As current chairman of the Association of Trial Lawyers of America’s “Railroad Law Section” I have closely followed the argument by the nation’s railroads that adding simple stop signs to highway grade crossings makes them safer, reducing the risk of injuries or deaths. The railroad arguments seemed logical to me–at least initially. Why wouldn’t making all drivers stop at all crossings improve safety, reduce collisions and be inexpensive? Well, think again! Adding the stop signs actually makes each rail crossing more dangerous than WITHOUT the stop sign! So is the finding of a major unbiased study published in April 2006 discussed below. After reading the study, I had further questions, so I interviewed the author himself.
The study, “Examination of Highway-Rail Grade Crossing Collisions Over 10 years in Seven Midwestern States,” was published in ITE Journal in April 2006, and authored by Richard A. Raub. Raub recently retired from Northwestern University Center for Public Safety as a senior research scientist, primarily in traffic safety, and was an original member of the Highway Safety Manual Joint Subcommittee, and holds a M.S. in transportation science from Northwestern.
Essentially, Raub studied highway-rail grade crossing collisions for a 10 year period (1994-2003) in seven midwestern states, [Illinios, Indiana, Iowa, Michigan, Minnesota, Missouri, and Wisconsin] using Federal Railroad Administration injury and property damage data from crossings with 1) crossbucks only; 2) stop signs; 3) flashing lights; and 4) lights and gates. His important findings:
“Overall the analysis shows that the highest collision rates using any basis occurred at locations where the warning device was a stop sign. The push to upgrade crossings from crossbucks only to sign signs may be creating a false sense of improved safety.”
“36 percent of the collisions [analyzed] occurred at crossings with only crossbucks as a warning device. Next highest were crossings with flashing lights, representing 28 percent of the collisions. Crossings with gates and STOP signs represented 23 percent and 13 percent, respectively.”
“Approximately 57 percent of crossings had only crossbucks. An additional 22 percent had flashing lights. These two devices accounted for almost 80 percent of all crossings in the seven states. Only 13 precent of all crossings had gates installed, with few exceptions, they were two-quadrant. Eight percent has STOP signs.”
“What is surprising about the findings is the extremely high collision rates for crossings with STOP signs. Even for Michigan, which had the lowest STOP sign crash rate, it was 21 percent higher than the rate for crossbucks only in that state….It is also important that even when STOP signs were installed at crossings that previously had only crossbucks, collision rates increased. This finding, although based on a small sample because of the very low likelihood of a collision at any passive crossing, further suggests STOP signs may not provide the assumed level of enhanced safety.”
Raub explains that it is not clear WHY stop sign controlled crossings have higher collision rates. One theory is that motorists regard stop signs at crossings to “have less meaning” than the law intends. Motorists may misjudge the speed of trains or the time needed to cross after stopping at the stop sign. Raub also did not analyze “yield” signs compared to stop signs, which one study showed may have more safety promise than stop signs.
Shapiro: It seems from the tenor of your study, that you may have been surprised by the Wisconsin statistics, and that this caused you to expand the analysis. Is that what happened?
Raub: I did not expect the crash rates (based on ADT) for Wisconsin crossings to be as high as they were. Because of that finding, I decided to expand the study to the 7 Midwestern states. The expanded analysis suggested that the initial findings were not an anomaly. For each state, when based on ADT, but also daily trains which is another measure of potential hazard, the crash rates were higher than for other warning systems (gates, flashing lights, and crossbucks only).
Shapiro: Since the publication of your article this year, what notable feedback have you received?
Raub: I have not received feedback since the publication. However, I shared the first study with a number of professionals both in safety research and railroads, and as a paper presented at the annual Transportation Research Board in 2004. The Federal Railroad Administration (FRA) Office of Safety has accepted the results and is has provide data to take the study national. George Black of the National Transportation Safety Board gave his whole-hearted approval. While as a county highway engineer, he installed stop signs which initially appeared to improve safety, but he now doubts their overall usefulness. Dr. Robert Gallamore, Director of the Northwestern University Transportation Center fully agreed with the findings. Feedback from the presentation also was positive. There were, of course, dissenting voices, one of which (anonymous to me) helped delay publication of the article for about one year.
Shapiro: Have you ever been asked to provide any expert opinions on the effect of stop signs in any railroad crossing litigation, and if so, by which side/party?
Raub: The contract we had with the Union Pacific Railroad which helped initiate the original study prohibits me from serving as an expert witness against them. This was what the University and the Center for Public Safety were willing to accept as the only condition.
Shapiro: What types of services are you offering in your position with this outfit now?
Raub: I am free to act without constraints except as noted. Note, too, that once I can get the crossing inventory data transferred from SAS (Statistical Analysis System) format which I had at Northwestern to .TXT so I can use it at home (or find the funds to buy SAS), I am planning to take a nationwide look at characteristics of crossing collisions. This step is encouraged by the FRA Office of Safety.
Shapiro: To safety experts that want to add stop signs at highway-rail grade crossings, rather than to consider lights and/or gates, how do you respond and how do you frame the issue?
Raub: While gates are “obviously” the safest choice, they also are very costly. Flashing lights, as specified by the Manual on Uniform Traffic Control Devices and the FRA, while less complex than gates also are costly. Both forms of active crossing warning systems cannot be supported from a cost-benefit perspective.
Currently there is underway a National Highway Cooperative Research (NCHRP), project 3-76B under the Transportation Research Board which is examining the potential benefits of other forms of low-cost warning systems to be used at passive crossing (crossbucks only) denoting the presence of a train. I was part of the bidding but unsuccessful team for this project. These low-cost systems may be beneficial.
From both a traffic engineering safety and a human factors perspective, I always have had reservations about the use of stop signs at any intersection with little or no cross traffic. The highway literature has suggested that a yield sign appears to have much greater effectiveness. Fortunately, a recently complete National Highway Cooperative Research (NCHRP), Report 470, project provides guidelines for use of yield signs at highway-rail crossings. To me, what now needs to be done is a well constructed study of motorist behavior at these crossings under three conditions: crossbucks only, yield, and stop signs.
Richard A. Raub is presently a principal in Raub Associates, 10136 SW Washington Street, Portland, OR 97225, (503) 297-2786 (H), (503) 833-2409 (cell), DickRaub@AOL.com.