ArticlePDF Available

The psychology of crowd behaviour in emergency evacuations: Results from two interview studies and implications for the Fire and Rescue Services

Authors:
Article

The psychology of crowd behaviour in emergency evacuations: Results from two interview studies and implications for the Fire and Rescue Services

Abstract

Existing psychological models of crowd behaviour were applied to examine emergency egress behaviour, and how this could facilitate the safe management of mass evacuations. Two interview–based studies of survivors’ experiences of different emergencies were conducted. It was found that far from mass panic occurring, being in an emergency can create a common identity amongst those affected. A consequence of this is that people are cooperative and altruistic towards others – even when amongst strangers, and/or in life–threatening situations. The analysis has direct implications for how the Fire and Rescue Services manage mass evacuations. In line with earlier critiques, the concept of mass panic is considered to be a myth unsupported by existing evidence. Crowds in emergencies can be trusted to behave in more social ways than previously expected by some involved in emergency planning.
Copyright 2009 by The Psychological Society of Ireland
ISSN 0303-3910
The Irish Journal of Psychology
2009 Vol. 30 No. 1 pp. 59-72
The psychology of crowd behaviour in emergency
evacuations: Results from two interview studies and
implications for the Fire and Rescue Services
Chris Cocking1, John Drury2, & Steve Reicher3
1Department of Psychology, London Metropolitan University
2Department of Psychology, University of Sussex
3School of Psychology, University of St Andrews
Existing psychological models of crowd behaviour were applied to examine
emergency egress behaviour, and how this could facilitate the safe management
of mass evacuations. Two interview-based studies of survivors’ experiences of
different emergencies were conducted. It was found that far from mass panic
occurring, being in an emergency can create a common identity amongst those
affected. A consequence of this is that people are co-operative and altruistic
towards others- even when amongst strangers, and/or in life-threatening
situations. The analysis has direct implications for how the Fire and Rescue
Services manage mass evacuations. In line with earlier critiques, the concept
of mass panic is considered to be a myth unsupported by existing evidence.
Crowds in emergencies can be trusted to behave in more social ways than
previously expected by some involved in emergency planning.
Introduction
The human tragedy apparent in mass emergencies and disasters is depressingly familiar
in society today. Media coverage after 9/11 and more recently the July 7th London
bombings of 2005 was characterised by reports of fear, shock and mass panic. However,
while ‘panic’ is a word frequently used to describe egress behaviour in disasters, a
closer inspection of the behaviour of those affected rarely supports this idea of mass
panic. Indeed, the behaviour of crowds during disasters is often much more social than
that with which they are sometimes credited, with co-operation and altruism often
predominating, rather than selsh uncooperative behaviour. While academics tend
to accept that the concept of mass panic during emergencies is largely a myth (e.g.,
Keating, 1982; Mawson, 2005; Quarantelli, 2001), this view still persists to some extent
in the applied eld and in popular discourse, with the implication being that crowds
in emergencies cannot be trusted to behave in a co-operative way that facilitates their
safe evacuation from danger. This paper will provide evidence supporting the notion
that crowds can behave in ordered and meaningful ways that seem appropriate to those
Address correspondence to: Chris Cocking, Department of Psychology, London Metropolitan University,
Calcutta House, Old Castle St, London, E1 7NY, England. Email: c.cocking@londonmet.ac.uk
60 Cocking, Drury, & Reicher
involved (given the information they have available about the current situation). It will
also offer practical suggestions for the Fire and Rescue Services to help ensure the safe
management of crowds in mass emergencies.
The myth of panic
Despite the frequent use of the word in coverage of disasters, the term ‘panic’ is rarely
examined in any analytical detail to investigate what those who are ‘panicking’ are
actually doing. Mawson (2005, p. 96) denes panic as ‘inappropriate (or excessive) fear
and/or ight’. However, he argues that when looking at behaviour during emergencies,
it is difcult to identify such instances of panic, and descriptions of panic are usually
made by outside observers with the benet of hindsight. Indeed, people in emergencies
often behave in meaningful ways according to the information available to them at the
time. For instance, studies of evacuations in res (e.g., Donald & Canter, 1990; Sime,
1983) show that people tend to leave venues by the way they entered, even if there are
other, closer, exits available. This can be seen as meaningful behaviour, as the smoke
and heat in res can create uncertainty, which may cause people to seek escape through
an exit that they already know exists if they are unfamiliar with other possible routes.
Donald and Canter (1990) also found that fatalities could occur in res because people
were often unwilling to deviate from familiar activity, leading them to continue along
known routes, even in the face of danger. He concluded that the concept of panic or
‘non-adaptive behaviour’ was inadequate to explain human behaviour in res, and that
the danger in such situations often lay in people’s inability to evacuate quickly before the
re spread to an extent where it became impossible to escape.
Recent studies (e.g. Reicher, 2001; Stott & Drury, 2000) have argued that society’s
view of crowd behaviour and management often rely wrongly on early reactionary crowd
theory, such as Le Bon’s (1895/1947) commentary on the crowds of the Paris Commune
during 1870-1. Le Bon emphasised the negative elements of crowds, frequently referring
to them as an irrational mob prone to emotion and suggestibility. If one person began
behaving in an anti-social or irrational way, then this would quickly spread to others
in a process termed contagion. Although the crowds he studied were those involved in
civil disorder, some of his key ideas have been applied to crowds during emergencies
to conclude that mass panic develops in the following ways. Firstly, when faced with
danger, people behave irrationally as they evacuate in a manner disproportionate to the
actual threat (Smelser, 1963). Collective bonds and norms also break down, meaning
that people behave instinctively as they compete with others in an uncoordinated way to
escape danger (Cantril, 1958; Quarantelli, 1954; Strauss, 1944).
This approach (known as the ‘panic model’) has been inuential in the design
of public spaces and engineering of procedures for emergency evacuations (Sime, 1990,
1995). It suggests that plans for the evacuation of buildings should focus on physical
factors such as the width of emergency exits to prevent jamming, rather than psychological
factors such as the role of information and communication. In short, rather than being
viewed as active, thinking agents, crowd members are considered to ow in the same way
as unthinking, inanimate objects such as ball bearings. This has implications for whether
emergency planners decide to provide information to the public during emergencies,
as there is often a concern that people will panic if they know the true gravity of the
Crowd behaviour in emergency evacuations 61
situation facing them. Therefore, information is often withheld, despite there being little,
if any evidence to support this assertion (Drury, 2004).
However, despite the frequent usage in popular discourse of the term ‘panic’ to
describe behaviour during disasters, studies of various different emergencies throughout
the last century have found a general absence of mass panic, despite the threat of death.
This includes: the atomic bombings of Japan in 1945 (Janis, 1951), res in the US
(Feinberg & Johnson, 2001) and UK (Donald & Canter, 1990), and crushes at concerts
(Johnson, 1987). More recently, Blake, Galea, Westeng, and Dixon (2004) studied the
behaviour of evacuees from the World Trade Centre on 11th Sept 2001, and found that
‘classic panic action or people behaving in an irrational manner was noted in [just] 1/124
(0.8%) cases’ (p. 5).
The social attachment model of crowd behaviour
An alternative model of emergency evacuation behaviour has been developed by Mawson
(1978, 2005): the social attachment model. He argues that in times of danger people
display afliative behaviours, where they attempt to move from unfamiliar situations
towards people and/or places that are familiar (such as friends or family). They will
also try to evacuate within this familiar group rather than as individuals. Family and/or
friendship ties remain strong in these situations with mutual cooperation predominating
within these groups, as opposed to selsh, uncooperative behaviour. This approach is
inuenced by Bowlby’s work (1969, 1973) into children’s distress at separation from
attachment gures (usually their mother). The presence or absence of attachment gures
is also believed to inuence distress during disasters. Indeed, Mawson (2005) reports
that during air raids on London in World War II, children often found separation from
parents more distressing than the air-raid itself, suggesting that attachment bonds endure
even in highly stressful situations.
There is support for the social attachment model in a number of studies of
behaviour during emergencies (Cornwell, 2003; Feinberg & Johnson, 2001; Sime,
1983). In each of these cases – evacuations from res in buildings – it was shown that,
rather than breaking down, social bonds within groups endure during emergencies, with
people tending to delay their individual ight to ensure safe evacuation of the group as
a whole. Sime (1983) interviewed survivors of a re in a leisure complex on the Isle of
Man, UK in 1973, and argued that families involved adopted a strategy for group rather
than individual egress, with the quicker family members delaying their own escape to
ensure the safe evacuation of slower individuals. However, Feinberg and Johnson (2001)
argued that this could have tragic consequences for the group as a whole. Their study of
the Beverley Hills Supper Club re in the US in 1977 found that individual fatality risk
increased with group size (although more groups survived than would be predicted by an
individualistic panic model if all social bonds had broken down). Nevertheless, implicit
within these ndings is the notion that while social norms remain largely intact during an
emergency, the larger the group one is in, the chances of individual escape decrease, as
group members wait until all are safe. It is also possible that physical size of the group,
as well as any existing attachment, can delay egress, as larger groups may take longer to
evacuate.
The social attachment model has advantages over the panic model as it explains
62 Cocking, Drury, & Reicher
the evidence for social behaviour and cooperation during emergencies. However a key
limitation of the social attachment model is that while it rightly points out that people
attempt to preserve the safety of existing attachment gures, by focussing on this, it
neglects the possibility of co-operative behaviour between those who had no existing
attachment bonds before the emergency began, or that attachment bonds could develop
quickly between strangers. It also assumes that while being around attachment gures
can diminish fear, if the individual is alone or with strangers, then, ‘even mild threats
can precipitate ight-and-afliation to familiar persons and locations at a distance’
(Mawson, 2005, p. 102), thus implying that ight and selsh ‘panic behaviour’ are more
likely when in a crowd of strangers.
However, coverage of disasters throughout history is replete with examples of
complete strangers behaving co-operatively towards each other, even under conditions of
great personal danger, leading Furedi (2007) to state: ‘even in today’s highly individuated
globalised society, calamities have a unique capacity to encourage acts of solidarity and
altruism’ (p. 2).
Johnson’s (1987) examination of a fatal crush at a concert by ‘the Who’ in the
US in 1979, found that rather than displaying selsh panic behaviour as was reported,
crowd members tried to help others where possible. When trampling occurred, it was
largely due to people not being able to help those who had fallen because of the pressure
of others upon them. It is of course possible that crowd members had an existing sense of
unity that encouraged more co-operation, as they were fans of the same band, but other
studies have found that this unity can develop amongst people whose only common
link appears to be that they are affected by the same disaster. For instance, Tierney
(2002) found that New Yorkers during and after the terrorist attacks of September 11th
2001 were generally altruistic and co-operative towards each other even at great risk to
themselves.
A self-categorisation approach
Psychological research into crowd behaviour has argued that far from losing all rational
thought and behaviour to a ‘mob mentality’, crowd members tend to operate within social
norms, often imposing limits on their behaviour, even during highly stressful situations.
The Elaborated Social Identity Model (ESIM) of crowd behaviour (Reicher, 2001) is
derived from studies of crowd conict (e.g. Drury & Reicher, 1999; Reicher, 1984), and
suggests that a common identity emerges amongst crowd members as a result of a shared
fate in the face of illegitimate attacks from an out-group. This common identity can
result in people helping and supporting others who may have been complete strangers
before the conict started. The ESIM tends to focus on incidents of crowd conict,
such as political demonstrations and disorder at football matches, but has recently been
applied to crowd behaviour during emergencies, such as res, natural disasters and
terrorist attacks. Here the model echoes the suggestion of the sociologist Clarke (2002)
that disasters can create a sense of ‘we-ness’, and hence solidaristic behaviours amongst
survivors. Clarke’s concept of we-ness has not been tested empirically, but the ESIM,
and its parent theory self-categorisation theory (SCT; Turner, 1982, 1985; Turner, Hogg,
Oakes, Reicher, & Wetherell, 1987; Turner, Oakes, Haslam, & McGarty, 1994) provide
the conceptual tools and some suggestive evidence of the psychological process.
Crowd behaviour in emergency evacuations 63
SCT suggests that cognitive representations of the self take the form of self-
categorizations, which may range from personal self-categorizations (what makes us
distinct from others) to shared, collective self-categorizations (what makes us similar to
others). Seeing oneself as personally interchangeable with other in-group members on
some relevant dimensions ‘depersonalization’ means not only seeing the in-group
as homogeneous, but also seeing other in-group members as part of self. This, in turn,
means caring about others and acting in their interests, even where they are not known.
According to SCT, a shared fate is one possible determinant of a shared self-categorization.
The ESIM studies described above showed how such a shared self-categorization can
arise when crowd members perceive themselves to be collectively under threat from an
external group. Extrapolating from this, we might suggest that, during a mass emergency,
shared self-categorisation can arise in a crowd of strangers to the extent that they perceive
the same shared threat. This could help explain the accounts of people helping strangers,
even at risk to themselves. For, if people categorise others as part of the collective self,
then a threat to others is also a threat to the self. This account would suggest that panic
is a feature of individuals, and in most crowds, order will be the norm, with panic only
present in a few unrepresentative individuals.
However, this account has yet to be examined empirically. Therefore, two
interview-based studies of survivors of mass emergencies were carried out to investigate
such incidents in more detail and gather qualitative data about people’s experiences. The
following research questions were considered:
1. Would there be any evidence of mass panic behaviour, and would any
individual distress and/ or fear spread to the crowd as a whole?
2. Furthermore, would orderly, altruistic and co-operative behaviour
predominate as opposed to disorderly, personally selsh and competitive
behaviour?
3. Could any co-operative behaviour be explained by the development of a
common identity in response to the shared threat faced in an emergency?
Method
Two interview studies were conducted with survivors of mass emergencies, where there
was a real or perceived threat of danger and/or death. The data from both studies were
drawn from larger data-sets that have been analysed elsewhere in greater detail (Drury,
Cocking, & Reicher, in submission; in preparation). The rst was an exploratory study
to investigate whether some situations were more likely to result in panic than others.
Interview data was gathered from a diverse range of different crowd emergencies,
with different physical constraints, levels of threat faced by participants, and scale of
casualties suffered.
However, as the interview study progressed, little evidence was found for mass
panic occurring, despite the clear threat of death in some of the emergencies studied.
There were a number of practical difculties in gaining access to survivors, and some of
the emergencies studied took place years ago, with all the potential problems of accuracy
in remembering. Some of the interviews therefore contained gaps.
Halfway through the project, however, the 2005 July 7th London bombings
64 Cocking, Drury, & Reicher
occurred. This presented an opportunity to gather more contemporaneous data, and to
further examine the suggestion that unied crowds do not panic, and that an emergency
itself can unify a crowd to some extent. On July 7th (hereafter 7/7), four bombs were
detonated during the morning rush-hour (three on trains on the London Underground – at
King’s Cross, Edgware Road and Liverpool Street – and the fourth on a bus at Tavistock
Square). The trains and bus targeted were full to capacity, and the vast majority of
travellers would have been commuters with minimal existing afliative ties to their
fellow passengers. The consequent fear of re and/ or death after the explosions may
also have resulted in a highly stressful situation with no obvious escape route. According
to the panic model, this may have resulted in survivors disregarding others’ welfare
to ensure personal survival. Hence if solidarity and cooperation, rather than panic and
personal selshness, were evident in the events studied, then the self-categorization
account would be supported over the afliation and panic models.
Participants
Comparative events interview study
Participants were recruited via adverts in the national media, and through existing
contacts. The events can be divided into ve categories: sinking ships, football stadium
and concert crushes, res, bombs and bomb threats, and a train accident. Twenty-one
survivors’ accounts from eleven different emergencies were gathered and analysed.
July 7th study
A total of 17 participants were recruited in a similar way to the previous study, but
a web-site1 was also set up after 7/7 asking those affected directly to send us their
personal accounts by e-mail (thirteen accounts were received). In addition, background
data was gathered from various on-line accounts from 117 survivors of the bombings
(65 were male, 34 female, and 18 anonymous). We then approached those who had
responded by e-mail to ask if they would be interested in being interviewed as well.
Four further interviewees were recruited via advertisements in the press, approaching
support organizations and ofcial bodies, and through snowballing personal contacts
to see if they knew others who would be willing to share their experiences. Six of the
interviewees were men and six were women. Of the remaining ve e-mail respondents,
three were women. Five of the interviewees and three of the e-mail-only respondents
were directly caught up in the blasts; the rest were eye-witnesses.
Interview procedure
We arranged for the interviews to take place where the participants would feel comfortable,
usually their home or a local park. The interviewee was rst asked to provide some
background, to set the scene, and then to tell the story of the events as they remembered
them. The rest of the questions covered the following issues:
Behaviour: e.g., ‘What did you do in response to these events? How quickly did
people respond and evacuate? Did people co-operate/ help each other out? Did anyone
behave selshly?’
Thoughts/ feelings: e.g., ‘What were you thinking/ feeling as incident progressed?
Did you feel in control of your actions/ feelings? Do you think that anyone panicked? If
Crowd behaviour in emergency evacuations 65
so, what did they do?’
Identities: e.g., ‘How would you describe those in the evacuation with you?
How did you feel towards them? Did you feel a sense of unity with each other? Was
there a common identity before-hand?’
Each interview lasted between 45 and 90 minutes, and the data was transcribed
in full and subjected to thematic analysis, with the themes determined by our research
questions.
Results
As much of the material gathered contained data that overlapped between the different
research questions, they are reported in two different sections, as opposed to addressing
each research question in turn, which is done in the Discussion section.
The myth of panic: Was there personally selsh behaviour?
Comparative events interview study
Eleven interviewees described the crowd’s behaviour as ‘panic’, while eight did not.
However, this does not necessarily support the panic model, as these accounts needs
more probing. What did people means by ‘panic’, and what did people actually do? For
instance, most interviewees (14 vs. 6) explicitly contrasted the over-emotional, panicked
behaviour of some individuals with the relatively orderly behaviour of the rest of the
crowd. Furthermore, when directly asked if they thought there was mass panic, most
interviewees (18 vs. 1) also explicitly denied that crowd panic took place.
The interviewee quoted below, a survivor of the Hillsborough football disaster
in 1989 where 96 Liverpool fans died in a stadium crush, was explicit in his use of the
term panic:
1. You had no choice, you went where the crowd took you ... as everybody else
did ... it was that scary, it was terrifying, but as I said once blind panic has set in I would
think that was that was the main part, every everyone really panicked, sheer panic, the
police panicked, the crowd panicked, everyone panicked (Hillsborough 1).
While he mentions panic six times in this extract, he does not explain in any detail what
it is people were doing that led to him describing it as panic. When asked to give an
example of what he felt was panicked behaviour, he gave the following account:
2. Never ever would I consider stepping on a dead body. I did that to save myself
because I panicked ... 2 I thought to myself look I’m either going to step on he or she ...
to get out of this and live or die (Hillsborough 1).
However, while this is an exceptional situation, a closer analysis begs the question as to
whether this is actually personally selsh panic or logical ight behaviour. He was faced
with a stark choice of stepping on someone who was dead (and so beyond help) to escape
the crush in the pens, or be killed himself. He also later described his own co-operative
behaviour to help other injured fans once he was free from the crush:
3. As soon as I could get my arms out I was helping people and pushing them
up ... it’s only when you look back you just feel ‘oh I could have done that’, I mean you
look back, I mean everyone did help each other and I don’t think there was anyone that
could really look back and say I didn’t do anything to help anybody (Hillsborough 1).
If the panic model was correct, we would expect little in the way of such co-
66 Cocking, Drury, & Reicher
operative behaviours. However, consistent with the suggestion that mass panic is rare,
accounts of helping (including physically helping people, allowing others to go rst,
and comforting others) were more common than personally selsh behaviours (barging
others aside, ignoring others in need, trying to push ahead of others etc.). Indeed, more
interviewees reported behaving helpfully (12) than said they did not (6), more interviewees
reported being helped (13) than not (0); and more reported observing others helping
others (20) than did not (1). Likewise, fewer interviewees reported behaving personally
selshly or competitively (3) than said they did not (14), fewer interviewees reported
suffering from others’ personal selshness (6) than said they did not (14). However,
more interviewees reported observing others being personally selsh (11) than did not
(5).
July 7th study
Afliation theory would suggest that mass panic would be more likely under the
conditions present on 7/7, as those affected would be in a situation of extreme danger
largely with strangers. However, of the 7/7 survivors interviewed, only one said they
‘panicked’, while four others said they ‘felt’ panicky, but this was usually feelings of
fear rather than any overt behaviour. Two respondents were explicit that they did not
panic. When describing others’ behaviour, only one respondent used the term ‘panic’,
and when asked what she meant, she replied that people were screaming. However,
she also described others’ behaviour as overwhelmingly ‘calm’. The other respondents
either denied that people panicked (ve respondents), said that people ‘started to panic’
but were ultimately calm or controlled (two), said that they did not see any panic (two),
or limited it to one individual or a small minority in the crowd ‘hyperventilating’,
‘screaming’ or becoming ‘hysterical’ (four). The following quote from a witness of the
Edgware Road bomb indicates the lack of panic:
4. It was so calm and relaxed it was almost like a re drill ... everyone was sat
down and the driver was saying ‘you might as well sit down as there’ll be a bit of a wait
till we get out’ and ... some people trying to get a bit further along the line but there was
no-one desperately running along the train, it was a very relaxed calm evacuation, and
I think the atmosphere and the instructions from the driver because he was very calm
about the whole situation obviously ... he’d seen the blast from his carriage but he was
calm and I think his calm instilled calm throughout the whole train yeah there wasn’t it
wasn’t a panic really bolt for the door by any means (July 7th 1).
Evidence for shared identity and cooperative behaviour
The third research question was whether there was a sense of common identity which
would at least partly explain the lack of panic and extent of helping behaviour. This was
tested by examining the data for references to common identity, and for a link between
common identity and helping.
Comparative events interview study
Participants placed great importance on unity with others involved, often spontaneously
mentioning this before we asked them about the concept. Therefore, we can have some
condence that the data reect a genuine sense of common identity that existed in the
Crowd behaviour in emergency evacuations 67
crowd. Thirteen participants referred unambiguously to a sense of unity or togetherness
with the rest of the crowd during the emergency. Their comments usually included
examples of other people’s motivations and behaviour, suggesting that the unity was not
something that existed only subjectively for them, but was felt by the crowd as a whole.
In most of the references to common identity, it is also described as developing over
the course of the emergency itself. Only seven participants described any sense of unity
before the incident, and these were all at football matches or concerts. Conversely, most
of the people who described a shared sense of threat also referred to a sense of common
identity developing over the course of the event, sometimes explicitly explaining the
feeling of unity in terms of the threat to the crowd as a whole.
The following extract is from someone evacuated from a hotel re in Boston,
US in 1971, and who had only arrived the day before from the UK, not knowing anyone
else in the hotel. Therefore, his chances of having a prior common identity with other
guests were unlikely. What is of interest, however, is the emergence of a sense of unity
with others in the course of the emergency, as all faced the same danger:
5. We were herded into groups of about ten or fteen people or so in the hall
talking and milling around amongst themselves at that point yes there was a little bit of
camaraderie that we’d all come through something that could have been potentially very
dangerous (Boston Hotel Fire).
This common identity was not just a perception amongst participants, as it also
appeared to inuence their behaviour to others, as the following quote from another
survivor of Hillsborough illustrates:
6. The behaviour of many people in that crowd and simply trying to help their
fellow supporters was heroic in some cases. So I don’t think in my view there was any
question that there was an organic sense of unity of crowd behaviour. It was clearly
the case ... that people were trying to get people who were seriously injured out of that
crowd, it was seriously a case of trying to get people to hospital, get them to safety ...w
I just wish I’d been able to prevail on a few more people not to put themselves in danger
(Hillsborough 2).
July 7th study
Most respondents from 7/7 were amongst strangers (only one interviewee and three
e-mail respondents were with friends or family) and so would have had minimal existing
afliative ties to their fellow commuters. However, as will be shown, there is evidence
of a sense of unity amongst survivors. Moreover, when asked to describe how this
unity developed, there was evidence that it emerged from a sense of shared fate. Nine
interviewees (plus two of the e-mail respondents) were explicit that there was a strong
sense of unity in the crowd. Indeed, as with the comparative event interview study,
some of them mentioned this before being asked about it by the interviewer, with the
following terms being used: ‘empathy’, ‘unity’, ‘similarity’, ‘part of a group’. When
asked to rate the strength of this feeling of togetherness, participants used scores such
as: 8/10, 9/10, 100%, 10/10, suggesting that this was a strong subjective feeling. Some
explicitly contrasted this positive feeling of unity in the emergency with the unpleasant
sense of competition and atomization with other commuters they would experience on a
normal rush-hour morning. Only one respondent reported a low sense of unity with others
68 Cocking, Drury, & Reicher
(scored as 3/10), and this was someone who was not directly affected by the explosions,
but was in the area of the bus bomb and heard the explosion from a distance. Therefore,
it is possible that he did not feel the same immediate threat of death that others felt, thus
reducing the sense of shared fate which would predict such a sense of unity.
The following quote from a female survivor of the King’s Cross bomb is taken
from a web-site where survivors offer mutual support3, and illustrates the sense of unity
that developed amongst those caught up in the blast:
7. One of the things which struck me about this experience is that one minute
you are standing around strangers and the next minute they become the closest and most
important people in your life. That feeling was quite extraordinary (July 7th 2).
There was also clear evidence of cooperation, with thirteen interviewees reporting
at least one instance of helping others – ranging from providing comfort, water, or rst aid.
While there were no reports of overtly personally selsh behaviour, seven interviewees
said they felt ‘selsh’ or guilty for being overly concerned for their own personal safety.
However, this may be evidence for survivor guilt, where people sometimes struggle to
come to terms with why they survived emergencies and others did not, or even feel they
were to blame for some aspect of the trauma they later suffer. Seven interviewees were
explicit that they witnessed no personally selsh or competitive behaviour in others.
However, two interviewees described one individual being concerned with his mobile
phone when he could have helped (although they did also report that others remonstrated
with the person for doing so), and two described people outside the events who showed
no concern for the plight of those affected. Only one described people ‘ignoring others,
walking past’. The following quote from a survivor of the Edgware Road bomb describes
the general co-operative atmosphere:
8. Some people took charge of the situation by looking for stuff and then other
people were just looking after people next to them and other people were just keeping
out of the way (July 7th, 1).
Discussion
The two interview-based studies generated a rich qualitative data-set which provided
evidence for the remarkable resilience that people are capable of under extreme pressure.
Moreover, this resilience appears to be not despite the crowd, but because of it, as it was
a product of the sense of collective identication that they derived from others involved
in the same incident. Each research question will now be addressed in turn.
Lack of mass panic
The evidence gathered supports the theory that there would be little evidence of mass
panic, and while the term ‘panic’ was often used by participants, it did not stand up to
scrutiny. Behaviour described as panic usually consisted of vocal expressions of fear or
distress amongst individuals rather than any physical behaviour, and this did not spread
to others. Furthermore, mass panic did not occur even under the conditions of the July
7th London bombings, where the afliation model too would have predicted it.
Orderly, cooperative behaviour would pre-dominate
People’s behaviour during the events in question tended to remain orderly and meaningful,
Crowd behaviour in emergency evacuations 69
with selsh, uncooperative behaviour being rare. Moreover, any behaviour in others that
was perceived by participants as personally selsh did not spread, and sometimes other
crowd members remonstrated with the ‘selsh’ individuals in question.
The development of a common identity in response to the threat would explain this
cooperative behaviour
Participants were often explicit that they felt a common unity with others affected
as a result of having a shared fate in response to the danger they faced, and that this
inuenced their cooperative behaviour. Furthermore, participants often spontaneously
mentioned this sense of unity before being asked by the interviewer, lending strength to
this hypothesis.
Conclusion and implications for practice
The widespread cooperative behaviour reported by participants in emergencies
contradicts the assumptions of personally selsh behaviour inherent in the panic model.
Furthermore, the evidence we found for the development of a common shared identity
amongst participants who had minimal attachment bonds to others before the emergency
began highlights the limitations of the attachment model in explaining all behaviour
in emergencies. Therefore we can concur with previous critiques (e.g., Keating, 1982;
Sime, 1990) that the idea of mass panic occurring in emergencies is largely a myth
unsupported by evidence, and that the term is neither a helpful nor accurate description
of human behaviour in emergencies.
This research has implications for the safe egress of large numbers of people in
emergency evacuations (see Drury & Cocking, 2007), and the following recommendations
are proposed to facilitate safely such evacuations:
1. Those involved in crowd management need to take the possibility of
emergencies seriously and be prepared for the worst case scenario, regardless of how
likely they believe it is to happen, as it may not be over-reaction, but rather under-
reaction that tends to lead to deaths in emergencies. There is evidence that people can
delay their own evacuation from danger (e.g., Sime, 1995), and that they can also assess
risk inaccurately, such as being in denial that negative events can occur or that they are
more likely to happen to others than themselves(e.g., Plous, 1993). Therefore, planners
need to be aware of these processes and pre-empt them, by considering how they would
respond to every possible type of emergency (especially the most extreme events).
2. Planners need to be aware that mass emergency evacuation behaviour is often
cognitive and meaningful. People’s behaviour in emergency evacuations is affected by
their knowledge of the nature of the emergency, the physical layout of the area, and that
they will seek further information and guidance during the incident. Therefore, more
rather than less information should be provided wherever practically possible during
emergencies. People in emergencies can digest and act upon information much more
effectively than they are often given credit for. Assuming that crowds will panic may
indeed delay efcient evacuations, if emergency planners withhold information from
crowd members in the mistaken belief that panic will occur if people become aware of
the threat they face. Indeed, a study of different methods of evacuation from the Metro
system in Newcastle, UK found that providing clear information to the public from
70 Cocking, Drury, & Reicher
a believable source about a threat, far from hampering efcient evacuation, actually
improved evacuation times (Proulx & Sime, 1991).
3. Forms of communication that can nurture a sense of collective identity should
be encouraged, and appealing to people’s cooperative nature before and during mass
emergencies can be an effective tactic to ensure a safe and efcient evacuation. A
shared identity can be encouraged in public spaces (such as underground stations) on
an everyday basis. This might be achieved via public information campaigns and the
wording of public addresses, advertisements, notices and so on. There is some evidence
(e.g., Muir, 2004; Ripley, 2005) that those who access emergency information in trains
and planes are more likely to survive such crashes.
Finally, and from a more general perspective, emergency planners should be aware
that crowds in evacuations can usually be trusted to behave well during evacuations. We
believe that our ndings are part of a growing body of evidence that crowd behaviour is
meaningful and cognitive, and using this knowledge could help reduce the risk of severe
injuries and/or fatalities happening during future mass emergencies.
Acknowledgements
This research was made possible by a grant from the UK Economic and Social Research
Council (ESRC), grant no RES-000-23-0446 To J. Drury (University of Sussex, UK),
S. Reicher (University of Saint Andrews, Scotland), D. Schoeld (RMIT University,
Australia) and, P. Langston (University of Nottingham, UK)
Notes
1. Available via; http://www.sussex.ac.uk/afliates/panic/lb/index.htm.
2. In this and future extracts, the symbols ‘...’ denote that material was edited for reasons
of space .
3. http://www.londonrecovers.com.
References
Blake, S.J., Galea, E.R., Westeng, H., & Dixon, A.J.P. (2004). An analysis of human
behaviour during the World Trade Center disaster of 11 September 2001 based
on published survivor accounts. Proceedings of Third International Symposium
on Human Behaviour in Fire, Belfast, September.
Bowlby, J. (1969). Attachment and loss, Vol. 1: Attachment. London: Hogarth Press.
Bowlby, J. (1973). Separation anxiety and anger: Attachment and loss. London: Hogarth
Press.
Cantril, H. (1958). The invasion from Mars. In E.E. Maccoby, T.M. Newcomb, & E.L.
Hartley (Eds.), Readings in social psychology (3rd ed.) (pp. 291-300). New York:
Henry Holt & Co.
Clarke, L. (2002). Panic: Myth or reality? Contexts, 1, 21-26.
Cornwell, B. (2003). Bonded fatalities: Relational and ecological dimensions of a re
evacuation. The Sociological Quarterly, 44, 617-638.
Donald, I., & Canter, D. (1990). Behavioural aspects of the King’s Cross Disaster. In D.
Canter (Ed.) Fires and human behaviour (pp. 15-30). London: David Fulton.
Drury, J. (2004). No need to panic. The Psychologist, 3, 118-119.
Crowd behaviour in emergency evacuations 71
Drury, J., & Cocking, C. (2007). The mass psychology of disasters and emergency
evacuations: A research report and implications for practice. Unpublished
manuscript. University of Sussex, Brighton. Available at; http://www.sussex.ac.uk/
afliates/panic/applications.html
Drury, J., Cocking, C., & Reicher, S. (in preparation). The nature of collective resilience:
Survivor reactions to the 2005 London bombings.
Drury, J., Cocking, C., & Reicher, S. (in submission). Every one for themselves? An
interview study of crowd solidarity among emergency survivors.
Drury, J., & Reicher, S. (1999). The intergroup dynamics of collective empowerment:
Substantiating the social identity model. Group Processes and Intergroup Relations,
2, 381-402.
Feinberg, W.E., & Johnson, N.R. (2001). The ties that bind: A macro-level approach
to panic. International Journal of Mass Emergencies and Disasters, 19, (3) 269-
295.
Furedi, F. (2007). Invitation to Terror: The expanding empire of the unknown. Continuum
Press: London, UK.
Janis, I.L. (1951). Air war and emotional stress: Psychological studies of bombing and
civilian defense. New York: McGraw-Hill.
Johnson, N.R. (1987). Panic and the breakdown of social order: Popular myth, social
theory, empirical evidence. Sociological Focus, 20, 171-183.
Keating, J.P. (1982). The myth of panic. Fire Journal, 147, 56-61.
Le Bon, G. (1947). The crowd: A study of the popular mind. Dunwoody, GA: Norman
S. Berg. (Originally published 1895.)
Mawson, A.R. (1978). Panic behaviour: A review and new hypothesis. Paper presented
at the 9th World Congress of Sociology. Uppsala, Sweden. August.
Mawson, A.R. (2005). Understanding mass panic and other collective responses to threat
and disaster. Psychiatry, 68, (2) 95-113.
National Institute for Clinical Excellence (2006). NICE Guidelines on PTSD. Online
leaflet provided for survivors of trauma. Accessed on 13/7/2007 via http://
www.7julyassistance.org.uk/downloads/Affected%20by%20the%20London%20
bombings.pdf
Muir, H. (2004). In times of crisis, how do passengers react? 15th Westminster lecture on
transport safety, Parliamentary Advisory Council for Transport Safety, London.
Plous, S. (1993). The psychology of judgment and decision making. New York: McGraw-
Hill.
Proulx, G. & Sime, J.D. (1991). To prevent ‘panic’ in an underground emergency: Why
not tell people the truth? In G. Cox & B. Langford (Eds.), Fire Safety Science:
Proceedings of the Third International Symposium (pp. 843-852). London: Elsevier
Applied Science.
Quarantelli, E.L. (1954). The nature and conditions of panic. American Journal of
Sociology, 60, 267-275.
Quarantelli, E.L. (2001). Panic, sociology of. In N.J. Smelser & P.B. Baltes (Eds.),
International encyclopedia of the social and behavioural sciences (pp. 11020-
11023). New York: Pergamon Press.
Reicher, S.D. (1984). The St Pauls’ riot: An explanation of the limits of crowd action
72 Cocking, Drury, & Reicher
in terms of a social identity model. European Journal of Social Psychology, 14,
1-21.
Reicher, S. (2001). The psychology of crowd dynamics. In M.A. Hogg & R.S. Tindale
(Eds.), Blackwell handbook of social psychology: Group processes (pp. 182-208).
Oxford: Blackwell.
Ripley, A. (2005). How to get out alive. From hurricanes to 9/11: What the science of
evacuation reveals about how humans behave in the worst of times. Time. 2/5/2005,
58-62.
Sime, J.D. (1983). Afliative behaviour during escape to building exits. Journal of
Environmental Psychology, 3, 21-41.
Sime, J.D. (1990). The concept of ‘panic’ In I. Donald, & D. Canter (Ed.), Fires and
human behaviour (2nd ed.) (pp. 63-81). London: David Fulton.
Sime, J.D. (1995). Crowd psychology and engineering. Safety Science, 21, 1-14
Smelser, N.J. (1963). Theory of collective behaviour. London: Routledge & Kegan
Paul.
Stott, C.J., & Drury, J. (2000). Crowds, context and identity: Dynamic categorisation
processes in the ‘poll-tax riot’. Human Relations, 53, 247-73
Strauss, A.L. (1944). The literature on panic. Journal of Abnormal and Social Psychology,
39, 317-328.
Tierney, K.J. (2002) Strength of a city: A disaster research perspective on the World
Trade Center attack. Available at http://www.ssrc.org/sept11.
Turner, J.C. (1982). Towards a cognitive redenition of the social group. In H. Tajfel (Ed.),
Social identity and intergroup relations (pp. 15-40). Cambridge, UK: Cambridge
University Press.
Turner, J.C. (1985). Social categorization and the self-concept: A social cognitive theory
of group behaviour. In E.J. Lawler (Ed.), Advances in group processes: Theory
and research (Vol. 2) (pp. 77-122). Greenwich, CT: JAI Press.
Turner, J.C., Hogg, M.A., Oakes, P.J., Reicher, S.D., & Wetherell, M.S. (1987).
Rediscovering the social group: A self-categorization theory. Oxford: Blackwell.
Turner, J. C., Oakes, P. J., Haslam, S. J., & McGarty, C. (1994). Self and collective:
Cognition and social context. Personality and Social Psychology Bulletin, 20,
454-63.
... In the process of pedestrian evacuation, tension not only affects individuals but also spreads through emotional infection among individuals. Cocking et al. [24] pointed out through empirical analysis that emotions can spread rapidly among people during an emergency evacuation, even among strangers. Fu et al. [25] suggested that there was contact infection in pedestrian emotion. ...
... Cocking et al. [24] believe that the emotional value of evacuation individuals is related to their self emotions and the emotions of others around them. In this paper, the dynamic equation of the tension emotional intensity of pedestrian j at time t is defined by ...
Article
Full-text available
Traditional dynamic models cannot fully describe the microdetails of the impact of emotional contagion on individual state and behavior when pedestrian evacuation is simulated. This paper addresses the problem by constructing a dynamic evolution mechanism among emotion, state, and behavior. First, the direction perception domain of pedestrians during evacuation is defined. Then, the dynamic emotional perception of pedestrians during an evacuation is studied, considering the emotional increment caused by personal walking speed and others in the direction perception domain. Next, emotional contagion is introduced into the improved cellular automata (CA) simulation model in the floor field (FF), entitled the “CECA model.” The transfer probability of pedestrians in different states is proposed by defining “susceptible emotional state” and “infectious emotional state.” Finally, the simulation results are compared with known models. The results demonstrate that the improved model can improve the evacuation efficiency of the system significantly. Simultaneously, the effects of emotional threshold, infection coefficient, calm coefficient, and perception radius of the evacuation system on the pedestrian evacuation process are simulated and analyzed, providing a basis for evacuation managers to formulate evacuation strategies.
... These emotions are contagious and rapidly spread through a crowd. Therefore, panic emotions, as a natural uncontrolled process, can spread from one individual to all air passengers [18,17]. ...
Article
Full-text available
The aim of this paper is to study a discrete mathematical model of panic spreading on an airplane inspired by SIRS model, and the optimal control strategies, applied to reduce the number of panicked passengers during the flight. The population is divided into three compartments: Panic-prone panicked, and recovered passengers. Two control strategies were used, which are psychological therapies, applied by trained flight personnel, to both susceptible and panicked passengers. The discrete-time Pontryagin maximum principle was used to characterize the optimal controls. All numerical simulations were performed using MATLAB. The obtained results demonstrated which of the strategies is effective.
... Retracing the widespread phenomenon of evacuation reveals that people may be under great stress in the face of emergencies, which leads to a general loss of self-control and decision making that produces non-social and irrational behaviour [43,44]. However, while people do respond emotionally and make different behavioural decisions in response to stress [45], evacuees behave more socially than in mass panic [46,47] and tend to make rational decisions and display pro-social behavioural decisions rather than lose self-control [48]. During the evacuation, depending on their social identity, people make a range of behaviours that conform to social norms [46,49] and decisions that are consistent with group behavioural tendencies [50]. ...
Article
Full-text available
In the research, decision-making capabilities are explored in relation to the prediction of evacuation efficiency to improve forecast accuracy on metro platforms. For this purpose, this study reviewed theories related to evacuation behaviours utilising the anomaly-seeking approach and the paradigm of relationship development. The conceptual framework of decision-making capability and evacuation behaviours was explored based on risk perception, level of emergency knowledge, survivability and emotion, and their relationship with the partial least squares equation was constructed. A predictive model of evacuation efficiency and its differential equations incorporating this relationship were also proposed based on the epidemic model. By developing and testing the conceptual framework and model, theoretical support is provided for evacuation behaviour, while assisting emergency management in developing plans and measures to respond to emergencies on metro platforms. This study realises the possibility of predicting evacuation efficiency from a decision-making capability perspective.
... Different studies show that high motivation, which often involves pushing and shoving, increases density [2,3], and reports from real-life scenarios indicate that pushing from behind can lead to life-threatening density and pressure resulting in injuries and fatalities [4]. Although there is broad evidence of cooperative behavior in emergencies [4][5][6][7], pushing may also occur during evacuations, which further increases the danger. [8] Several simulations of pedestrian crowds have therefore tried to integrate this behavior [9,10] but without providing a systematic psychological basis. ...
Article
Full-text available
Pushing behavior impairs people’s sense of well-being in a crowd and represents a significant safety risk. There are nevertheless still a lot of unanswered questions about who behaves how in a crowded situation, and when, where, and why pushing behavior occurs. Beginning from the supposition that a crowd is not thoroughly homogenous and that behavior can change over time, we developed a method to observe and rate forward motion. Based on the guidelines of quantitative content analysis, we came up with four categories: (1) falling behind, (2) just walking, (3) mild pushing, and (4) strong pushing. These categories allow for the classification of the behavior of any person at any time in a video, and thereby the method allows for a comprehensive systematization of individuals’ actions alongside temporal crowd dynamics. The application of this method involves videos of moving crowds including trajectories. The initial results show a very good inter-coder reliability between two trained raters demonstrating the general suitability of the system to describe forward motion in crowds systematically and quantify it for further analysis. In this way, pushing behavior can be better understood and, prospectively, risks better identified. This article offers a comprehensive presentation of this method of observation.
... For instance, even though the virtual occupants had the same appearance in the office and school, office workers had more positive perceptions of occupants in the office, whereas teachers had more positive perceptions of occupants in the school. This finding had to do with participants' psychological ties with others, as they tended to have closer relationships with familiar people or those sharing common identities 39,74 . Additionally, office workers considered hiding places more in the office, while teachers considered hiding places more in the school. ...
Article
Full-text available
Active shooter incidents represent an increasing threat to American society, especially in commercial and educational buildings. In recent years, a wide variety of security countermeasures have been recommended by public and governmental agencies. Many of these countermeasures are aimed to increase building security, yet their impact on human behavior when an active shooter incident occurs remains underexplored. To fill this research gap, we conducted virtual experiments to evaluate the impact of countermeasures on human behavior during active shooter incidents. A total of 162 office workers and middle/high school teachers were recruited to respond to an active shooter incident in virtual office and school buildings with or without the implementation of multiple countermeasures. The experiment results showed countermeasures significantly influenced participants’ response time and decisions (e.g., run, hide, fight). Participants’ responses and perceptions of the active shooter incident were also contingent on their daily roles, as well as building and social contexts. Teachers had more concerns for occupants’ safety than office workers. Moreover, teachers had more positive perceptions of occupants in the school, whereas office workers had more positive perceptions of occupants in the office.
... Second, shared experience may lead to a sense of shared identity (Bradshaw & Muldoon, 2019;Kearns, Muldoon, Msetfi, & Surgenor, 2017). This may be something as benign as the experience of commuters on a train that breaks down (Neville, 2011), or as serious as the experience of commuters on a train that has been bombed (Cocking, Drury, & Reicher, 2007). Third, historians, sociologists and (latterly) psychologists have suggested that the various forms of physical coordination which occur at collective eventsmarching in step, chanting slogansmay help consolidate and help people to imagine themselves as sharing a common identity (e.g. ...
Article
The increasing growth of high-rise residential buildings in recent years and the existence of hazards such as fires, earthquakes, and terrorist incidents have increased the importance of evacuating buildings to save lives. Meanwhile, statistics show that a significant percentage of the population and hence residents of buildings are permanently disabled. However, the study of the evacuation of a mixed-ability population is more complicated than a homogeneous population. This study examined phased evacuation, emplacement strategy, and ramp installation using a mixed-ability population simulation in a 26-story residential building. The study population includes 2% of people with mobility limitations and 98% without disabilities. Unlike previous studies, the results revealed that phased evacuation could not reduce evacuation time in a mixed-ability population. It was also found that emplacement strategies could shorten the evacuation time by up to 10%, though suffering from an ethical problem. Meanwhile, combining ramp installation with others strategies offered the best and broadest impact on a safe evacuation and could improve evacuation time by up to 30% and reduce congestion on exit routes.
Article
Full-text available
When a disaster happens, and emergency caution sounds, it is anticipated that people will stay levelheaded, composed, and collected. Being calm amid a fatal event is exceptionally valued and definitely enhances your survival probabilities. But detailed reports and inadequate research accessible propose that persons under danger can behave in a manner that put them and others in more risk. This paper explores a few of the unconstructive behaviors which can occur amid predicaments and interventions to mitigate them.
Article
The importance of the human factor during catastrophic events is increasingly being highlighted in the field of disaster management as human behaviors critically affect the propagation of disasters and the extent of damage they cause. Despite the attempted efforts, the complexities and the inherent difficulties in modeling human behaviors still prevent from an effective tool for predicting them during disasters. This paper contributes to this course of research by proposing a methodology to construct a predictive indicator, namely the Behavioral Pattern during Disaster Indicator (BPDI). First, a systematic literature review on the identification and the relations between behavioral factors, behavioral patterns and humans’ reactions during catastrophic events is performed. The review presents also the main typologies for modeling the human factor in the context of disaster management. Then based on the review findings, the paper proposes a statistical methodology that comprises correlation and ordinal regression analysis for constructing a predictive indicator. The methodology is applied to a case study that was created in the context of this research. A questionnaire survey that yielded 573 responses is used for identifying the effect of 18 factors on the adoption of one or more among three behavioral patterns. Several conclusions are drawn through the methodology’s application with the most significant one being the potential of a lack of significant correlations between behavioral factors and patterns during catastrophic events. The final output of this research, i.e., the BPDI indicator can be an effective tool that is easy for disaster management authorities to apply towards predicting human behaviors during disastrous events.
Article
The research of crowd evacuation plays the predominant role of guaranteeing the controllability of the stampede disaster and decreasing accidents casualties. Herd behavior is one of the characteristic crowd evacuation behaviors, which has the definite influence on evacuees’ decision-making and evacuation efficiency. In this paper, the crowd evacuation observation experiments were conducted to investigate herd behavior by considered 16 working conditions, including change the number of participants, alter urgency levels and set barrier. The participant allotment and crowd movement time on different sides and exits are obtained. The judgment area and crowd density are proposed to understand the herd behavior and choices of decision maker. The critical states of herd behavior in direction and exit choices are confirmed according to the video observation and data analysis. The results indicate that herd behavior moderately leads to the different distribution and movement time of the experimental crowd. Decision maker does not present herd behavior when crowd density is relatively high or low. When crowd density in direction choice judgment area is within the range of [1/12, 1/6] (person/m²), and the crowd density in exit choice judgment area is within the range of [1/27, 2/27] (person/m²), decision maker basically triggers herd behavior. This study provides insight into the critical state of herd behavior which deepens the understanding of crowd evacuation and contributes to the reference of the crowd accident management.
Chapter
This chapter represents an unfashionable engagement with the work of the collective behaviour theorist Neil Smelser (1962, 1969). Smelser studied under and worked with Parsons (see Parsons & Smelser, 1998), and his sociological work was heavily influenced by his functionalist framework, as he acknowledges. Although he did not hold that society was often in a completely harmonious and stable state, he retained the notion of such equilibrium as the baseline from which the emergence of collective behaviour needed to be explained. His theory therefore hinges on the idea that social movements are the results of strains in the social structure, defined as ‘the impairment of the relations among, and consequent inadequate function of, the components of social action’ (Smelser, 1962). According to Smelser, ‘people join radical movements because they experience social dislocation in the form of social strain, especially when such strain springs from rapid social change’ (Smelser cited in Weeber & Rodeheaver, 2003). The notion of strain is arguably the most critical component of his model of the determinants of collective behaviour (Weeber & Rodeheaver, 2003), although the existence of strains alone was not considered enough to explain why collective behaviour occurred at the times and in the forms that it did. Importantly for this book, Smelser assumes ‘that perceived structural strain at the social level excites feelings of anxiety, fantasy, hostility, etc’ (1962, p. 11).