May 25, 2022

052 - More realism in evacuation modelling with Anne Templeton

052 - More realism in evacuation modelling with Anne Templeton

You have seen these lovely evacuation simulations, the ones with a bunch of agents moving together or clumping at an exit. Ones that we use to determine ASET condition, and which are present in almost every large PBD project...

Maybe even you are running such simulations. So, with that experience in mind - have you ever wondered if what you see makes sense. We all feel that humans in groups behave differently than a bunch of units in a crowd. But to what extent that 'different' could be important? That is the question with which I have approached dr Anne Templeton from the University of Edinburgh. Anne is a renowned scientist in the field of crowd psychology. She was kind enough to tell me the difference between psychological and physical crowds and why sending students to a pub helped here quantify that. Tap into this episode if you would like to learn the new stuff we are finding about human behaviour, and how that may change our future modelling and evacuation planning.

And make sure to check out Anne's webpage which is absolutely full of resources!


[00:00:00] Wojciech Wegrzynski: Hello, everybody. Welcome to fire science show session 52. Today, we're jumping back into the world of human behavior and sort of evacuation modeling. I have invited a really great guest, Dr. Anne Templeton of University of Edinburgh where she leads her. Group a research group on identities and collective behavior, actually. That's uh, she also has a page that's called like that identities, So that's actually quite a good sort of resource already plugged in the, into the start of the podcast.

[00:00:33] Wojciech Wegrzynski: Uh, with Anne we're going to discuss. group behavior and collective behavior and crowds, which is very interesting thing when we model. Our evacuation processes. We usually observe these large groups of people moving together. But. Our approaches cannot distinguish between physical crowds and psychological crowds. And what that means you learn in the episode.

[00:00:57] Wojciech Wegrzynski: So we are, modeling [00:01:00] the groups but not really modeling the group behavior of it. And this group behavior can change the response rate of the group. It can change the walking speeds and can effect multiple parameters that we usually investigate with high-fidelity models. So it's very interesting to be able to understand these complexities. And, uh, really the jumping into this interesting field of social psychology.

[00:01:23] Wojciech Wegrzynski: To see where it's going. I'm not sure if today. We are the state where this can be modeled explicitly altogether, I guess not, I guess it's still a work in progress. But this work is very exciting and the progress is great. So, I guess you'll enjoy this episode a lot, because this is the things that are talked in here will be part of our work sooner or later. And definitely deserve more interest from the world of fire science. So let's not prolong this anymore. Let's spin the intro and jump into the episode.

[00:01:57] Wojciech Wegrzynski: [00:02:00] Hello everybody. I'm here today with Dr. Anne Templeton from University of EdinburghHello Anne,. Great to have you.

[00:02:26] Anne Templeton: Hi, thank you so much for having me.

[00:02:28] Wojciech Wegrzynski: it's a pleasure. It's a joy to have you and learn some stuff on social psychology and crowd psychology. Actually, that's a, this is a topic that, I don't know almost anything about it. So I'm more than happy to learn a lot firsthand today and share that journey with viewers.

[00:02:46] Anne Templeton: We're going to delve right in.

[00:02:48] Wojciech Wegrzynski: Yeah, let's go. Give me hardcore. No, no, no. Simplifications go, go with equations. Now equation has a lot of work in podcasts, but everything else works. So the first thing, [00:03:00] when we're talking about behaviors of large groups of people, I always wondered if the first thing that comes to my mind, It stops to be in individual behavior.

[00:03:10] Wojciech Wegrzynski: And when it starts to be a group behavior, you know, when, when I stopped being human, I become a sheep and move with everyone else. And I think that's probably a good point to start the discussion on the crowd. Psychology.

[00:03:24] Anne Templeton: Well from a kind of psychology perspective, we always talk about physical crowds and psychological crowds. And the distinction is that a physical crowd, a physical crowd is something that you might see in a shopping center or a transport hub, where you have people who happen to be there. At the same place at the same time, maybe with their friends or family, but otherwise they're not really connected with the others in the crowd.

[00:03:52] Anne Templeton: They just happened to be there a psychological croud on the other hand, that's where the group processes come in. And that's really what we focused on in [00:04:00] quotes psychology. So that is where people in the crowd feel like they're part of a group with others and they feel. that others are in the same group as them.

[00:04:10] Anne Templeton: And that's where we start to see collective behavior. And in sort of psychology, we talk a lot about social identity theory. This is the idea that we have personal identities. These are things are idiosyncratic to, to ours as individuals. So for example, I have a squint. eye, I am not a morning person. That kind of thing at dark too much.

[00:04:34] Anne Templeton: My eyes, no.

[00:04:35] Wojciech Wegrzynski: Yeah.

[00:04:36] Anne Templeton: But then we have social identities and these social identities refer to our memberships as part of social groups and that social group can be anything. So it can be, you know, you're a fan of a particular sports team that you are a runner that you're an academic, you're an engineer that you're a firefighter. And these social groups are important because they tell us a little [00:05:00] about how we should act. In certain situations. These groups have definitions, they have values, have got behaviors associated with them. Now that's not to say that we're not acting as individuals. We'd all lose our kind of ability to reason to be rash.

[00:05:20] Anne Templeton: It's just that what's important in that context when we're in those group situations and we feel like part of a group is we're guided by what it means to be part of that group. And so what's important to the group is important to us because we're part of that group.

[00:05:33] Wojciech Wegrzynski: So that would mean a. crowd of a given number in the shopping mall and the crowd of the same number in an office where everyone knows each other will be from your perspective, a completely different crowds.

[00:05:43] Anne Templeton: Yeah.

[00:05:44] Wojciech Wegrzynski: And I'm giving them almost the same brief equation time, treating them as, as, as old as the same mass of people.

[00:05:53] Anne Templeton: Yeah.

[00:05:53] Wojciech Wegrzynski: so, okay. This, I understand why you guys classify it under complexity, because it very quickly becomes very [00:06:00] complex. Uh, when you're a member of this, let's say random crowds, not, not, uh, group. do you, as scientists still observe these groups, behavior? I don't know. Maybe group is forming, like, uh, And lost when they, are stuck together on the island and they form a group, a bunch of strangers, maybe,

[00:06:22] Anne Templeton: That is a blast from the past. I haven't seen Lost in a long time. Excellent series. it's a great point. So for us, that kind of distinction between the physical and psychological crowd is in the psychological crowds. There's that sense of being part of the same group. So we do see collective behavior and physical crowds because you still see families and small groups of friends, for example, right.

[00:06:43] Anne Templeton: But that's very much subgroup behavior. It's not kind of widespread And your example there of a kind of, in an office environment, if you have an evacuation and if that would be different from say something in shopping center, I think that's really interesting because there's been a ton of [00:07:00] work there.

[00:07:00] Anne Templeton: Been done on evacuations and emergencies, where you've had people who were in physical crowds that have become psychological crowds in the emergency because of that common feet of the emergency. So there's been great work by new John Drury, Chris Caulkings, Steve Reicher they did some fantastic work with survivors of the London bombings, uh, July 7th, 2005 London bombings.

[00:07:27] Anne Templeton: And, uh, have you ever been to London? Have you ever been in London underground?

[00:07:31] Wojciech Wegrzynski: Yep. Fantastic place.

[00:07:32] Anne Templeton: yeah, it is, but so at commuting time, which is when the attacks happened, when you're on the tube, you don't look at anybody else, you don't interact, you don't want to catch someone's eye. It's awkward. Right. You're just there trying to get to your job just going about and trying to find a seat.

[00:07:49] Anne Templeton: And so that's a physical crowd. That's commuters. They're not connected. So actively unconnected. But when the emergency happened, when the bomb was detonated, [00:08:00] people started to come together because they felt like they were part of the same group. And so John Drury, Chris Caulkins do your brochure. They conducted interviews media.

[00:08:10] Anne Templeton: All of a reaction slant bombings, and they found again and again, that when the survivors were explaining, helping behavior, like helping others too, if that can we going back into dangerous areas, providing first aid, tying tourniquets, providing emotional support water, all sorts of stuff. They all said it was in the basis of feeling part of the same group.

[00:08:30] Anne Templeton: So it's that feeling part of a group was motivating people to provide that support and help us evacuate. Not everyone felt part of the same group in all emergencies. I'm not seeing that there's going to be variability, but we see this continual theme of the public and emergencies coming together. That transformation from the physical crowd to the psychological crowd in this collective behavior on the basis of that group membership.

[00:08:54] Anne Templeton: And I think that's, that's

[00:08:55] Anne Templeton: amazing.

[00:08:56] Wojciech Wegrzynski: yeah, it is. And now outside of it being, interesting and [00:09:00] fun, let's talk about how, how useful it is. Like, do you, as a scientist, observed that. Uh, a physical crowds would have a different, I don't know, response characteristic or different like overall behavior over a group crowds

[00:09:17] Wojciech Wegrzynski: in the macro-scale .

[00:09:18] Anne Templeton: I think that

[00:09:19] Anne Templeton: that's

[00:09:19] Anne Templeton: a great question. So, I can talk about work. I've done on collective behaviors, just on walking behavior, home feeling part of the same group

[00:09:26] Anne Templeton: impacts. so staff that during my PhD at the university of Sussex I conducted a few experiments on walking behavior.

[00:09:36] Anne Templeton: So pedestrian movement and this kind of came about because we see pedestrian models, computer models of pedestrian behavior who are simulating, how people navigate space and that's all grand. But we thought from across psychology perspective, one thing that was missing was group processes is feeling part of a group with others.

[00:09:57] Anne Templeton: And so one of the experiments I did. [00:10:00] Was I went into a second year of psychology statistics lecture, and I recruited a bunch of participants to take part in a psychology experiment. I said that they were being recruited because there were Sussex psychology students. And we had these black baseball caps that had like Sussex psychology and the psychology logo on them.

[00:10:24] Anne Templeton: And we've recruited them to walk to the student bar. Now, the reason that I did that was because I wanted to see what the movement behavior was when people felt like they were part of the same group. And I kicked directly compare that walking behavior. to footage I'd take. So I filmed it from above and I could directly compare that behavior to fish taken previously of the same people coming out of lecture theater when they hadn't been primed to feel part of the same group.

[00:10:54] Anne Templeton: And so we could see this difference between when there was a fiscal crowd of friends who've been in lecture theater, [00:11:00] but otherwise pretty unconnected compared to what we were praying them to feel part of the same group as Sussex psychology

[00:11:06] Anne Templeton: students.

[00:11:07] Wojciech Wegrzynski: So you turned them into group by giving them hats like a common identity.

[00:11:11] Anne Templeton: Yeah, we primed that existing identity. So they're already Sussex psychology students, right. That's why they were in that statistics lecture. So we just made that salient. So that was important for how these sold themselves and saw others at that time. And we gave them the hats, not only to remind them of their own identity, but so they could see who else was in their group.

[00:11:33] Anne Templeton: And for when I was, filming footage from above a bridge, I could see who was taking part in the experiment for tracking them. And what we found was that people who felt part of a group when we had this psychological crowd, they altered their behavior. So in terms of physical proximity, they moved much more closely together.

[00:11:53] Anne Templeton: And when we look at subgroup size, And the fiscal credit, when we hadn't done any manipulation, we [00:12:00] had maybe people in pairs, maybe threes or size of groups of three or four, but then in the psychological crowd, it was much larger. You were seeing like eight to 10 people. And actually it was really hard to divide them into subgroups as they're moving.

[00:12:14] Anne Templeton: So they're always chatting to each other coordinating. And so there were much apart from being much kind of physically closer together. They also reduced their speech cannot walk together and they walked a further distance, not from ATB, but actually between their different steps, because there were always kind of coordinating around each other to stay together as a group.

[00:12:34] Anne Templeton: So even just on that basic level, when you feel like others, a part of your group, we saw there was this collective coordination of them. Just try and stay to.

[00:12:43] Wojciech Wegrzynski: And what if crowd is consisting of bunch of medium sized groups? Like you'd have a first year lectures students second year, third year, and they will all associated with each other, but not with the others.

[00:12:56] Anne Templeton: Yeah. So I think that's what we saw in their physical crowd, [00:13:00] because it was the same participants, same people, right? From the same lecture. It was just a few weeks before I make you see people who were in those small groups. So you can see those kinds of friendship groups. And so they were more physically.

[00:13:13] Anne Templeton: You could see very distinctly these clear cell groups and they want faster and less distance because they were just kind of like going to their end point and not really coordinating around each other as

[00:13:23] Anne Templeton: much.

[00:13:24] Wojciech Wegrzynski: And, we're investigating anything in terms of how they respond to information. Like I really wonder if group has, is processing that the information as a group or each individually is processing, like there's a fire alarm.

[00:13:39] Anne Templeton: Well, that that's a really good question. So we didn't look at it in that experiment, but I've recently been doing, interviews and surveys with residents of high-rise buildings. And we've been specifically looking at that rule of group processes and how they decide to, to respond in emergencies. And we found quite a lot of stuff has come [00:14:00] out with these interviews and surveys.

[00:14:01] Anne Templeton: But one of the things that keeps coming up is that people who feel. In the same group, as others in the building are more likely to seek them out in the building. The more likely to go and look for them, you know, stick their head into the corridor, ask others what to do. They can seek information from each other about how to react and share information with them as well.

[00:14:24] Anne Templeton: And they will even go further into the building in the event of a fire. to help other group members to help people you see as being in their group, but also to get information from them. And so we see this kind of, this is, this is self-report. This is people saying either what they would do in an emergency in the event of a fire.

[00:14:42] Anne Templeton: Or what they had done previously. Let me see this again and again, and it's not only that people will go on and look for information that they'll delete that commuting they'll give information is very much related to how much they feel in a group with others. And they will trust information from people.

[00:14:59] Anne Templeton: They feel a part [00:15:00] of the group more than people they feel. And that's tricky, right? Because it might be that you have, I don't know what a fire warden, for example, giving you information and you're getting contradictory information from your direct neighbor about the fire. And if you trust your neighbor more, if you see them as more part of your group, more self relevant, they have more direct information that's relevant to you.

[00:15:23] Anne Templeton: There's a chance you can listen to them. If you have lower trust of that firearm.

[00:15:29] Anne Templeton: It's quite an

[00:15:29] Anne Templeton: extreme example.

[00:15:31] Wojciech Wegrzynski: no, no, it's not extreme actually. It's, it's very down to earth because I think we've entered, um, Useful engineering teritory. I often end up doing evacuation studies for high-rise buildings. When the high rise building is, just being delivered. It's not sold yet. And the way how they sell, high-rise buildings, at least in here. Usually this there's company, a that would buy like 10 floors there's company B that will buy seven companies [00:16:00] would buy 20 or something. I didn't know. but th th the evacuation scenario, especially when we go into selective evacuation, like floor by floor to limit, the waiting time, it's insensitive, you know, to this, possibly. Group identity or group interaction, you know, w there's a good chance we would evacuate floor 15 of company, A with floor 30 of company B together. and there's a good chance. We would evacuate one company for like very long. Time. And there's a good chance this people, because they work in the same company, they could already communicate, between them and, either realize it's, it's a fire they need to escape or, or figure out is a training.

[00:16:44] Wojciech Wegrzynski: And they don't want to be bothered with this stupid gnosis to go out, get out of the building. And in both cases, it challenges our design concept. Like I think it's us who did that? Take the, designing this in a way that, [00:17:00] is very artificial compared to the, to how real world, works. You see any way, how, or are you working maybe on the way that this could be incorporated in the design of,

[00:17:10] Wojciech Wegrzynski: tall billing of equations stills?

[00:17:12] Anne Templeton: I think something that I keep finding is that. Social influence is so important. So who's looked to for information and how trusted the information is, keeps coming up again. And again, when people are saying how they would respond to the emergency and on one hand that is, you know, doing evacuation to seek information from, from people in your group or to give information or help them to leave.

[00:17:37] Anne Templeton: But another thing is when you have residents of high rise buildings, they're given safety information, right? What to do in the event of a fire. And that could be a leaflet to the door. It could be a training session. It depends. And what I found, especially in interviews was that often if the information was coming from something like attendance management association, who the participant [00:18:00] didn't really see as being on their side, they were just as kind of other, other, putting all leaflets through the door, they would ignore it.

[00:18:05] Anne Templeton: But Gordon, the bed believe that we've got strict. But if it was like, if it's training say or information from fire and rescue services, the pay retention, because they felt like fire rescue services were acting on their behalf, you know, actingly for the group. And so we see this of. Who's listened to whose information is listened to and trusted is really impacted by those perceptions of who is in the group.

[00:18:31] Anne Templeton: And who's acting on behalf of the group. And this also is particularly important. You mentioned there about, you know, kind of phase evacuation, a lot of the high rise buildings that we have have stay put policies in place and where people have to stay, where they are in the event of a fire, because the buildings are supposedly built to be safe and people can stay. There is all well, I found that there was so little trust in that. And so if you have, this leaflet [00:19:00] saying, stay, put and totaled otherwise, and you have a neighbor, your friend friends saying, no, we need to evacuate. You're going to listen to your neighbor. Now there is a lot more that goes into it, like trust in the building.

[00:19:11] Anne Templeton: For example, knowledge of the guidance, all this stuff is important. I don't want to disregard that. But whose information is trusted again, is impact appropriately relations. And that is so important for thinking about how to roll out evacuation guidance, or even what to do in the event of a fire type guidance.

[00:19:27] Anne Templeton: If we're looking at models of evacuation behavior, think putting these group process in are so important, especially because we keep seeing in emergencies, people who are previously unconnected can come together as a group, and that then affects their behavior.

[00:19:41] Wojciech Wegrzynski: brilliant. So essentially we can, as engineers, we can put a lot of effort into designing, strategy for a building. Phase evacuation could be one, one example, stay put policy could be another one that you've brought, but I can see more like, uh, use or do not use [00:20:00] elevators. You know, there can be a building where elevators are using evacuation, but people will not use them because they are all lives.

[00:20:07] Wojciech Wegrzynski: they were told that you not use elevatoring in case of a fire. Yeah. There, there may be some sort of evacuation, floor where. People can survive fire with enhanced, capabilities. So there's a lot of, possible scenarios where you would like people to do a very certain thing.

[00:20:27] Wojciech Wegrzynski: Otherwise they are in danger potentially.

[00:20:30] Wojciech Wegrzynski: And now,

[00:20:31] Wojciech Wegrzynski: Now we need to make the people go there. And you say that the trust of the trust to information of the trust who gave you the information may be actually a critical factor in there. So for us as engineers designing the information is one way, but the way how we pass it, can mean yes or no for the results.

[00:20:52] Wojciech Wegrzynski: Uh, So I, I think what you're saying is not only interesting, but I think this should be used in designing how we convey [00:21:00] information. Like if you know, group has leaders or people who people trust, it's this probably reasonable to, to use the leaders to convey the information

[00:21:11] Anne Templeton: Well, it's an interesting question. I mean, talking about leaders, what we often see is that there isn't necessarily a clear. Leader as such in terms of like, you know, you might see that some insecurity was, was a leader or there was like a kind of leader church group, something like that. We don't tend to see that.

[00:21:32] Anne Templeton: We tend to see a lot of this kind of collective. Um, coordination. That's quite to self-organization where people are collectively deciding how to respond. Obviously that's a bit different if you think someone has more information, if they know the guidance more like that, that does have an influence, but we don't really see kind of clear leaders come out as much.

[00:21:52] Anne Templeton: It's a lot of coordination. And however, in terms of. Kind of guidelines that [00:22:00] include these group processes through there's two things. So one is that I also did interviews with fire rescue services, myself Claire Nash. She actually conducted the interview. She's brilliant, , and fire and rescue services.

[00:22:12] Anne Templeton: So we can not to think it was 22 interviews, but an hour long, each individual interviews and fire rescue services, , they were saying. they see how much the public wants to help and how much they want to buy, provide support. And they also see this importance of building good relations with the residents so that they are listened to when an emergency happens.

[00:22:37] Anne Templeton: And not only so like building those. Can I have positive relations with residents and keeping that kind of community building, which is so important, but also they were right on the money when they were saying that they don't only give guidance. See, this is what you should do. They also say, this is why you should do it so that people are able to make informed choices and not so important as well.

[00:22:58] Anne Templeton: Okay. People to here. [00:23:00] Holly Carter it public health. England has done a lot of work on this with professor John Drury at University of Sussex. She's done some fantastic experience. Field studies where she has altered the type of communication that's been given by first responders. This was an, CBRN decontamination experiment.

[00:23:22] Anne Templeton: And she altered the type of communication that was given to the people going through the decontamination from the first responders. And she found that depending on the type of communication you. Influenced the extent to which the people followed the guidance. And this is because the guidance that was given impacted the perceived legitimacy of those first responders, like how logistically they were in their rules, the Sheriff's social identity, but the responders, therefore how much we feel like they're part of the same group as us.

[00:23:56] Anne Templeton: And then how quickly they went through the decontaminate. Especially with [00:24:00] a bunch of other things as well, but the kind of key take home point from this study that I really love. And I think her work is so convincing. She trains us again and again, is that when you tell people exactly what they need to do and why they need to do it, and you give them updates about what's happening, they see the guidance is more logistic.

[00:24:20] Anne Templeton: They see the people giving the guidance as being part of the same group you're working together and they're more likely to follow the guidance. So at its most basic level, I think that telling people what they actually need to do and why they need to do it. So they buy into it is so important. And often we see in emergencies, that information is maybe withheld from the public about what to do, or there's a delete before we get that kind of response.

[00:24:47] Anne Templeton: And often doesn't necessarily say why you need to do the information it's not highly informative. And that can, that can have very negative consequences because we also see in emergencies [00:25:00] that again and again, that people will start to collectively coordinate through and response. They'll start helping each other.

[00:25:07] Anne Templeton: There's this great need to do something to fight help. And so on one hand they can be doing amazing things, but if they haven't got the full information, Then that can be tricky, that can be tricky to work with. So I would argue it's really important to give the information to people so they know how to respond.

[00:25:23] Anne Templeton: They know exactly what they're meant to be doing. The group processes so much comes into that

[00:25:27] Anne Templeton: as well.

[00:25:28] Wojciech Wegrzynski: And, if we turn into not positive, but negative effects like, uh, have you also observed this group behavior that led into some sorts of disasters? I don't know. Maybe people.

[00:25:42] Anne Templeton: Oh,

[00:25:43] Wojciech Wegrzynski: Choosing to stay put based on the, well, you've mentioned that before that, that you can ignore some wording because the group sets.

[00:25:50] Wojciech Wegrzynski: So, but do we have like evidence that a group behavior could have been a ruling factor in some disaster, because that's also a way how we can prevent them. Right.[00:26:00]

[00:26:00] Anne Templeton: In terms of, in terms of theory, I think. It could happen and we see people delaying evacuation to help others. But then who's to say that that wouldn't actually lead to a more effective evacuation because you get more people out. It might be slower, but you get more people out in the end. So it's tricky.

[00:26:19] Anne Templeton: One thing. That does tend to happen is you see this in aftermath of Grenfell for example, the public really, really wants to help. And so they'll give all the resources and unless it's coordinated because of that needs to help, it can be the, can you have a bunch of resources? You're not really sure what to do with, if you look at.

[00:26:40] Anne Templeton: But for the Ariana Grande concert in Manchester arena, where there is the attack. So the care slack report that came out in the aftermath of that, where they looked at accounts from crests, ecologists looked at council first responders, everyone. And what are the kind of key take home messages of that was how much the public.

[00:26:59] Anne Templeton: I [00:27:00] wanted to help and wanted to apply first date and everything, but didn't necessarily know how to do it. And so the argument that the carousel report made really clearly was give them the resources to do it properly. I think that's the key thing. If you have people who want to help make sure they're able to do it safely and in the right ways, and that involves coordinating with them because often the public will know a lot more about what's going on as well, because they'd been there for a moment.

[00:27:27] Wojciech Wegrzynski: You wrote a paper about more realism in pedestrian behavior models?

[00:27:33] Wojciech Wegrzynski: Uh, I love the title. my first thought we really want realism. That's, that's a, that's, that's a tough choice, you know, because really it adds complexity. And, from my perspective sometimes reduces utility. Of the simulation, like, sometimes in precise simulation can give you more knowledge than a very precise and beautiful study, which leads nowhere.

[00:27:57] Wojciech Wegrzynski: that, that actually could, could be my, my first [00:28:00] question about the modeling, uh, what's the value in more realism in pedestrian behavior models? How do you view that?

[00:28:07] Anne Templeton: I mean, that's a great question. And I fully agree it was a bold title. I was actually updating my website a while ago and had to put in paper titles and I remembered one that is humans do not always act selfishly. And I was like, I'm seeing a theme I've just bold titles, um, for, to answer your question in terms of. Getting more realism and models. Now, I don't know if we can ever get full realism and model for X. There are so many assumptions that we put in, but I think something that's missing is. Evidence from crowd psychology about group behavior. So we have run lots of experiments on how people react and close.

[00:28:44] Anne Templeton: We have so many surveys that we've had with people at credit events. We've got this huge, huge wealth of knowledge about how group processes impact behavior, but until quite recently, That empirical research wasn't being put into pedestrian models. We [00:29:00] did have a lot of, group behavioral. It was comforted models, but a lot of it was based on assumptions of behavior or atheoretical accounts or behavior or old theories that have been largely outdated and described Sydnor a better explanation.

[00:29:16] Anne Templeton: And so something I really wanted to kind of do in my line of work was looking at, okay, so how can we look at the role of group processes in a way that informs pedestrian models? And until a few years ago, they were very different disciplines, but now there are lots of people who are interested in this, which has been fantastic to see this trend and change, but it's tricky because we have to think, okay, so.

[00:29:39] Anne Templeton: What parts of group behavior do we want to put in the models? Because we can't put absolutely everything that gets really complicated quickly. We need to kind of find a starting point. And what evidence do we have that we can kind of measure the models against to see how well they're actually simulating this behavior.

[00:29:57] Anne Templeton: And so that's why I tried to design experiments [00:30:00] like the walking behavior experiment. I told you earlier, when we, when we were paying people to feel like they were part of the same group, the idea is that we want to kind of look at, if we manipulate this one thing about group processes, what effect does it have on behavior?

[00:30:15] Anne Templeton: And you can supplement that with surveys, for examples, get motivations, et cetera. But really it's just finding a way of running experiments, Rican say, all right, if we just manipulate this one thing, what happens? And then how do we put that into model? How do we even put that into an algorithm? How do we then test it?

[00:30:35] Anne Templeton: What I love doing is saying, all right, I've got this feature CRO behavior, got these stats or proximity. This to distance from each other in proximity, we've got speed. Um, all those kinds of information about mental who was in what group, if we run the same scenario in a model, can we get this similar behavior and then come to the simulation?

[00:30:56] Anne Templeton: Can we match it up to real world behavior? [00:31:00] And so far, I think it was still got a long way to go. But it's a really interesting challenge, how we condense all this huge theory. That's what I was trying to get at the paper is how we condense this huge amount of theory and all this knowledge from social psychology and especially across psychology into a model.

[00:31:17] Anne Templeton: How do we test these bits in a manageable way? And that's. I absolutely love this work. This is one of my favorite things, because it has involved working with pedestrian, modelers, computer scientists, mathematicians, fire engineers, even designing like shared vocabulary because we found that we had very different definitions of the same word.

[00:31:36] Anne Templeton: So it should be, conversation is very different. To try and work out. Okay. How do we do this? And it's led to so many exciting collaborations, like within Ricoh, who you had on one of your previous podcasts, didn't work with him on evacuation behavior. We use virtual reality experiments to try and manipulate.

[00:31:53] Anne Templeton: Who in this scenario is seen as in group or out group who's part of the group or not, how much do we attend to them? How does that influence our behavior? [00:32:00] This work has slipped by a John Drury URC Sussex again, I'm doing that work with Fergus Neville. Uh, or at St. Andrews.

[00:32:08] Anne Templeton: And that's a really nice interdisciplinary multi-disciplinary project that is just a joy to work with. Cause that's really trying to get. Data that we can look at and then lead on down the line. We say, all right, how can we use this to inform models and test models against it?

[00:32:23] Wojciech Wegrzynski: That affected I would identify like a sub model where someone takes decision a sub model of walking, a sub model of que formation. So is it on all of these fields where this crowd psychology or social identity can influence the

[00:32:43] Anne Templeton: Let's see, I think this is a great question because before I used to think, okay, you have this locomotion model just as pedestrian movement model, then you have this kind of psychological layer and that's work that we did with Garth Costa and Isabella von Sivers And actually, when you think about it, so these group [00:33:00] processes impact or perceptions as well as their behavior.

[00:33:03] Anne Templeton: And so you can, it influences how we perceive the information we receive, how we process that information and how, how that then affects our outcomes. So I think it really needs to be embedded and many steps of the model. And it's tricky to work at where the best place. To do that. So for example, in terms of imagine you're modeling, high-rise building a fire goes. You might assume, okay. Everyone immediately evacuates or everyone follows state. But actually what we find is that people aren't really sure quite how to respond. And so they'll look to each other for information and then decide how to act. So they see the stimulus, they see this fire through then seek information about it.

[00:33:46] Anne Templeton: Try and work at her threatening is more behavior to actually take. And then the act. So we kind of need it at those different stages. I wish I had a concrete answer to like, here is

[00:33:57] Anne Templeton: exactly where I'm still puzzling out.

[00:33:59] Wojciech Wegrzynski: It's a [00:34:00] developing field, like rapidly developing field. So I don't think anyone expects like solid answers. Like we, we understand It is so different. from what are we doing today? You know, it also, there, there needs to be concrete evidence for us to switch our behavior because we know today. If you're a consultant, there's a very specific way you approach a project related to evacuation. You're essentially interested in maybe three things. One would be the total evacuation time. Second would be bottlenecks. The third would be the waiting time. So you want to optimize for those three? Maybe, maybe a little more, but, this are the points you, you would look for in, in your evacuation model.

[00:34:42] Wojciech Wegrzynski: and now. I would like to have more realism in the simulation, the question is how that translates to the answers of the, of the model. But also it can translate to the fact that my questions are stupid and I [00:35:00] should have been asking my model something different, you know, because, maybe I should not be interested in, the evacuation time.

[00:35:06] Wojciech Wegrzynski: Maybe I should be interested in, finding probability in which a certain chain of events occurs that blocks my, have a question completely. And that's a problem I should solve. And the not prove for the. Sorry I'm that it takes one hour to walk 30 floors down. so, so, so so I, I think the way how this field develops it, it opens a lot of like, it opens a lot of doors, but again, I, I'm not sure if everyone should jump through them already immediately because, we need to first understand what we want to gain from that.

[00:35:40] Anne Templeton: Absolutely. Absolutely. And I think you started off there where you made a point, probably like. This current psychology wasn't even really kind of acknowledged as such until pretty recently. So I remember it feels 2014. I went to the Pedestrian Evacuation Dynamics conference for the first time, [00:36:00] and I felt like such an outsider because I was kind of social psychologists .

[00:36:05] Anne Templeton: Amazing mathematicians Great be modelers, but. The only kind of social psychologist. And I was there presenting work with Isabella von Sivers where we were like, great processes are important and here is how you can start to put them in a model. I'll make all so many questions about why is this important?

[00:36:22] Anne Templeton: Why do we need this? The model is already working. Why isn't these assumptions? So cry behavior. And there was a huge uphill battle to kind of show why this is important. And luckily, crowd psychology has been going from strength to strength, which has made the argument quite easy, but really you need that dialogue.

[00:36:39] Anne Templeton: And I mean, that was the pen 2014. This year, I gave a keynote with a PED about group processes and that. Never would have imagined that granted us, cause they were trying to like four grudge, early career researchers, but I never would have thought that would happen back in 2014. So it shows how far we've come.

[00:36:59] Anne Templeton: And I [00:37:00] work with Movement Strategies with OFR Consultants, people in industry who are developing these models and taking them to, large stadiums to large housing complexes across the UK because they're interested in it. But. We are all still collectively puzzling out, how do we apply this? What is it we actually want to gain from these models?

[00:37:21] Anne Templeton: How, what do we want to know about group processes? And so on one hand, I think that. From crowd psychology. We can inform these models so much, but also I've learned so much from these conversations and they've opened up so many questions that I hadn't thought of and ways to research group processes, new methodologies, for example, virtual reality.

[00:37:43] Anne Templeton: Right? So all these new methodologies I think, are really informing the shape that crowd psychology has taken in the direction that it's taking. One thing I've been trying to do when. Looking at, okay. What do we actually want to do in this model and how you get processes [00:38:00] impact? It is thinking about what are kind of key outcomes are.

[00:38:03] Anne Templeton: So whether that is evacuation time, right. Or if it is, response time from getting information. And working out how we were grouped processes might fit into that in terms of theory and building experiments to manipulate it. So in the experiments we have could have behavioral outcomes of if that curation time over response time that can then inform the models.

[00:38:24] Anne Templeton: So it's really working out, okay, what is this discipline do like an evacuation behavior of what are the kind of keep it so important to them? How can we work together? Like how can we inform each other? And there's a lot to do, but it's a really exciting time to

[00:38:37] Anne Templeton: be.

[00:38:38] Wojciech Wegrzynski: I think if I reverse where they've said previously, what can you get from current model, which has a pre evacuation time distribution as simple agent movement, que formation model in the end, what I've said the time, the time you've waited for two exits. the total evacuation time, the capacity of your stairwell, [00:39:00] that's it.

[00:39:00] Wojciech Wegrzynski: You cannot get anything further from it. And it's not that you can improve that model by adding a one simple, small sub model and it will solve and suddenly give you an insight. I think when you step into crowd psychology. Like you need to implement almost the whole thing other, otherwise it doesn't make sense, but then you get, get the capability of understanding, not just, finding a, an answer, which is the number.

[00:39:30] Wojciech Wegrzynski: And when you understand what's happening with our challenges with your building. You can take informed design decisions. You can, design procedures for the building. I don't think we're doing a really great job in that today, you know, designing procedures for buildings and, in the end, if, you know, if. design a smoke control system in an atrium or whatever. And I tell you, it's going to give you 10 minutes evacuation time for me to go to like 12 or 15 minutes. It's [00:40:00] usually a huge redesign with a large investment costs. But if you understand the crowd psychology and you're in your building and you could act to limit the time required to escape from 12 minutes to 10 minutes, maybe you are way more efficient than I am in, you know, spending the money on safety of the building.

[00:40:19] Wojciech Wegrzynski: And this can be a feature of safety of buildings. So I, I see, the use of this field. Absolutely.

[00:40:27] Anne Templeton: I think. What will you see again and again, in emergencies from crowd psychology is that the public can be an amazing resource for good. They're so good at sharing information, sharing resources, supporting each other. I don't know what the best way is to harness that other than to make sure that they have the resources and knowledge needed to do it well.

[00:40:49] Anne Templeton: And I think more work is needed to do that. So I also, I don't want to say that their current models, those kind of evacuation models aren't working because the art, they clearly are. You have some [00:41:00] really, really great models out there that can predict behavior really well. But if we want to make them even better for the sake of safety plan, If we're saying that group processes keep impacting this behavior, we should be modeling them as well.

[00:41:16] Anne Templeton: That's my main argument is that word's doing well, but we can keep doing even better if we get more knowledge.

[00:41:22] Wojciech Wegrzynski: I'm here also close to the world of buildings, but the world of crowd mulling is, is much, much larger, large outdoor events. Um, wildfire, evacuation, I guess these are also the fields where, this group behavior may actually be even stronger. Yeah. I mean, Let's go wildfire. You have families that would escape together.

[00:41:41] Wojciech Wegrzynski: You have neighbors who would be escaping together. I think these bonds could be even stronger than, than your floor in their office. Right?

[00:41:49] Anne Templeton: I wonder, I think it's an interesting question. One thing I've found with high rise buildings is that. You can't always see her, everyone else's [00:42:00] reacting because you're in different floors. You might have a different layout. And so people will go and seek others to see how they're reacting. On the other hand, if you have something like an attack on our busy street, a busy city street, you can see how everyone was responding on that street is really very clear, right?

[00:42:19] Anne Templeton: And I think there might be something. This is an interesting research question for future is. The extent to which you can see others reacting and how much you feel like they're part of the group, how that then impacts response. So I think for me, it doesn't matter what group people are in. If they feel like part of a group receive this kind of heightened trust of those people in the group, going to them for information.

[00:42:44] Anne Templeton: But there is something there, an interaction with the kind of physical environment and their ability to do. I think that that's something interesting. So something of kind of keeping in mind, especially as we've been moving to these built, high-rise building evacuations. I don't have a concrete answer, but those are my kind [00:43:00] of initial speculations

[00:43:01] Wojciech Wegrzynski: I would take it even further. You know, I feel that really interests me personally, struggles, road tunnels, and, how to make people leave their vehicle and escape tunnel. That's a, I think a hell of an issue, which I don't have a really great answer to. So also maybe, crowd psychology could help us understanding. Collective behavior of drivers in a tunnel and how to reach them, especially in a situation where you would like, if I keep the people who do not see a fire, you know, that that sometimes happen. If you have a very long tunnel it's, the drivers can stop far away from the fire and they don't even know they're under the threat direct threat.

[00:43:43] Wojciech Wegrzynski: And you would like them to leave the vehicle and go away

[00:43:45] Anne Templeton: I think that self relevance of the threat is really important. And that perception of is this immediate threat or not, I think is, is key. what we've seen in crowd psychology is when there's that perception of the threat and the common fee, that's where we see the collective [00:44:00] behavior in. But if people don't see it as self.

[00:44:03] Anne Templeton: So you don't see it as affecting them. You might see that that connection to the others might not be happening. I know there's been great work done in tunnels about types of signage and how visible is and how much that influences others. One thing that comes to mind to, to use the example of, or even talking about quite a lot, but I think it's a great example is in the London bombings, that was in a tunnel, people were stuck between stations.

[00:44:27] Anne Templeton: It was dark. They couldn't see what was going on and. They were helping each other to evacuate the road, going, like taking people out towards the platforms, going back in helping others, because they felt that everyone was in the same boat that they didn't know if another bomb was going to go off merchant services couldn't get in.

[00:44:46] Anne Templeton: It was very much, we are under threat right now. And so how do we help each other to get out in something like a tunnel? If you can't see the. I think that's a very different situation because the threat isn't necessarily [00:45:00] present at that exact time. Unless you have a way of communicating. Why is a threat and what, what to do?

[00:45:06] Wojciech Wegrzynski: Yeah. we are designing both alarm systems for tunnels now. So it gives us a direct line of communication between the, center of operations and the tunnel, up to a level where you can call people by a gentleman in red shirt, please evacuate that. That's. Kind of helpful, but, yeah. Thank you.

[00:45:24] Wojciech Wegrzynski: That, that, is, you know, food for thought. That's, really good. And I hope, that your field grows and develops and, it leads us to, much better models in, in pedestrian dynamics.

[00:45:35] Wojciech Wegrzynski: maybe you want to share some resources or tell the audience the next step to, to learn more about this.

[00:45:42] Anne Templeton: Absolutely. So all I'm going to tell him that this is a shameless plug for my website,

[00:45:49] Anne Templeton: but my,

[00:45:49] Wojciech Wegrzynski: a little price to pay for a great guest in the podcast.

[00:45:53] Anne Templeton: I just made it. And this is so exciting to me that I have a website it's Identities and Collective behavior. So I'm updating that with [00:46:00] lots of resources and work republish work. My students are doing talks and things like that that have been recorded all available. If you want to learn more, there are so many fantastic researchers and cross psychology like has done great work.

[00:46:13] Anne Templeton: On, uh, communities in Grenfell uh, the aftermath of the emergency and how they came together to support each other, sort of a surrogate has some fantastic work and collective action. If you're interested in that and cross psychology and John Drury has led. Uh, really game changing team of researchers at the university of Sussex who are doing all sorts of stuff in cross psychology, particularly in evacuation research interests in his look at his work.

[00:46:37] Anne Templeton: Chris Caulking has done a ton of work with emergency services, looking at their interactions with the crowd and the public in emergencies. Holly Carter public health. England has done a lot of work with emergency services and those communication approaches hoping impact behavior. Steve Rusher has been, I would just keep naming

[00:46:54] Anne Templeton: names, but there are so many people.

[00:46:57] Wojciech Wegrzynski: Um, I'm taking notes[00:47:00] that's really a really, really great. okay. thank you. Thank you so much for coming to the podcast was joy to have you here. And I hope many people are no more, open to crowds, crowds, psychology. And then then more informed about the role of social science and in fire.

[00:47:17] Wojciech Wegrzynski: And I think it is profound. Thanks so much for coming and,

[00:47:20] Anne Templeton: Thank you so much for the invitation I have thoroughly enjoyed this minute. Great conversation. So thank you

[00:47:25] Anne Templeton: so much for the

[00:47:26] Anne Templeton: invitation.

[00:47:26] Wojciech Wegrzynski: And that's it. Your crash course into the world of social psychology and crowd behavior. I hope you've enjoyed it as mentioned in the intro and by Anne there's a webpage where she collects all her research and publication resources to share. It's and I've checked it. It's really great.

[00:47:50] Wojciech Wegrzynski: You should try the swell. If anything, in these episodes was interesting to you. My take away so that the group behavior can be completely different [00:48:00] than. Single person behavior. And once we start. Modeling or seeing them in the groups, the crowd behaves in a slightly different way. It doesn't mean it's worse.

[00:48:10] Wojciech Wegrzynski: It doesn't mean it's dangerous. Doesn't mean our models do not work, but it's just different. And this difference can be assessed on the path through, but the modeling. And I think that's, that's a good takeaway. There are a phenomenon that we omit on purpose,, which we may actually try to understand.

[00:48:28] Wojciech Wegrzynski: And model to get our models better. As I've said in the episode.

[00:48:32] Wojciech Wegrzynski: It's maybe not the model's that I'm not sufficient. Maybe our questions are stupid. Maybe we should form our problems related to evacuation better. And I think as the crowd psychology. Field develops as social sciences, improving fire safety. We'll have better and better models. To understand the behavior of people in fires and truly assess the, , their safety and draw the best ways , to help them out in this horrible situation. [00:49:00]

[00:49:00] Wojciech Wegrzynski: Anyway that this is for today. I hope you have enjoyed this episode. And the next one as usual waiting for you next Wednesday. Bye.