Objectives: Taxis, one of the main transportation modes that occupy the roadways in Seoul, are semipublic transportation modes for transporting passengers safely and promptly. Considering that one fifth of passenger vehicles on the roads in Seoul are taxis and the crash rate of taxis is double the exposure to traffic, it is important to identify risk factors of taxis from that of private cars. In this paper, crash causes and characteristics in both taxi crashes and private car crashes are investigated to identify the risk factors in accordance with the injury severity.Methods: An eight-year light-vehicle crash dataset was utilized, in which injury levels were defined as severe vs. non-severe. Three binary logit models that estimate the severity of crashes, the injury severity for at-fault drivers, and the injury severity for victims were modeled for taxi crashes and private car crashes. Independent variables were extracted and included in the models to evaluate the odds ratio of each predictor variable.Results: The results indicated that violation of traffic signals and signs was the highest contributor among all violation types for taxi crashes and parties involved (at-fault driver and victims), while driving on the wrong side of the road resulted in the highest increase in the odds ratio for private cars. Head-on collision and nighttime driving increased the likelihood of severe injury risk for all models, while age was the most prominent factor for the injury level of victims. Use of seatbelts had a major impact on the at-fault drivers, especially for taxis.Conclusions: This study identified the risk factors that affect the crash- and party-related severity level when casualties involved taxis and private cars. By employing both crash- and party-level models, the study not only identifies the risk factors among taxis and private car crashes but also provides a comprehensive picture of the injury profile of all vehicular occupants, which helps to devise safety measures that enhance the safety and reduce the injury severity for parties involved in crashes.
Keywords: Taxi crashes; binary logit model; private car crashes; risk factor.