Skip to main content
Best News Website or Mobile Service
WAN-IFRA Digital Media Awards Worldwide
Best News Website or Mobile Service
Digital Media Awards Worldwide
Hamburger Menu




Commentary: What caused the Halloween crowd crush in South Korea

Organisers would benefit from understanding some basic crowd psychology, says the University of Sussex’s Prof John Drury.

Commentary: What caused the Halloween crowd crush in South Korea

Onlookers, police and paramedics in Itaewon on Oct 30, 2022, where a crowd crush killed more than 150 people. (Photo: AFP/Jung Yeon-je)

BRIGHTON: More than 150 young lives snuffed out. High school students, teachers, military personnel, a singer and actor, a 24-year-old celebrating his birthday – they had all gone to Itaewon, one of the most popular districts in Seoul, for a night of fun on Saturday (Oct 29). They ended up getting crushed in a deadly crowd surge.  

Crowd crush disasters like the one at Itaewon this Halloween usually involve the following factors.


First, there are too many people in the same space. By “too many” I mean anything above five people per square metre, which is dangerous. When there is a sudden movement or surge within a crowd of this density, this can lead to pressure on people at the other side of the crowd.

More than 150 people were killed in the crush in a crowded Itaewon alley in Seoul. (Photo: AFP/Anthony Wallace)

In very large crowds, people entering the crowd cannot see how crowded it is at the front. This characterises features of the Hillsborough disaster of 1989, for example, in which 97 people died after a human crush during a football match between Liverpool and Nottingham Forest at Hillsborough Stadium in Sheffield, South Yorkshire, England.

They moved forward and had no idea that people were being crushed in another part of the crowd.

Second, there is an obstruction. At Itaewon, people were in a narrow 4m-wide downhill alley that looks similar to the area where people were trapped at the Love Parade disaster in Germany in 2010. Twenty-one people were crushed to death or suffocated while more than 600 others were injured in a congested tunnel at the music festival.

Density increased as more people came into the area, and likely those in the middle tried to get out; but the walls on each side made extrication difficult.

Third, in many cases there is a crowd collapse, which itself is fatal and represents a further obstruction. We found this at the Bethnal Green tube shelter disaster of 1943, where 173 fleeing from an air raid were crushed to death while attempting to enter the shelter.

At this event, someone slipped on the stairs and a portion of the crowd fell on top of her. Others at the back moved forward into this obstruction.


Commentators often describe events like Itaewon as a “stampede” caused by “crowd panic”.

The term “stampede” is misleading. There was no running at most of the fatal crowd crush events, and running does not seem to be the cause of tragedy at Itaewon.

Videos of the Itaewon incident certainly show fear and distress. But this occurred after the crushing had begun. It was a consequence, not a cause of the disaster.

So, “panic” is an unnecessary explanation. The psychology of the crowd was not abnormal, it was completely normal.

In a situation where density is too great, where there are obstructions and where is a crowd collapse, normal routine behaviours (moving towards one’s target, moving to try to get out, staying with loved ones) can contribute to the crush.

The question now is how events like these can be prevented. It’s worth noting that there are regular crowd events that exceed safe levels of density, but which go off without such problems. I’m thinking of the annual Haj at Mecca.

One factor that organisers need to take into account is the desired spaces for the crowd at any event.

Where do people desire to be and gravitate towards? These will be the likely pinch-points. Having established that, and having a measure of the location capacity, it is then necessary to regulate the numbers in that space.

There are different ways of doing this, but I would recommend communicating with the public both before the event and on the day to get them to understand the reasons for the restriction.

People in a crowd don’t usually understand the risks of density until they are right in the middle of it, so it’s a good idea to explain this carefully and in advance.

Map showing where crowds were during a Halloween event in Seoul's Itaewon district on Oct 29, 2022. (Infographic: CNA/Rafa Estrada)


When the tragedy occurred on Saturday, about 100,000 people were estimated to be in Itaewon. But there were only 137 police officers in Itaewon at the time, authorities have said.

They also said that they did not have guidelines to handle the huge crowds.

Typically, "safety management plans" for gatherings of more than 1,000 people have to be submitted to the government in advance for review by police and fire departments.

But in the case of the Itaewon Halloween festivities, there was no designated event organiser – individual bars, clubs and restaurants held their own events and partygoers flocked to the area.

It’s not sufficient to simply say that Itaewon was a public space and there was no organiser. The land is presumably managed by the local authority, and they should therefore be the ones with a safety plan.

There should have been a professional safety group that would have anticipated possible risks by having an estimate of likely numbers, as well as personnel in place on the day to manage the locations where there was a risk of overcrowding.

Back in 2005, in Brighton, there was an unticketed music event that was held on the beach – a public space. The organisers expected 60,000 people but 250,000 turned up. It was a massive strain on facilities and on staff, and all emergency exits were blocked by the crowd. There was a real danger from the overcrowding, though on this occasion disaster was averted (probably because of the cooperative behaviour across the crowd). The next time the organisers ran the event a few years later – again on public land – it was fully ticketed, with fencing around it.

Organisers would benefit from understanding some basic crowd psychology. We often assume that everyone desires “personal space”. But this isn’t always true.

For live events and celebrations, the opposite is the case. Here, we seek proximity to like-minded others, as this is where the “atmosphere” is.

People will seek out, endure, and enjoy what are objectively dangerous levels of density at many events, and organisers need to be aware of this.

Yes, all these fatal crowd crushes are preventable. We now have the science and experience – whether crowd dynamics (maths and engineering) and psychology – to ensure safer events.

John Drury is Professor of Social Psychology at the University of Sussex. He specialises in the study of collective behaviour.

You may also be interested in:

Source: CNA/aj


Also worth reading