The first minutes: everyone waits to see who believes it first
You're indoors, and something is wrong: a strange smell, a rumble, an alarm that could be a drill. The first thing you do is look at other people. They do the same. Everyone sees calm faces around them and decides they're overreacting. But that calm isn't an assessment of the situation - it's a reflection of everyone else's calm.
In 1968, Latané and Darley had participants fill out a questionnaire in a room where smoke came in through a vent. Those sitting alone were quickly driven out by the smoke: about 3 out of 4 went to report it. If 2 planted participants sat nearby and kept writing as if nothing were happening, about 1 out of 10 reported the smoke. Same room, same smoke - only the company was different.
This is what the first phase looks like: not too much movement, but too little. People finish gathering their things, finish a message, ask, "Are we sure this isn't a drill?" This is where minutes slip away that will later prove decisive.
Then the signal becomes unmistakable, and everyone stands up at once. Many head for the exit they came in through: it's familiar, even if another one is closer. From this point on, the mechanism changes completely.
Then it's not fear, but crowd density that decides
While there are 3-4 people per square meter, you can still choose where to put your foot. At around 5-6 people, contact becomes continuous: your body is squeezed from all sides, and a push from the back rows travels through people like through water. The crowd starts to sway in waves, regardless of what any individual wants. The exact threshold is disputed: it depends on the geometry of the space and where the crowd is moving.
Density scale up to 6 people per square meter. At around 5, a crowd flows rather than walks
What looks like panic from the outside feels different from within: you're being carried along. No one has a plan, and there's no room to make a decision. If someone falls, the gap is instantly closed by bodies. In a crush, the danger is pressure on the chest, not other people's cruelty.
Movie-style panic - everyone for themselves, everyone trampling everyone else - is rare in real incidents. Analyses of fires and terrorist attacks repeatedly describe the opposite: strangers share water, pull each other out, and guide one another to an exit.
In a crush, people become savage with fear and trample each other.
More often, the opposite happens: strangers help. The danger comes not from cruelty, but from pressure that no one controls in a dense crowd - not even those who are pushing.
Why might people delay leaving in the first minutes?
What's most dangerous in a dense crowd crush?
Sources
- Latané & Darley, "Group Inhibition of Bystander Intervention in Emergencies," Journal of Personality and Social Psychology, 1968
- Helbing, Johansson & Al-Abideen, "Dynamics of Crowd Disasters: An Empirical Study," Physical Review E, 2007
- work by Drury and Reicher on human behavior in emergencies
Next in the “Crowd Psychology” series
3 articles in full on the site, 3 more in the app
