Do people really behave differently in a crowd?

Australia Broadcast Center

24-05-2025

Sana Qadar: There is this long-standing idea that when people get in a crowd, when they're among a mass of other people, they get more irrational.
Associate Professor Milad Haghani: By the mere fact that he forms part of an organized crowd, a man descends several rungs of civilization. Isolated, he may be a cultivated individual. In a crowd, he is a barbarian.
Sana Qadar: Wow.
Associate Professor Milad Haghani: That is, creature acting by instinct.
Sana Qadar: These are the hugely influential words of Gustave Le Bon, a French social psychologist working in the late 1800s, early 1900s, who is considered the father of crowd psychology.
Associate Professor Milad Haghani: Basically, what he theorized is that by the sheer fact that people are in a crowd of people, they lose the sense of self. And he later on concluded that because of this, crowds are basically trouble in the making.
Sana Qadar: It's an idea that's reverberated through history.
Associate Professor Milad Haghani: Le Bon has had many fans amongst dictators. Mussolini was a fan, for example. And he has said here and there in writing that he has learned a lot about crowd behavior by studying the book that Gustave Le Bon wrote. So because of this, Le Bon is often referred to as the father of crowd psychology. But it's safe to say that crowd psychology has not had the best father.
Sana Qadar: But was Gustave Le Bon wrong? Like, aren't crowds sometimes dangerous? How does crowd psychology really work?
Associate Professor Milad Haghani: Nowadays, we know much more about crowd behavior. There are modern day social psychology theories that have debunked the theory of Le Bon. And we know much about crowd behavior that is more connected to reality.
Sana Qadar: This is All in the Mind. I'm Sana Qadar. Today, from the 1989 Hillsborough disaster, to the London bombings in 2005, to the Halloween crowd crush in Seoul, South Korea in 2022. We find out what modern research tells us about how crowds really behave and how to keep yourself safe in one.
Associate Professor Milad Haghani: Many authoritarian figures in the history have resonated basically with Gustave Le Bon's message.
Sana Qadar: Why did he hate crowds so much? Where was he getting these ideas from? Do you know?
Associate Professor Milad Haghani: Um, it's speculative, my answer. But I think it's just a dog whistle to the authoritarians of the time. You know how sometimes us researchers say something to appease politicians so we get funding? I think this could be something equivalent with the provisions of the time, whatever it was, whatever advantage he was going to get, because he knew that would be a dog whistle.
Sana Qadar: This is Milad Haghani, an associate professor of urban mobility at the University of Melbourne. And he says more than a century later, Gustave Le Bon's assumptions about crowds still turn up in decision-making by leaders.
Associate Professor Milad Haghani: Prime examples of that we can trace to the management of the COVID-19 disaster in the United States.
News archive: (Donald Trump) And I don't want people to be frightened. I don't want to create panic, as you say.
Associate Professor Milad Haghani: If you remember, Donald Trump did an interview with Bob Woodward about the management of the COVID-19 pandemic, where he was asked, you knew about the serious danger of the pandemic months earlier than you took actions. Why did you not warn people earlier? And the response he gave was that I did not want people to panic.
News archive: We don't want to instill panic. We don't want to jump up and down and start shouting that we have a problem.
Associate Professor Milad Haghani: Right. And that speaks to the same line of ideas that has been borrowed from Le Bon's theory, that the sheer fact that people have the information can make them impulsive and irrational and can make them act panicky. And that has been time and time again used as an excuse to withhold information from people at times of emergencies.
Sana Qadar: Now, that idea about not wanting to cause panic, that's not totally without merit. Some research has shown that emotions like anxiety, fear, and panic can spread in a contagion-like effect. But Milad also makes the point that sometimes what's dismissed as panic is really just people acting quite rationally in threatening circumstances. More on that a bit later. But here's just one other thing I want to clarify. If we go back to Gustave Le Bon's original writing in his book The Crowd, A Study of the Popular Mind, to be fair, Milad says it's possible Le Bon was really just thinking about rioters and mobs when he developed his ideas. But he says that nuance wasn't elucidated. The way Le Bon wrote his views, simply being in a crowd of any sort, made individuals prone to losing rationality and acting impulsively. But there are all kinds of examples in the century or so since he was writing that show us this isn't the whole picture.
Associate Professor Milad Haghani: One of the most contemporary psychological theories that explain the behavior of crowds is the social identity theory that explains when people are in a crowd, not in just a physical crowd, a crowd that has some sense of shared identity, it creates a sense of in-group versus out-group and that can moderate their behavior. That is why you see the sense of rivalry in soccer stadiums, for example, in the UK. That can explain crowd behavior better than the sheer fact that just because people are in crowds, they act irrationally. But also it's got implications for the management of emergencies. As I mentioned, not every crowd has a sense of shared identity. If you are waiting for your train in the morning, peak hour on a platform that is very crowded, you do not necessarily have a sense of shared identity with other people. However, if something of risky nature happens in that moment, all of a sudden a crowd that is of purely physical nature, the crowd becomes a social crowd. The sheer fact that they are in a dangerous situation together, it creates a sense of shared social identity. It has been documented in relation to many mass emergencies.
News archive: Good evening, this is Peter Cave with a special edition of PM. There have been a series of explosions on the London Underground and there are also...
Associate Professor Milad Haghani: In particular, for example, in relation to the 2005 London bombing that had affected commuters in the morning peak hour.
News archive: Within the past five minutes, I've seen a line of about 20 people walk past me who are walking wounded. People with cuts all over their face and blood dripping down their faces and bandages.
Sana Qadar: During the 2005 London bombings, four suicide bombers attacked the city's transport network. Three bombs went off on the Underground and a fourth on a double-decker bus. 52 people were killed that day, July 7th, and nearly 800 others were injured.
Associate Professor Milad Haghani: Retrospective interviews and research with the survivors of the attack has shown that people not only did not act selfishly and irrationally, rather, they made conscious decisions to help and to assist other individuals and basically acted in ways that are completely opposite to the notion that people in an emergency lose their mind. And this is not an isolated example in relation to many other cases of emergencies. This has been documented that a physical crowd can suddenly become a social crowd that has a sense of shared identity. And that is a more modern way of viewing the crowds and how it can be managed in times of emergencies. It has created the sense that crowds can be looked upon as allies that can assist in times of emergencies, as opposed to an element that needs to be controlled or an element that needs to be denied information at times of emergency on the assumption that if they have the information, they are going to panic and that is going to be a bad outcome.
Sana Qadar: So the London bombings is a really positive example of crowd behavior, what can happen. But you talked also about soccer crowds. We've seen a lot of soccer gatherings descend into hooliganism. What's the difference there?
Associate Professor Milad Haghani: One of the most challenging crowds, obviously, can be found in football stadiums, soccer stadiums in the UK. And the modern day social identity theory has also helped UK authorities and especially the UK police to change the way they approach crowds. When you look back at 20 to 30 years ago, there were a lot of crowd disasters happening every few years in the stadiums in the UK. The Hillsborough disaster is the most prominent one.
News archive: I see people lying down. I see people there shaking their heads at John's ambulance people there.
Sana Qadar: If you're not familiar with that event, it is the deadliest disaster in British sports history.
News archive: More than 30 ambulances ferried the dead and wounded to surrounding hospitals.
Sana Qadar: It happened in 1989 at the Hillsborough Stadium in Sheffield, England, during an FA Cup semi-final. Ninety-seven people, men, women and children, died in a crowd crush that developed.
News archive: The victims were killed when a big crowd at the FA Cup semi-final surged in packed terraces.
Associate Professor Milad Haghani: Since then, the approach in managing crowds has significantly changed and has been aligned with the modern day crowd psychology. And that is thanks to the effort of some of the social psychologists that are crowd experts and are based in the UK who have helped sharing these ideas with the UK police. And they have shifted basically from a heavy-handed approach of crowd control to seeing crowds as allies and managing the crowds in a form that is more of a dialogue, in an understanding that they are not dealing with trouble necessarily, but the way they approach the crowd can create trouble as a result of their actions.
Sana Qadar: That is essentially what happened in the Hillsborough disaster. An inquest later concluded that some police and stewards were so focused on preventing possible hooliganism that their actions made things worse. They penned fans into tight packs as a way to prevent possible trouble, which had the effect of trapping people in a confined space.
News archive: Because they knew we had more supporters. And the supporters suffer like that.
News archive: It's now clear that the anti-hooligan fences, a standard feature at all British grounds, were the cause of many of the 94 deaths as people were crushed against them.
Sana Qadar: Milad says a lot of lessons were learned from the events that transpired that day. And as a result, we haven't seen disasters like Hillsborough in the UK since.
Associate Professor Milad Haghani: However, some other countries might not have necessarily kept up with the science. The most prominent example is the crowd disaster that we had in 2022 in Indonesia, for example, where tear gas was used in the stadium, even though FIFA bans using the tear gas in the stadiums. That resulted in a lot of behaviour that would not have otherwise shown by the crowd. People getting stuck at the exit doors and getting caught in a bottleneck that eventually led to the death of many people.
Sana Qadar: In that more recent crowd crush, at least 125 people died.
News archive: Police say the gate was partially open, allowing for a frenzied scrum of hundreds to exit only one or two at a time.
Associate Professor Milad Haghani: So that approach, that heavy handed crowd control, top down view approach is still being practised in some parts of the world. And we do still see the disastrous outcomes.
Sana Qadar: So basically how authorities treat crowds can influence how they behave.
Associate Professor Milad Haghani: Absolutely. You need to understand the sense of shared identity, the sense of us-ness. You know, when you are in a crowd where you feel a sense of shared identity with others, one unfair action by a police member, for example, can create a sense of hostility because other crowd members would see that as an offence to themselves as well. Of course, in cases of unlawful behaviour, actions need to be taken. However, the unnecessary provocation of the crowd, as soon as something happens, deploy the tear gas, for example, that is going to escalate the situation. So the art of de-escalation is being practised in some areas of the world, not necessarily in some other areas. And as a result of that, we see the differential outcomes.
Sana Qadar: But we do also know that emotions are contagious. Isn't that partly what plays out in, you know, on a large scale in crowds when things go wrong?
Associate Professor Milad Haghani: Not necessarily. Not necessarily. Yes, emotions are contagious, but not in the way that it's been framed in old-fashioned socio-psychological theories. Not in the way that it results in deaths. Yes, when something happens that lifts up the mood of a crowd, we might get some extra level of happiness by seeing our fellow soccer fans being happy. That's some sort of contagion effect. However, this effect has not been documented to have resulted in any of the crowd disasters. The disaster in Itaewon, Seoul, for example.
Sana Qadar: That disaster is the crowd crush that developed on Halloween in 2022, when tens of thousands of mostly young people arrived in a part of Seoul called Itaewon, which is a nightlife district. More than 150 people were killed, and three police officers were later imprisoned for professional negligence. With a court ruling, the disaster could have been prevented or minimised if police had prepared properly for the number of people that were expected that night.
Associate Professor Milad Haghani: It did not happen because the panic spread throughout the crowd. It was the opposite. People found themselves in a very stressful situation, and as a result, they got anxious. But it was not the anxiousness of the crowd that caused the situation. It was not the spread of anxiousness through contagion effect that caused the situation either.
Sana Qadar: Why is that so key? You've kind of touched on it, but why is it so key to understand that order of events? It's not anxiety and panic that causes the issues. It's the circumstances, the physicality of the space they're in, the action of authorities that then leads to these emotions.
Associate Professor Milad Haghani: Well, the immediate consequence of this type of framing is that it stops you from getting to the root causes of the problem. Because if the default assumption is that people die in crowds because they act irrationally, because they panic, then that exonerates authorities from looking back at what they did during that event that resulted in the deadly consequences that they observed. And that has major consequences for future planning of crowds, for future practices. And it basically stops us from learning about the lessons that are hidden in these types of disasters.
Sana Qadar: This is All in the Mind. I'm Sana Qadar. Today, challenging our assumptions about crowd behavior. Milad Haghani is an associate professor of urban mobility at the University of Melbourne. He says while panic, anxiety and emotion don't in themselves produce crowd crushes, a view that is shared by many other experts in crowd behavior. There are certain characteristics of a crowd that can make them more prone to a crush developing. And one type of riskier crowd is a religiously motivated crowd.
Associate Professor Milad Haghani: There are many reasons for that. Religious crowds are highly motivated. They need to do an act of worship. They need to do some form of, for example, dipping in the water, getting themselves to the location, depending on what type of religious gathering we are talking about, whether it is Hajj or the Kumbh Mela or other religious events. So they are highly motivated. And when you are in that mindset, you are willing as an individual to accept a higher level of risk in the hope of getting the outcome that you want. And when you're dealing with a very, very large number of highly motivated people like this, you face a great deal of risk and uncertainty. And it's a type of crowd that becomes really difficult to protect, basically.
Sana Qadar: A very quick explainer, if you're unfamiliar, Hajj is the Muslim pilgrimage to Mecca in Saudi Arabia. Attendance is often capped, but about 2 million people attend each year. The Kumbh Mela is a Hindu festival and pilgrimage. It changes location each year, rotating between four different sites, each of which are tied to a river that's considered sacred. And it's actually the world's biggest gathering, drawing tens of millions of people.
Associate Professor Milad Haghani: And it's a type of crowd that becomes really difficult to protect because of the fact that they see their act for a noble and justified cause, and they are happy to take risk for it. And this usually happens in open space areas, not in indoor areas that are structured. And that also adds to the complexity of managing who is going to what direction and who is going to do what. It is not for the lack of preparation and precaution. During the Hajj, you see a lot of precautionary measures in place. During this year's Kumbh Mela, there was no element of technology that was not in place to prevent these kinds of crowd disasters. AI cameras that are equipped with AI analysis and can communicate to a control room center. All of that was in place and still could not prevent deadly consequences that we saw during the event.
Sana Qadar: This year, at least 30 people died at the Kumbh Mela, which ran from mid-January to late February.
Sana Qadar: But you're saying if those crowd control strategies, even using AI, couldn't stop crushes in places like the Kumbh Mela, what extra precautions or planning are needed to keep religiously motivated crowds safe?
Associate Professor Milad Haghani: That is an excellent question. And that brings me to the modern way of keeping crowds safe. And the way that I try to advocate for in my research as well. So we have the traditional top-down approach of managing and controlling crowd. We know that that is not necessarily the solution because there is so much authorities can do when crowd disasters are in the making. In those kinds of situations, the actions of people can make a difference. They can understand the risk if they can come up with subtle actions that would mitigate the risk that can change the outcome. And that is something that I quite heavily pursue in my research as well.
Sana Qadar: In his research, Milad has looked extensively at how people can be equipped with strategies to keep them safe and which strategies work best. Over the years, he's done research using computer simulations, as well as running experiments with real people. Let me describe one series of experiments that he conducted over four days in September 2022. So what he did was gather about 250 recruits at the University of New South Wales campus in an event space with eight-metre high ceilings. That was key so his cameras could clearly record the movements of people on the ground. Now, Milad and his team used cardboard to create pretend walls and hallways, carving up that big space into rooms, so to speak. So participants had a more confined space to operate in. In one of the many scenarios he tested out, Milad gave about 20% of the crowd a specific instruction on how to exit.
Associate Professor Milad Haghani: So it could be something as simple as, instead of following the majority of people in your navigational decisions, try to follow the minority of people. Or instead of being rigid in your decision making, try to be more flexible in your navigational decision making.
Sana Qadar: One question, was there just one exit for people to get out of or were there multiple exits?
Associate Professor Milad Haghani: Each experiment has its own setup. It was rarely the case that there was only one exit because you need to introduce certain levels of complexity and room for people to make navigational decisions, basically. And then the same task gets repeated again for a couple of times. And you see later on in the data that the system, same system became a little more efficient. And then you do the same thing with another 20% of the crowd. The system becomes a little further efficient.
Sana Qadar: When you say efficiency, do you mean the crowds were exiting the building more quickly or more safely?
Associate Professor Milad Haghani: That's one of the key metrics. There are a range of metrics that you can use in order to assess the safety of crowds, how efficiently and how quickly, for example, they discharge a venue is one thing. The number of near collisions that they have. There are a range of indicators you can use, but they usually align with each other. So when one improves, the other ones also improve as well. And then you realize that once 50 to 60% of the crowd is equipped with a good decision-making strategy, the crowd is already gaining the full benefit that they can.
Sana Qadar: He gives the analogy of vaccinations. If a certain number of people are inoculated, you get herd immunity. Same thing kind of happens here. And so here are the messages or strategies that he found were most effective for helping the crowd exit safely.
Associate Professor Milad Haghani: If people had the mindset that when navigating the crowd, they should follow the direction that the minority of people are going to, that had a positive impact on the system, because that means a more even distribution of the crowd across the system. If people were more open to change their navigational decisions when they face crowdedness, as opposed to saying that, no, we have chosen this direction and we are going this direction, that had a positive impact on the efficiency of the system.
Sana Qadar: And in situations where he simulated an emergency exit
Associate Professor Milad Haghani: People reacting to the risk quickly and not hesitating, that had a positive impact on the system.
Sana Qadar: You might think that last one is a bit strange. Like, duh, people would leave quickly in an emergency situation. But there are countless real world examples that tell us often people will hesitate, whipping out their phones to record whatever is happening, instead of immediately leaving. One of Milad's most interesting findings, though, is how to get families or other groups to evacuate more safely.
Associate Professor Milad Haghani: When you have a crowd of solo individuals and compare it to the same size of the crowd that is composed of families and friends, the second one is far less efficient.
Sana Qadar: Is that because people want to stick together?
Associate Professor Milad Haghani: That's the whole root cause of it. But how can that be fixed?
Sana Qadar: People aren't going to leave their children behind if they're in a crowd.
Associate Professor Milad Haghani: They're not going to leave their children behind. You cannot ask people to separate. These kinds of strategies have to be sensible, have to be easy to remember, and have to be practical. People have to accept them, otherwise they are useless. I tested this strategy, for example, I asked people to hold their hands, the groups of families or simulated families, and do not let go of each other's hands. And that did not help the crowd at all. In fact, it made it a little bit worse. But then I said, this time, instead of holding hands, you hold the back of another group member's clothes, basically form a line, form a snake. And once they did that, the whole inefficiency introduced by the presence of social groups disappeared and some more efficiency was gained. The crowd even became more efficient than a crowd of solo individuals. And that gave me just the idea that these strategies that we need to communicate to people, they don't need to be so complex. They can be ad hoc strategies that can even be communicated to the crowd on the site as something is happening. This is not something that takes too many words to communicate to people.
Sana Qadar: That strategy worked because groups basically took up less space if they moved towards the exits in a line formation. So is this something you should do in real life if you find yourself in a crowd with your family or friends? Well, Milad says it would work best if everyone or a significant portion of the crowd was also informed of this strategy and was also deploying it. But even if that's not the case, it can still be a useful strategy for exiting quickly and together. So based on your own research and whatever other research exists on this kind of thing, for people listening who are wondering, you know, if they find themselves in a crowd, what they should do, what would your advice be?
Associate Professor Milad Haghani: Some of the most generic ones could be to read the room and the congestion being built in the space around them. That is for the prevention of crowd crushes, basically. And if they see that the density of the crowd is increasing, the crowd is going up without any control, best thing you can do is to remove yourself from that situation and try to stay somewhere else before joining that crowd that is getting denser and denser. That could be one thing. If you're navigating yourself in a highly crowded situation, try to be open-minded about revising your decision. If you see that in this direction, the crowd is getting really heavy and there is an alternative way of exiting, even if it's a little bit further, and it requires you to walk for longer. If people have that mindset, the entire crowd becomes more efficient. And I have tested that.
Sana Qadar: And if you happen to find yourself in an emergency situation, he says, don't stop to whip out your phone and record what's going on. Get out of there immediately.
Sana Qadar: How much our crowd crushes a risk here in Australia? We've talked about a lot of examples from overseas. How much is this a worry in Australia?
Associate Professor Milad Haghani: Sana, we are lucky in that one, we have a really good crowd safety culture overall. According to my observations, people are relatively good and crowd safety savvy. They are patient. They do not necessarily behave impatiently when they find themselves in a slow moving crowd. And also on the management side, our venue operators, our local councils, they are really safety savvy and they are keeping in touch with the latest science. That is a really positive thing that can keep us hopeful. However, that doesn't mean that we get to become complacent because we have had near miss accidents in Australia. Especially of concern is New Year's Eve celebrations in our major cities, including Melbourne and Sydney. I've made an example of religious crowds and how they are motivated and that creates additional elements of risk. The New Year's Eve crowds are also somehow comparable. Not that they are religiously motivated, but there are elements that translate into by and large similar behavior. Imagine a crowd of several hundreds of thousands of people. And there is one particular moment during the night and that's the fireworks. And after that, the crowd is not interested. Everybody wants to go home. And they are highly motivated to go home because they think that transport options could be limited. That is why I described the dynamic, you know, somehow similar to religiously motivated crowds. And that creates a high level of risk. And over the last few years, we have seen cases where we got close to accidents.
Sana Qadar: One example is a crush that started to develop in Sydney just before midnight on New Year's Eve in 2022. The crowd surged towards a vantage point for a better view of the fireworks, pushing through an open security gate and into a ticketed area in the rocks. Thankfully, security guards quickly managed to contain the situation. No one was injured.
Associate Professor Milad Haghani: We have been very lucky that these accidents have not translated into deadly outcomes. And as a result, it has created a certain level of consciousness around keeping the crowds during the New Year's Eve safe.
Sana Qadar: So while we are pretty lucky here in Australia, Milad's main takeaway is that safety is a shared responsibility.
Associate Professor Milad Haghani: It is not just the responsibility of organizers and authorities. It is also our responsibility as well to be vigilant. In fact, our vigilance makes a bigger difference in terms of safety outcomes. And being safety savvy and vigilant does not mean that it's something that can ruin our experience. It can be something very subconscious, just having these things that we just discussed in the back of our minds. That can make a huge difference in how us as individuals and also people around us are kept safe in crowded environments.
Sana Qadar: That is Milad Haghani from the University of Melbourne. We'll link to a bunch of his research and other studies as well on crowd behavior in our show notes and on our website if you're interested in learning more. And that's it for All in the Mind this week. Thanks to producer Rose Kerr and senior producer James Bullen. Our sound engineer this week was Roi Huberman. I'm Sana Qadar. Thanks for listening. I'll catch you next time.