The Federal Railroad Administration (US DOT) maintains a file of carrier-reported railroad accidents and incidents that meet stipulated threshold criteria for damage cost and/or casualties. A thoroughly-cleaned five-year time series of this data base was subjected to unbiased statistical procedures to discover (a) important causative variables in severe (high damage cost) accidents and (b) other key relationships between objective accident conditions and frequencies. Just under 6000 records, each representing a single event involving rail freight shipments moving on mainline track, were subjected to statistical frequency analysis, then included in the construction of classification and regression trees as described by Breimann et al. (1984). Variables related to damage cost defined the initial ''splits,'' or branchings of the tree. An interesting implication of the results of this analysis with respect to transportation of hazardous wastes by rail is that movements should be avoided when ambient temperatures are extreme (significantly < 20(degrees) or > 80(degrees)F), but that there should be no a priori bias against shipping wastes in longer train consists. 2 refs., 2 figs., 12 tabs.
Rail transportation risk and accident severity: A statistical analysis of variables in FRA's accident/incident data base
1991
27 pages
Report
No indication
English
Railroad Transportation , Transportation Safety , Accidents , Rail Transport , Risk Assessment , Cost , Damage , Data Base Management , Equipment , Fault Tree Analysis , Frequency Analysis , Injuries , Regression Analysis , Statistical Data , Temperature Dependence , Meetings , Tables(data) , Accident severity , EDB/320202 , EDB/990200 , EDB/570000
Light Rail Accident Involvement and Severity
Transportation Research Record | 1996
|Light Rail Accident Involvement and Severity
Online Contents | 1996
|Light Rail Accident Involvement and Severity
Transportation Research Record | 1996
|Light Rail Accident Involvement and Severity
British Library Conference Proceedings | 1996
|