Analysis of Severity of Vehicle Crashes at Highway-Rail Grade Crossings: Multinomial Logit Modeling

Abstract

The purpose of this paper is to develop a nominal response multinomial logit model (MNLM) to explore the impact of various explanatory variables on three different crash severity levels of vehicle-related crashes at highway-rail grade crossings (HRGCs) in the U.S.A. Crash data (from 2005 to 2012) and crossing inventory data for this study are obtained from the Federal Railroad Administration. The highway-rail crash and crossing inventory data are merged together by using USDOT crossing identification number to create a data set consisting of 7,414 observations after data cleaning. Different variables are considered for estimating the coefficients of the model. The MNLM results indicate that when rail equipment with high speed struck a vehicle, the chance of resulting in fatal injury increases. Crashes occurred at higher temperatures are also more likely to be injury and fatal as compared to those occurred at lower temperatures. In addition, the study reveals that male vehicle drivers with the age being 25-60 or above 60 years, pick-up trucks, concrete and rubber surfaces are more likely to be involved in more severe crashes (including injury and fatal crashes). On the other hand, truck-trailer vehicles, foggy and snowy weather conditions, development area types (i.e. residential, commercial, industrial and institutional), and higher daily traffic volumes are more likely to involved in less severe crashes. Finally, it is presented that reducing train speed and vehicle speeds, as well as educating drivers will minimize the chances of resulting in more severe vehicle crashes at HRGCs.