SIGNIFICANT FACTORS IN RAIL FREIGHT ACCIDENTS : A STATISTICAL ANALYSIS OF PREDICTIVE AND SEVERITY INDICES IN THE FRA ACCIDENT INCIDENT DATABASE

The Federal Railroad Association maintains a file of carrier-reported accidents and incidents that meet threshold criteria for damage cost and/or casualties. Using a five year period from this data base, an investigation was conducted into the relationship between quantifiable risk factors and accident frequency and severity. Specific objectives were to identify key variables in accidents, formulate a model to predict future accidents, and assess the relative importance of these variables from the perspective of routing and shipping decision making. The temporal factors YEAR and MONTH were found to be significant predictors of risk; accident severity was greatest for accidents caused by track and roadbed defects. Train speed was an indicator of accident severity; track class and training tonnage were inversely proportional to accident severity. Investigation of the data base is continuing, with a final report expected by late summer. 15 refs., 1 fig., 10 tabs.