Accident Cause by Railroad




This report includes each accident cause reported by each railroad involved in an accident/incident. Accident causes are categorized by human factor, mechanical, signal, track and miscellaneous.
If multiple railroads are involved in a single train accident, each railroad involved in the accident is required to submit a Form 54 Railroad Equipment Accident/Incident. Each railroad provides their own information and interpretation of the accident, including accident cause. This report contains all accident causes reported to the FRA for all accidents. Thus, each accident may have multiple accident causes reported. As such, this report does not provide a total because it would be overstating the number of accidents. 
For more information, please reference Multiple Reports for a Single Train Accident and Understanding the Data Included/Not Included below. 

Reportable Accidents
Railroads involved in a rail and/or rail equipment accident/incident, that meets the monetary threshold are required to report FRA Form 6180.54 Rail Equipment Accident/Incident. The monetary threshold changes annually. Accidents with a damage cost greater than or equal to the annual monetary threshold are required to be reported to the FRA on Form 54. This data is reportable data and is published on FRA’s safety data site. Accidents with a damage cost under the monetary threshold are required to be recorded, and accident information must be share with the FRA as requested. This data is not published on FRA’s safety data site. This data is called accountable data.
Train Accidents (not at Highway-Rail Crossing) vs. Accidents (including Highway-Rail Crossings)
The type of accidents reported on FRA Form 6180.54 include derailments, collisions, highway-rail crossings, RR grade crossings, obstructions, explosions-detonations, fire/violent ruptures, and other impacts.
The Train Accident (not at Highway-Rail Crossings) Summary (2.03) report excludes reported highway-rail crossing incidents. Generally, “train accidents” refer to accidents/incidents where one or more trains or train equipment is involved, whereas “highway-rail crossing incidents” refer to accidents/incidents where a train/train equipment and a non-train entity (car, truck, bus, pedestrian, bicyclist, etc.) are involved at a highway-rail crossing. All other reports include highway-rail crossing incident data.
Since all highway-rail crossing incidents are reported to the FRA on Form 6180.57 Highway-Rail Crossing Accident/Incident report, there is duplication if you sum all data from Form 54 and Form 57. The train accident (not at highway-rail crossing) reports on this site allow you to sum data from Form 54 and data from Highway-Rail Grade Crossing Incidents without duplication.
Multiple Reports for a Single Train Accident and Understanding the Data Included/Not Included
If multiple railroads are involved in a single train accident, each railroad with equipment involved in the accident is required to submit a Form 54. Each railroad provides their own information and interpretation of the accident. Since multiple reports for a single train accident are reported to FRA, the result is that the full dataset contains multiple reports for the same accident. If each report is counted, then the number of accidents would be overstated.
FRA selects one report to represent the accident in its summary reports. FRA then manipulates the data for each report for one accident by adding additional columns, where the selected railroad’s track type, primary accident cause code and track class is applied to all reports.
The data in these reports display one report for each accident, instead of all reports. Let’s use an example to understand what this means: Railroad A and Railroad B are in an accident. Railroad A reported a human factor accident cause and Railroad B reported a signal accident cause. FRA elects to use Railroad A’s report to represent the accident. In this report, the accident would be displayed as a human factor accident and not a signal accident. Thus, the data in train accident (not at highway-rail crossing) reports and accident (including highway-rail crossing) reports would count the accident as a human-factor accident cause instead of a signal cause. Similarly, if Railroad A reported the accident occurred in a yard and Railroad B reported industry, the accident would be counted as a yard accident.
The railroad reports contain all reported accident causes and track types. So, data for Railroad A’s human factor and Railroad B’s signal accident would be counted. Similarly, Railroad A’s yard accident and Railroad B’s industry accident would be counted. However, summing data would overstate the number of accidents. Thus, there is no Grand Total because there are multiple reports for one accident. Though there is only one record used to represent the accident, there are additional columns on each record to report the total fatalities, injuries and damage cost for all railroads involved in a single accident.
There are six reports on this site that display and summarize the data and should be used for different purposes. To review data for the number of the total number of unique accidents, select Train Accident (not at Highway-Rail Crossings) Summary (2.03)Accident Cause (including Highway-Rail Crossings) (3.10) or Accident Type (including Highway-Rail Crossings) (3.08). These reports have one railroad report selected to represent each accident, so the counts are not overstated (using our example, these reports contain Railroad A’s data). The fatality, injury and damage cost include data for all railroads in the accident (e.g. data for Railroad A and Railroad B). Please note there is no View Data by Railroad – it would only display data for Railroad A, even though Railroad B is also involved in the accident. If you choose to filter by railroad, please understand you are only getting data for Railroad A and Railroad B would be excluded.
Accident Track Type by RailroadAccident Cause by Railroad and Accident Detail Listing (3.18), on the other hand, include data for all reports for all accidents. So, data for Railroad A and Railroad B is included. However, if Railroad A hit another train owned by Railroad A, then there would be two accident reports filed by Railroad A for one accident. So, Railroad A’s reported accident cause and track type would both be displayed and summed, although there was only one accident.
Please refer to the Rail Equipment Accident/Incident Data (Form 54) dataset if you would like to review all reports for all accidents.
General Information 
It is recommended to use Google Chrome or Microsoft Edge and to block third-party cookies in your browser. It is recommended, under your Display Settings on your computer, to set the display resolution to 1920 X 1080 and to set the size of text apps and other items to 100%.
For the publication timetables, please see the Data Release Schedule. For a crosswalk of legacy site reports to this new site, please see the Report Cross-Reference.
To download the data into Microsoft Excel or a csv, scroll down to select the Download icon in the lower righthand corner, and then select Crosstab.
Please note that blank values in the table mean that no data has been reported; they are not zeroes.
For the full rail equipment accident/incident dataset, please reference Rail Equipment Accident/Incident Data (Form 54). For the unique rail equipment accident/incident dataset,  please reference Rail Equipment Accident/Incident Data (Form 54) Subset – Unique Train Accidents (Not at Grade Crossings).
For more information on this data and other forms, please refer to the FRA Guide for Preparing Accident/Incident Reports. For more information on the filters in this dataset, please refer to the Report Filter Glossary.