vehicle type, and accident type. The findings show that the multinomial logit model used in this paper provides a better understanding of risk factors related to two-lane rural highways.