Examination of Highway-Rail Grade Crossing Collisions Nationally from 1998 to 2007


Richard A. Raub

Document Type

Journal Article

Publication Date


Subject Area

operations - traffic, operations - frequency, infrastructure - stop, infrastructure - traffic signals, planning - methods, planning - safety/accidents, planning - safety/accidents, mode - rail


Warning systems, Warning devices, U.S. Federal Railroad Administration, Traffic fatalities, Stop signs, Statistical methods, Statistical analysis, Road safety, Railroad grade crossings, Mathematical statistics, Level crossings, Highway safety, Highway railroad grade crossings, Highway rail intersections, Grade crossings, Flashing traffic signals, Fatalities, Fatal accidents, Death, Databases, Crossbucks (Railroad crossing), Crashes, Collisions, Accident rates, Accident frequency


Most studies of highway–rail grade crossing collisions have looked at specific classes of warning devices (e.g., gates). Some limited work on crossings with passive warnings is available. This study, using FRA data, examines 10 years of collisions nationally at public crossings with the four primary crossing warning systems: crossbucks only, stop signs, flashing lights, and gates. The study examines and compares collisions and collision rates. Overall, crashes and fatalities at highway–rail crossings have been decreasing while the number of crossings, vehicles, and train movements has been increasing. The number of public highway–rail crossings, the number of collisions, and the percentage of collisions resulting in a fatality are not well correlated. Some important characteristics of collisions are these: a train striking a vehicle is more likely to cause a fatality, male drivers outnumber female drivers by three to one, drivers younger than 25 are overrepresented, and drivers older than 65 are more likely to be killed than those in other age groups. Additionally, collisions are more likely to occur during the day and in clear weather. As timetable speeds increase, drivers are less likely to run into a train. The highest collision rates based on traffic volume occur where stop signs are used. Collision rates vary by state, and the ranking in rates by device also varies substantially from state to state. Finally, though useful, the databases maintained by FRA could be improved to be more comprehensive, both for research and for identification of potentially dangerous crossings.