Once you are in a crowd crush, there is little you can do to save yourself. If possible, Still recommends gradually working your way to the outside of the crowd by stepping sideways as the crowd moves backward.
Panic can happen even without a crowd, in wide open space, as we’ll see. But in almost every case, it is a symptom of a larger problem. In fact, the reason that so many disaster researchers are loath to talk about panic is that the word is a conversation killer. The crowd panicked, end of story. But there is a problem underneath the panic. But the problem was almost always preventable. Just like hurricanes don’t have to kill people, crowds don’t have to crush.
The closer you look at the crowd, the less irrational the behavior looks. If caught in a suffocating crush of humanity, is it irrational to try to survive, even if the only way to do that is by clawing on top of people? Certainly not. So does this mean that the crowd’s behavior is irrelevant? Is a stampede simply inevitable at a certain crowd density? Is panic a myth after all?
Stampedes are primarily a function of time, space, and density. But there is an X factor. If a high-density crowd is moving through jamarat, a fatal crush may result. But it won’t necessarily. As with a herd of cattle, something else has to happen to start the stampede.
The Prerequisites of Panic
One way to solve the panic riddle is to consider when panic does not happen. Before Britain entered World War II, there was a long period of anticipation in London. Evacuations of children began. Sandbags lined the roads. People carried gas masks, and movie theaters closed down. British military planes droned above the populace, day after day. Authorities worried that German attacks on civilians, when they came, would cause widespread panic. In the Lancet, the editor of the British Journal of Medical Psychology wrote: “Since air raids may produce panic in the civilian population it is well to consider the factors that facilitate or diminish panic, and what steps, if any, may be taken against it.”
But when the bombs finally started falling, people behaved unexpectedly. In her captivating dispatches from London to the New Yorker after the war began, Mollie Panter-Downes described the public’s defiant stoicism: “The British are either the calmest or the stupidest people in the world,” she wrote. Appealing to the national sense of humor and identity, the Ministry of Information launched a clever series of advertisements depicting “correct British behavior” under stress: “What do I do in an air raid? I do not panic. I say to myself, ‘Our chaps are dealing with them,’ etc.” (Note the wonderfully blasé use of the word etc.) After the first major raid killed four hundred people, train commuters bragged to one another about the size of the bomb craters in their neighborhoods, Panter-Downes wrote, “as in a more peaceful summer they would have bragged about their roses and squash.”
Forty years later, the expectation of panic consumed U.S. Authorities after a nuclear power plant accident at Three Mile Island in Pennsylvania. It was an unprecedented event, and good information was slow in coming. It was not even clear who was in charge. If ever a situation was ripe for panic, this would presumably qualify. At first, the governor advised everyone in a ten-mile radius to stay inside with their doors closed. Later, the governor announced that pregnant women and preschool children within a five-mile radius should evacuate. The National Guard was readied. Air-raid sirens sounded in the downtown area of the state capital. But the evacuation turned out to look a lot like any evacuation before a hurricane. The elderly were the least likely to evacuate. And the people who did leave did so in an orderly manner. The predicted anarchy from panicking drivers did not materialize.
What kept people calm? Ed Galea, the evacuation expert in the United Kingdom, had long wondered if culture influenced a public response to an emergency. The English are notoriously self-possessed, after all, and for all the countries’ differences, American culture is closely linked to that of Britain. Perhaps public reasonableness was a matter of nationality.
In January of 2005, Galea ran an experiment to try to find out. Would Brazilians respond the same way as Brits to a surprise fire alarm? Before running the experiment, Galea took bets among his British colleagues about what would happen. Half said the Brazilians would never move as quickly as the Brits. They would sit, finish their coffees, and then consider, just consider, making an exit. The other half of his colleagues had an even less attractive view of Brazilians: they predicted they would break into some kind of Latin American hysteria dance—panicking and running in all directions.
First he tested the Brits. He ran an unannounced drill in the beginning of the school year at the library at the University of Greenwich, which proceeded in a very orderly fashion. Then Galea flew to Brazil. When he got there, he found that the Brazilian authorities had about as little respect for their own people as his colleagues. They were convinced the drill (an unusual event in Brazil, unlike in fire-safety-conscious Great Britain) would cause a panic. They were so distraught that Galea almost had to cancel the experiment. One high-ranking official actually said he was worried that people would bite their tongues off. And how would that look? Dozens of tongueless innocents fleeing from a library—all for the sake of research! Finally, Galea got approval to go ahead with the drill at a library.
The Brazilians, as it turns out, were just as orderly and rational as the Brits. He found no statistically significant difference in their response times. And miraculously, no one bit his or her tongue off in either hemisphere.
In 1954, a young sociology PhD candidate at the University of Chicago named Enrico L. Quarantelli pushed aside all the conventional wisdom and painstakingly mapped out when panic occurred and when it didn’t. His resulting paper, published in the American Journal of Sociology, was dry but groundbreaking. Through 150 interviews following three different disasters, Quarantelli drew up a sort of recipe for panic.
Panic occurs if and only if three other conditions are present, Quarantelli concluded. First, people must feel that they may be trapped. Knowing they are definitely trapped is not the same thing. In fact, in submarine disasters, such as the horrific sinking of the Russian submarine Kursk in 2000, humans are not likely to panic. The crew knows there is no way out. At submarine depths, even if they were to swim out of the hatch, they would not survive.
But if people worry that they might be trapped, that is a trigger for panic—even in wide open spaces. “War refugees caught in the open by strafing planes can develop as acute a sense of potential entrapment as individuals in a building during an earthquake who see all exits becoming blocked by debris,” Quarantelli wrote.
Second, panic requires a sensation of great helplessness—which often grows from interaction with others. What starts as an individual sense of impotence escalates when people see their feelings reflected around them. One person caught in explosions in a factory explained it this way to Quarantelli: “I can truthfully say that when I heard the moaning and crying of the others, I did get quite panicky.” Perhaps the Blitz and the Three Mile Island accident, like most disasters, did not cause panic because people did not feel very helpless. They could take shelter or evacuate, after all. And following the Lake Wobegon effect, the psychological phenomenon named after Garrison Keillor’s above-average town, most people probably suspected that they would be among the lucky ones.
The final prerequisite to panic is a sense of profound isolation, Quarantelli found. Surrounded by others, all of whom feel utterly powerless, we realize we are exquisitely alone. We understand that we could be saved—but no one is going to do it. Panic is, in a way, what happens when human beings glimpse their own impending mortality—and know that it didn’t have to be so.
Quarantelli’s analysis is unsatisfying in some ways. “A sense of helplessness” is hard to define or measure. “A sense of isolation” is even more nebulous. As we’ve seen, most people feel a strong sense of solidarity in a disaster. Hajj pilgrims feel an overwhelming sense of unity. So if Quarantelli is correct, what causes a sudden sense of isolation?
Laboratory Pan
ic
To find out, I returned to the aviation experts, people who understand human behavior better than almost anyone else—largely because the government requires them to do so. The crowd crush behavior seen on the hajj is, once in a rare while, seen on planes as well. Remember the Manchester plane disaster from Chapter 5, the mysterious 1985 accident modeled by Galea? A Boeing 737 carrying 131 passengers on a charter flight to Greece began its takeoff out of Manchester. The crew heard a thud and, thinking they had blown a tire or hit a bird, abandoned takeoff immediately. Nine seconds later, as they were cruising down the tarmac, the crew got a warning of a fire in the left engine. At the same time, a gust of wind carried the fire into the fuselage. Flames and thick, black smoke began pouring into the cabin.
But the situation still did not look catastrophic. After all, the plane had stopped moving and the pilot had called for an evacuation. There were no injuries from any kind of crash impact. Airport fire and rescue trucks had already arrived and begun spraying foam all over the plane.
The problems began at the exit in the middle of the plane, over the wing. A young woman seated in 10F, just beside the exit, had trouble opening the hatch. She was unfamiliar with how it worked, and her amygdala was in charge. She spent precious seconds pulling on what turned out to be her armrest. Finally, her female friend in the next seat stood up and pulled at the release handle. The hatch, which weighed forty-eight pounds, fell into the plane across the chest of the original passenger, trapping her in her seat. The other passengers then intervened, lifting the hatch off of her. Finally, forty-five seconds after the plane had come to a halt, the door was opened, according to the investigation report issued by the British government.
The passengers still had at least a couple of minutes to get out. But then the situation deteriorated further. Alarmed by the smoke and intense heat, many passengers rushed toward the exits and stumbled, falling onto the floor of the aisle. Others then started climbing over the seat backs to get past the pileup. One passenger standing farther up in the plane looked back and saw a tangle of human bodies in the center section. It appeared no one could move forward in the chaos. “People were howling and screaming,” he later told investigators.
This is the deadly “faster-is-slower” effect, as crowd experts call it. Above a certain speed, people moving for an exit will actually get out much later than if they had moved more slowly. An arch of bodies is created around an exit as everyone tries to get out at the same time. The friction leads to a clog, which slows down the entire evacuation. Imagine trying to walk out a narrow door moving shoulder to shoulder with five other people. You would knock into one another as you tried to gain access to the opening, spending time and energy squeezing past the other bodies before even getting to the door. If one of you tripped, the jam would become exponentially worse. The faster you tried to leave, in other words, the smaller the doorway would essentially become.
In the Manchester crash, the area by the overwing exit was not yet engulfed in flames, but it nevertheless became a death trap. People clogged the path, with some bodies half out of the exit. The last passenger to leave the plane was a young boy, who was pulled from the exit by a firefighter, five minutes after the plane had stopped. Fifty-five people died that day. Another fifteen were seriously injured.
A better word for panic might be overreaction, says G. Keith Still, the expert on hajj crowds. Something happens, some sudden stimulus, to cause a dangerously dense crowd to overreact. In the Manchester crash, it might have been the rush of heat or the slowdown caused by the confusion at the exit. In some cases, all it takes is one or two people to panic, dooming everyone. In a survey of 457 U.S. passengers involved in serious evacuations in the late 1990s, a large number of people reported misbehavior by a few. Twenty percent saw other people climbing over seats; 29 percent saw people pushing. Just over 10 percent said they saw passengers arguing with other passengers. And 6 percent actually admitted pushing someone else. At dusk on February 1, 1991, a Skywest plane collided with a USAir Boeing 737 plane at Los Angeles International Airport. Everyone on the Skywest plane died immediately. On the USAir plane, nineteen passengers died from inhaling smoke—ten of them while waiting in line to use the right overwing exit. The investigation found that the delay was caused in part by a scuffle between two passengers.
In the hajj, a stampede can be caused by people suddenly changing direction. Like a loud noise or a rumor of a bomb, a sudden surge of pressure can cause an overwhelming sensation of isolation and helplessness. Likewise, in a herd of animals, a stampede can be caused by something as legitimate as a flash of lightning—or as trivial as a cigarette lighter flicked by a foolish cowboy. So psychology matters, even though physics matter more.
Firefighter Tommy Walker found himself inside a ministampede most unexpectedly one Sunday afternoon in Kansas City, Missouri. He and a friend were off duty, sitting in a booth eating lunch at a busy pizzeria when a water heater blew up in the basement. No one knew what had happened. They had just heard an explosion, louder than fireworks, which shook the building. People started heading for the door. At first, the crowd was quiet, Walker remembers. There was no smoke or fire, no other threat aside from the initial noise. But very quickly, as the crowd clogged up at the small doorway and the exit slowed, the dynamics shifted. Walker watched from his seat as people started trampling over one another in their rush to get out. He could have been watching the Manchester evacuation. “People started yelling. People tripped over chairs, and they were just walked over. Good God, I kind of liken it to a cow going down in a cattle truck,” he says. “Those cows usually perish.”
Suddenly, Walker’s friend, a fellow firefighter, got up to leave too. The aisle was jammed with people, so his friend started walking over the table. Walker couldn’t believe what he was seeing. He grabbed his friend. “Sit the fuck down!” he told him. “There is no smoke, there is no fire, and that crowd is a lot more dangerous than anything else here.” His friend sat down.
The crowd thinned out a few moments later. There was still no actual fire, smoke, or any threat at all. Walker and his friend walked outside. There, on the sidewalk, were the wounded. Ambulances began to arrive. A handful of people had broken bones and sustained other injuries—not from the explosion, which turned out to be minor, but from the crowd. Says Walker: “Literally it was a nothing incident that turned into complete chaos in five seconds.”
It is hard to re-create panic in a laboratory without violating basic ethical principles. But after the Manchester crash, aviation safety researchers tried their best. In a series of experiments, they ran volunteers through evacuation drills in a mock-up of the plane. They couldn’t do it. People remained fairly calm, as they do in most evacuations.
But then something amazing happened. The researchers offered the volunteers $10 if they were among the first to get out of the plane. Things changed. This time, people began tripping and falling as they raced off the plane, piling up in the doorway and halting the evacuation—just as the Manchester passengers had done! So even without the threat of death, people could become their own worst enemy. The volunteers thought they might not make it out in time to get the money; perhaps they also started to feel helpless as the line out the door slowed; and then they may have felt the isolation that accompanies an every-man-for-himself rampage. The prize injected competition into the crowd, just as fire creates competition for air and space. The competition motivated people to move faster, which increased the density of the crowd. Then physics took over.
Tragedy at IKEA
Money, as it turns out, often has deadly consequences for crowds. The list of fatal crowd crushes over the past twenty years is stacked with events that should have been frivolous—giveaways, sales, store openings. On September 2, 2004, three people were killed at the opening of a new IKEA store in Saudi Arabia. In England, just four months later, a man was stabbed in a crowd crush at the opening of another IKEA. Some six thousand customers had shown up for the midnight opening of the
largest IKEA in the country. Witnesses reported people punching each other in the face to get a sofa. The chaos finally ended when IKEA shut down the store and riot police dispersed the crowd. “I think it’s fair to say we misjudged it, and maybe were a bit naive,” an IKEA spokesperson told the Guardian. “But some shoppers behaved like animals and started shoving and pushing everyone else. There’s not much we can do to stop that.”
Is it true that there is not much IKEA can do about customers who behave “like animals”? There is a reason stampedes happen over and over again in the same locations, says crowd expert Still. The problem lies in the design of the space and the management of the crowd. Simply put: too many people are moving through too small a space too quickly. There are plenty of practical solutions for these problems, at jamarat or at IKEA, says Still. One of the easiest fixes is to allow more time for the crowd to pass through. Also, traffic should be one-way, to avoid the turmoil of counterflow. Putting a column in front of an exit is an elegant way to help prevent clogging. In every case, communication between organizers and with the crowd is critical. Officials need to constantly monitor the crowd’s movement and quickly relieve pressure points. The New Year’s Eve celebration in Times Square in New York City is a model event, Still says. Police funnel over half a million revelers into separate viewing pens. Once people leave a pen, they cannot come back. That rule reduces the amount of traffic flow. At police headquarters, meanwhile, about seventy officers monitor live feeds of the crowd from dozens of cameras.
We understand how to prevent crowd crushes. That’s not the problem. The problem is convincing the people in charge to make the changes. In 1971, at the Ibrox stadium in Glasgow, Scotland, sixty-six people were killed in a grisly scene in a stairway after a soccer game. But the very same stairway, which was dangerously long and steep, had been the scene of major accidents in 1961, 1967, and 1969, which had left two people dead and dozens injured. It was only after the fourth disaster in 1971 that the stairway was removed.
Unthinkable: Who Survives When Disaster Strikes - and Why Page 19