Investigation of factors influencing the effectiveness of railroad-highway grade crossing warning devices

This research was concerned with investigating factors believed to influence the effectiveness of rail road-highway traffic control. The objectives of the study were to:

1. Determine the influence of device suitability on railroad- highway grade crossing safety;

2. Determine the influence of conformance to standards on rail road-highway grade crossing safety; and

3. Identify characteristics associated with locations with high and low accident rates.

A sample of 307 grade crossings was selected from the National Inventory of Railroad-Highway Grade Crossings. Accident histories and general crossing characteristics for the years 1975 through 1980 were compiled. A determination of device suitability was made using Federal Highway Administration (FHWA) guidelines and an evaluation was made concerning each warning device's conformance to the Manual on Uniform Traffic Control Devices (MUTCD). An accident rate was defined which accounted for differences in highway and train volumes and the time period. Analysis of variance techniques were utilized to determine the differences in accident rates due to different warning device type, device suitability, and conformance to MUTCD standards. An analysis was also conducted to determine physical and operating characteristics associated with low and high accident rates.

The major findings of this research were:

1. Differences in effectiveness, as reflected by accident rates, were identified for the warning device categories. Generally, gates were found to be most effective and passive devices were least effective.

2. The data did not support a conclusion that conformance or non-conformance of warning systems to MUTCD standards significantly influences device effectiveness.

3. Devices at locations which apparently warranted higher level devices were not found to have higher accident rates than devices judged to be suitable for their location's condition. The results suggest that the analysis based on FHWA guidelines and the available information failed to accurately assess device suitability.

4. High train speeds were associated with low accident rates and low train speeds were associated with high accident rates. This result suggests that driver behavior is influenced by train speed.

5. Low accident rate observations tended to have high train volumes, and high accident rate observations were associated with low train volumes. Apparently, accidents do not vary linearly with train volume. This suggests that driver perceptions of hazard are influenced by the frequency of train arrivals and that these perceptions influence safety.

6. Extensive use of the National Grade Crossing Inventory revealed limitations in that data base in terms of completeness and accuracy.