Eyes can play deadly trick at train crossings

Published 9:00 pm Sunday, December 7, 2003

In the United States, trains collide with vehicles 3,000 times a year, killing more than 300 people.

In almost every case, grade-crossing accidents — such as when a motorist tries to avert the train’s warning gate — are blamed on driver stupidity.

In fact, it is a lot more complicated than that.

Scientists say that many of these accidents are caused by deadly misperceptions, the visual and behavioral quirks known to science but not to ordinary drivers. And making rail-crossing encounters even more treacherous are train and crossing designs that fail to take into account how people perceive and behave.

It’s not enough to stop, look and listen at railroad crossings, these experts warn, because what you think you see can kill you.

"The attitude of the Federal Railroad Administration is that almost every accident that ever happened at a railroad crossing is the driver’s fault," said cognitive psychologist Marc Green, a partner in the Toronto consulting firm Visual Expert.

One peculiarity of human perception is that large objects in motion appear to be moving more slowly than they really are. We can observe this phenomenon at any airport, said Herschel Liebowitz, emeritus professor of psychology at Pennsylvania State University. Jumbo jets seem to drift down to the tarmac during landings, while smaller jets seem to race toward the runway, even when the larger plane is going faster.

Liebowitz, who first described the size/speed effect and other grade-crossing perils in 1985, field-tested his theories by riding in the cab of a locomotive and questioning railroad personnel: "It was almost immediately obvious what the problem was. … People misestimated the speed of trains."

The problem is compounded by perspective.

When we look down a railroad track, we don’t see the rails, or the telephone poles running along the tracks, as parallel. We see them converging in the distance at what artists and scientists call the vanishing point. As Liebowitz explains, we have learned to associate that apparent convergence with distance, and so we are likely to assume that the train is farther away than it really is.

Collisions also have what Liebowitz calls a "deceptive geometry" that can prove fatal.

Green explains the problem: Typically, you glimpse the train with your peripheral vision. Never as clear as central vision, peripheral vision is especially poor at gauging velocity. Even as the train moves toward you, and you move toward it, the train’s image maintains a relatively constant position on your retina, at the edge of your visual field. The result: "You don’t see it moving," Green said, and you assume it is still a safe distance away.

Then, when you are about to collide, the train’s image on your retina suddenly expands in all directions — a condition called looming. But at that point, you probably can’t stop in time, and neither can the train.

By Green’s estimate, perception — or misperception — is a factor in more than 80 percent of highway accidents, including those involving trains.

Many public officials charged with train-related safety are aware of the science of rail accidents but continue to take a Darwinian view of them.

"We know that the problem is there are just too many impatient drivers who fail to obey traffic regulations at either active or passive crossings," said Warren Flatau, a spokesman for the Federal Railroad Administration.

The tendency to blame the victim in grade-crossing accidents exasperates cognitive psychologist Green: "That lets the authorities off the hook. Then they don’t have to redesign the system."

The United States has made progress in train safety, experts say.

"Railroad crossing deaths in the U.S. have come down from 786 in 1975 to 315 in 2001. That’s a pretty good achievement, "said Eric Wigglesworth, an Australian accident researcher who won the Order of Australia — comparable to British knighthood — for his contributions to Australian public health and safety.

"I think this was largely due to the U.S. government’s rail-highway crossings program which since 1978 has injected $4 billion into crossings improvements," Wigglesworth said.

The elimination of thousands of grade crossings and the increase in so-called active crossings, especially in heavily populated areas, have been important advances.

Twenty years ago, only about 50,000 of the country’s 225,000 public grade crossings were protected by flashing lights, bells and/or gates that drop down when a train is about to pass. Today, there are about 62,000 of these active crossings out of 154,000.

Such active warning systems are expensive — even simple ones can cost $150,000 to install. And active crossings have human-factors problems of their own, which may help explain why they account for half of all grade-crossing collisions.

Liebowitz described one major problem 20 years ago. The low-tech gates then in use at active crossings dropped at a fixed time, whether the passing train was traveling quickly or slowly. As a result, drivers often had to wait long after the gates had dropped for a train to arrive.

But drivers are impatient, and those who have had to wait time after time for the train may decide that they can safely ignore the warnings.

"It’s called ‘the cry wolf phenomenon,’ " Green said.

Research published in the early 1990s convinced the Federal Railroad Administration that some motorists who had to wait more than 40 to 50 seconds at an active crossing would try to drive around the gates. Now, most active crossings are equipped with electronic "predictors" that gauge an approaching train’s speed and lower the gates for as few as 20 seconds before the train arrives.

Educating the public is one way to prevent accidents.

Operation Lifesaver, which coordinates rail safety messages throughout the United States, includes the larger-looks-slower phenomenon and the perspective illusion in its literature and also reminds drivers and those tempted to take their chances on railroad tracks that trains need a long time to stop.

Critics point out that most safety programs fail, in part, because most drivers who try to beat a train across the tracks believe they are safe. It is only recently, according to Najmedin Meshkati, a professor of industrial and systems engineering at the University of Southern California, and others, that the rail industry has begun looking at driver behavior as it is, rather than as it should be. Meshkati’s own work focuses on how decision-making styles and crossing complexity may contribute to collisions.

Among the rarely recognized dangers: the difficulty of hearing warning bells when sitting in a modern soundproofed car and the likelihood that new immigrants won’t recognize the traditional American crossbuck, the simple X-shaped railroad sign that warns drivers to yield to an oncoming train. And decision-making at grade crossings is even tougher at night.

"The demands that are placed on the driver at a railroad crossing are extraordinary," said Wigglesworth, the Australian researcher who successfully campaigned for improved warning systems there. "It’s not surprising there are so many accidents."

Ten years ago, Wigglesworth persuaded Australian lawmakers to address the danger of putting the same warning signs at both passive and active crossings. "The mix of active and passive systems kills people," he said, and explained why: Say you are a driver approaching an active crossing, marked with a crossbuck, warning bells, lights and gates. If the lights don’t flash and the gates are up, you know it is safe to cross the tracks.

Then you come to a passive crossing. It too is marked with a crossbuck. But, this time, the lack of red lights and lowered gates doesn’t mean "Go ahead." It means just the opposite. The crossbuck tells you to stop and look because a train may indeed be racing your way.

Using the same traffic signs when different actions are desired breaks one of the Ten Commandments of human-factors engineering, Wigglesworth explains: "If you use the same stimulus, you must expect the same response."

The Federal Railroad Administration is examining why drivers behave dangerously at grade crossings, an important first step. And it is studying so-called "intelligent transportation systems" that would warn train personnel, even stop the train, when a vehicle or person is on the tracks.

For now, long-armed gates that effectively keep cars from driving across the tracks are a good interim solution.

But people tend to behave in predictable ways, and grade crossings will continue to claim lives until their designers recognize that.