By Daniel Nilsson
Following large fatal fire incidents, ‘panic’ is often mentioned in the media. Examples include newspaper articles from the early 1900s, such as the one published after the Iroquois Theatre fire in Chicago:
‘PANIC BALKS ESCAPE; Maddened Audience Unable to Reach the Exits’
(New York Times, December 31, 1903)
Although researchers have repeatedly criticized the use of the term panic (Sime, 1980; Sime, 1984; Fahy, Proulx & Aiman, 2009), it is still being used. An example is the online article after the Kiss Nightclub fire in Brazil:
“Brazil nightclub fire survivors speak of panic”
(BBC News, January 28, 2013)
Panic is also mentioned in many handbooks from different countries, including the USA (Croker, 1917), the UK (Phillips, 1951) and the Soviet Union (Roytman, 1975). The assumed consequence of panic is often described as an uncontrollable rush for exits that leads to people being trampled to death (Phillips, 1951; Roytman, 1975). Also, it is sometimes claimed that information about the fire — people hearing the word ‘fire’ (Croker, 1917; Roytman, 1975) or overhearing a call to the fire brigade (Home Office, 1934) — can cause panic.
These examples suggest that there may not need to be any real danger for panic to occur and that people only have to believe that the situation is dangerous for panic to ensue. It is also suggested that panic can spread like a disease from person to person in a crowd (Phillips, 1951; Roytman, 1975). However, statements like these are typically provided without supporting evidence. Instead, data suggest that many of these ‘facts’ are simply myths. For example, research has shown that people recall a voice alarm more accurately and do not display significantly different behaviour if fire is mentioned (Nilsson & Frantzich, 2008).
The main cause of the belief that panic occurs is that the behaviour of others is viewed without a complete understanding of their situation. In many cases, panic is attributed to a specific individual based on observed behaviour, but the individual typically has a perfectly sensible explanation for that behaviour. Fahy, Proulx and Aiman (2009) present one useful example, basing their description on behavioural data from fatal residential fires (Brennan, 1998). Fahy, et. al. (2009) describe how a group of elderly people trapped by fire and smoke on the third floor of a building were described by an observer as ‘panicking’. At one point, an elderly man even threw a chair out of a window, which might seem quite illogical and ‘panicky’. However, he threw out the chair to test if it would break, which it did. He then concluded that he would be seriously injured if he attempted to jump from the window. From an external observer’s perspective, it might look like the man was acting in panic. However, throwing out the chair was hence a clever scientific experiment, which suggest rational decision-making in spite of a stressful situation.
In media accounts as well as in handbooks, panic is often used to explain why a fire results in many fatalities. However, it is not always this straightforward; many factors dictate the outcome of a fire: fire severity, overcrowding, availability of escape routes, etc. Panic can therefore not be used in retrospect as an explanation for behaviour, but instead requires comprehensive investigation of what actually happened. Investigation, in turn, requires a proper definition of panic.
Many proposed definitions attempt to link panic to irrational behaviour. As pointed out by Sime (1980, 1984), the use of ‘irrational’ is unfortunate, since it suggests people perceive all relevant information during evacuation and act contradictory to this information. However, it has been argued that this type of behaviour is not supported by available data (Fahy, Proulx & Aiman, 2009).
One definition of panic often used in research is by Quarantelli (1954), who used a study of a large incident database. In going through that database, Quarantelli was only able to identify a small number of cases where people behaved in a way that could be called panic. Based on an in-depth investigation of people’s behaviour in these specific cases, he defined panic as:
… an acute fear reaction marked by loss of self-control, which is followed by non-social and non-rational flight behaviour …
According to Quarantelli, panic is a short-lived phenomenon, involving seconds rather than minutes, and includes loss of self-control. He also claims that panic is non-social in contrast to anti-social. This means that people do not consider social bonds, e.g., family ties, but also that they do not display anti-social behaviour, e.g., deliberately hurt others to get ahead.
One example from Quarantelli’s large database is the behaviour of a woman who, when hearing a large bang, fled from her house in her bathrobe, leaving her baby behind. The woman explained that she believed a bomb had hit her house, and only turned back to get her baby once she realised this was not the case. The woman did not deliberately hurt her baby, but temporarily lost track of the fact that both she and the baby were in the house; that is, she temporarily did not consider social bonds.
According to Quarantelli, flight behaviour is non-rational rather than irrational. This means that people do not consider or look for surrounding information, rather than perceive everything, and thus make illogical decisions. Hence, panic is more like tunnel vision than behaviour contradicting what people know.
Another example from Quarantelli’s database is the behaviour of a man who was pushing a wheelbarrow when he saw a flaming plane diving towards his general position. The crashing plane made the man run to get away from danger, but he did not seem to consider properly in which direction the danger — a crashing plane — was traveling. The man displayed non-rational flight behaviour.
The definition by Quarantelli suggests that panic can occur, but his research suggests it is an extremely rare event. It is also something that is short-lived and occurs on an individual basis — it is an individual and not a group that ‘panics’. If Quarantelli’s definition holds true, panic is therefore a phenomenon that has limited relevance in most practical situations. People might feel temporarily anxious and briefly act accordingly in extremely stressful situations, but they soon calm down.
One substantial problem is whether myths about panic influence design in a negative way. For example, a designer might try to limit information available to the public to reduce the risk of panic. This will lead to a slower response and potentially larger consequences in case of fire. Other measures, such as wider exits, might have a positive influence on safety since they will reduce total evacuation time. However, it does not make sense to take measures to avoid something that has limited practical relevance for a properly designed building. Rather than designing buildings to avoid panic, designers should instead consider how people will actually behave during an evacuation.
In spite of the criticism of the use of panic (Sime, 1980; Sime, 1984), the term sometimes reappears in the literature, e.g., Helbing, Farkas och Vicsek (2000). This suggests that the discussion about panic being ‘myth or reality’ is not yet over. However, it is clear that panic cannot be discussed without a common understanding of the term — a universally accepted definition.
Daniel Nilsson is with Division of Fire Safety Engineering, Lund University
Brennan P. (1998). Victims and survivors in fatal residential building fires. Proceedings of the 1st International Symposium on Human Behaviour in Fire, University of Ulster, UK, 157–166.
Croker, E.F. (1917). Fire prevention. New York: Dodd Mead & Company.
Fahy, R., Proulx, G. & Aiman, L. (2009). ‘Panic’ and human behavior in fire. Proceedings of the 4th International Symposium on Human Behaviour in Fire, Cambridge, UK, 387–399.
Helbing, D., Farkas, I. & Vicsek, T. (2000). Simulating dynamical features of escape panic. Nature, 407, 487–490.
Home Office. (1934). Manual of safety requirements in theatres and other places of public entertainment. London: His Majesty's Stationery Office.
Nilsson, D. & Frantzich, H. (2008). Design of voice alarms – The benefit of mentioning fire and the use of a synthetic voice. Proceedings of the 4th International Conference on Pedestrian Dynamics, Wuppertal, Germany, pp. 135–144.
Phillips, B.G. (1951). Escape from fire – Methods and requirements. London: E. & F.N. Spon Ltd.
Roytman, M.Y. (1975). Principles of safe fire standards for building constructions. New Delhi: Amerind Publishing Co. Pvt. Ltd.
Sime, J.D. (1980). The concept of panic. in D. Canter (Ed.), Fires and human behaviour. Chichester: John Wiley & Sons, Ltd.
Sime, J. D. (1984). Escape behaviour in fires: ‘Panic’ or affiliation? PhD thesis, University of Surrey, Guilford.
Quarantelli, E. L. (1954). The nature and condition of panic. American Journal of Sociology, 60(3), 267–275.
Daniel Nilsson is with Division of Fire Safety Engineering, Lund University