Panic

Panic in fire-behaviour literature generally refers to nonadaptive, irrational or nonsocial flight that reduces the escape possibilities of a group. It involves more than fear, anxiety or rapid movement. Definitions include competitive or injudicious action, and the concept may be applied to a group as well as to an individual.

Research reviewed by Bryan and Proulx found this classic form of panic to be rare in fire incidents. Purposeful, altruistic and reasonable behaviour occurred more often, even when actions later proved unsuccessful. Fear or an inappropriate outcome alone therefore does not establish panic.

The distinction matters for fire engineering. Delays can result from ambiguous information, route unfamiliarity, social commitments or physical crowding rather than irrational flight. Treating every evacuation problem as panic hides those causes and the measures that could address them.

Definition and misuse of the term

Bryan discusses several definitions of panic as a form of harmful flight. Schultz’s definition combines fear with nonrational, nonadaptive and nonsocial behaviour that reduces the group’s escape possibilities. This definition concerns an observed pattern of action and its effects, rather than fear alone.

Fear and purposeful action are not proof of classic panic. Based on Bryan (2002), p. 3-330 and Proulx (2002), pp. 3-344–3-345.

The term is often used after multiple-fatality incidents without physical or social evidence of competitive flight. Bryan notes that participant accounts did not support the classic group-panic explanation in the cases he reviews. A severe outcome does not retrospectively prove the behaviour that supposedly caused it.

The Beverly Hills Supper Club fire is a specific example in Bryan’s review. Interviews and questionnaires from participants did not substantiate classic group panic involving competitive flight towards the exits. That finding challenges a proposed explanation of the incident; it does not imply that the evacuation conditions were adequate.

The distinction between description and explanation changes what evidence is needed. A report of fear describes an emotional state, while a report of pushing describes an action. Neither alone establishes the information, perceived options or physical constraints that produced the action, as Sime’s discussion emphasises.

Outcome can also bias retrospective interpretation. Bryan notes that behaviour is especially liable to be mislabelled when an incident has an unfortunate result. The relevant comparison is therefore between the action and the circumstances perceived at the time, rather than between the action and facts learned afterwards.

Sime's interpretation, as discussed by Bryan, treats panic as a description rather than an explanation. Calling behaviour panic does not identify the information, relationships or physical conditions that produced it. A causal assessment must examine those conditions directly.

The word can also be used differently by witnesses, public officials and researchers. Anxiety during danger may be described as panic even when the person acts deliberately and helps others. A shared technical definition is needed before observations can be compared.

Human behaviour in fire includes many responses that fall outside the technical definition. Investigation, warning, gathering family, route selection and assistance are common parts of evacuation. These activities should not be relabelled as panic merely because they delay immediate departure.

Fear, anxiety and adaptive action

Fear and panic are not equivalent. Bryan reports Sime's distinction between classic panic and behaviour accompanied by concern or anxiety. A frightened person can still evaluate options, follow directions and assist others.

Adaptive behaviour contributes to safety, while nonadaptive behaviour produces an unwanted result. However, an action can be well intended and still fail because the threat was misunderstood or more developed than perceived. Outcome and intention must be considered separately.

An unsuccessful attempt to fight a fire illustrates the difference. If the person believed the action would work, the behaviour may have appeared adaptive from their perspective. Warning others or calling the fire service might have been more effective, but that does not automatically make the attempt irrational panic.

Proulx reports that normal relationships and altruism often persist during emergencies. People may help others, warn them or gather with family before moving. Such actions can consume time while remaining purposeful and socially organised.

Even rapid movement does not prove panic. A person can move quickly in response to a clear threat and an appropriate instruction. The assessment must examine whether the action was competitive, injudicious and harmful to the group's escape rather than relying on speed alone.

Ambiguous information and delayed response

Initial fire information is often unclear. An alarm can be mistaken for a test, and smoke odour may not reveal the fire's location or severity. Occupants commonly seek further information before committing to evacuation.

Pre-movement time includes this recognition, interpretation and preparation. It can contain investigation and movement within the occupied space, so it is not necessarily inactivity. A long delay before travel is not evidence of panic.

Bryan describes recognition, validation, definition, evaluation, commitment and reassessment. These processes help explain why people wait, change their interpretation or select different actions as information develops. They offer more explanatory detail than a general panic label.

Delayed departure can reflect uncertainty rather than panic. Based on Bryan (2002), pp. 3-318–3-319 and Proulx (2002), pp. 3-348–3-350.

Other people can influence whether a cue is treated as an emergency. Passive companions can reinforce an interpretation that conditions are normal, while staff instructions can prompt action. The response is socially shaped rather than necessarily irrational.

Repeated false alarms can further reduce credibility. Proulx explains that occupants may then delay seeking information or responding to an alarm. This is an effect of previous experience and warning management, not proof that occupants are indifferent or panicked.

Affiliation and altruism

Affiliative behaviour leads people to seek others with whom they have emotional ties. Families may gather before departure, and groups may remain together while moving. These actions can change timing and route use without implying nonsocial flight.

Reentry provides another example of behaviour that can be misunderstood. Bryan reports people returning to fight fire, retrieve property, notify others, assist responders or rescue pets. Reentry can be dangerous and obstruct flow, yet it is often deliberate and purposeful.

Reentry also changes the physical movement problem. A person returning through a doorway can oppose the outward stream, even when the purpose is to find a relative or warn someone. Bryan describes this as a reason why an apparently one-way escape route may carry movement in both directions.

Such a return can be deliberate and socially motivated while still increasing exposure or obstructing others. The same action therefore has an intention, an effect on the person returning, and an effect on the evacuation stream. Keeping these separate is more informative than classifying every return as either helpful or panicked.

Keating interviewed 100 participants in fires in single dwellings and reported no instances of panic behaviour. Bryan presents that study as evidence of predominantly helpful and altruistic responses. The finding belongs to that interview sample; it does not establish that competitive flight is impossible in other conditions.

Group assistance can slow movement to the pace of a slower member. Proulx gives the example of a family with a child or older person. A hydraulic model may interpret the resulting speed as inefficiency, while the behaviour reflects group commitment and assistance.

Understanding affiliation helps explain why instructions to act only as isolated individuals may fail. Warning and route arrangements need to account for people seeking and assisting one another. This is a different engineering response from trying to control an imagined irrational crowd.

Crowd density and physical danger

Crowd incidents can involve crushing and asphyxia without classical panic. Proulx describes how density and restricted space can produce forces that individuals cannot resist. Physical danger can therefore arise even when participants are not behaving irrationally.

High densities can produce involuntary movement and shock waves through a crowd. People at the back may not understand that forward movement has stopped or that people at the front are being compressed. This communication failure can worsen conditions without a shared decision to compete.

Crowd danger can arise from density and communication failure, not necessarily panic. Based on Proulx (2002), pp. 3-345–3-346.

Proulx distinguishes competitive rushing away from something, called panic by Fruin, from rushing towards an objective, termed a craze. These specialised descriptions should not be applied indiscriminately to every rapid crowd movement. The physical situation and direction of demand matter.

Crowd density also changes ordinary walking relationships. Speed falls as space between people disappears, and flow through a restriction can produce queues. A crowd-flow analysis should represent those physical conditions rather than attributing every queue to panic.

Continuous routes and management of demand can address some crowd risks. Proulx's review discusses route capacity, dispersion and avoidance of extreme density. These measures target observable physical causes rather than relying on an untested behavioural stereotype.

Proulx also distinguishes crowd management from crowd control. Management concerns the design and operation of a setting in relation to the crowd’s purpose and expected behaviour. Control describes a more disruptive response when those arrangements fail, rather than a substitute for adequate routes and information.

The front and back of a dense crowd can have very different information. People behind may continue moving without knowing that those ahead cannot advance. Pressure transmitted through the crowd can therefore increase without each participant intending to harm others or understanding the danger at the obstruction.

Protective decisions and route choice

Bryan’s decision-process account examines how occupants recognise cues, validate an interpretation, define the situation, evaluate options, commit to action and reassess. Those processes direct attention to the information available at the time. They allow an incident to be examined without first assigning a panic label.

Route choice can reflect familiarity rather than irrationality. Proulx reports that occupants often use familiar entry routes rather than emergency exits they have never used. The resulting path may be longer or more congested, but it has an understandable behavioural basis.

People can also reassess an initial choice. New smoke, staff information or the actions of companions may make another route preferable. Bryan's reassessment process captures this change without treating the first or second choice as automatically irrational.

Competition for an exit can develop when people believe that not everyone will reach safety in time. Bryan describes how reward structure and perceived exit availability can influence cooperative or competitive responses. Such competition requires analysis of conditions rather than a blanket assumption of panic.

Evacuation should therefore be studied as a sequence of information, preparation and movement. Problems at one stage can resemble those at another if only total time is observed. Separating the stages helps identify whether the issue was warning, affiliation, route capacity or actual competitive flight.

Evidence and interpretation limits

Interviews and questionnaires provide evidence about what occupants did and why. They can reveal altruism, uncertainty and purposeful reentry that are invisible in a simple outcome account. Their findings remain tied to the incidents and participants studied.

Evacuation drills offer observations under particular warning and management conditions. They may omit the perceived danger of a real fire or include expectations produced by training. Results from drills and victim accounts should be compared with their contexts retained.

A rare phenomenon should not be described as impossible. The reviewed research challenges the assumption that classic panic is usual, but it does not establish that every crowd or individual always behaves adaptively. The appropriate conclusion is to require evidence of the behaviour and its causes.

Similarly, altruism should not be romanticised as always beneficial. Warning or assisting others can improve outcomes, while reentry or unsuccessful suppression can increase exposure. Purposeful action can have harmful consequences without becoming classic panic.

Models should state which behaviours they represent. Hydraulic approaches can describe movement through route elements but usually omit social influence, investigation and assistance. They cannot be used alone to infer that observed delays or route choices were caused by panic.

Engineering implications

Warning systems should provide information that occupants can interpret. Directive voice messages and staff directions can reduce uncertainty in the cited studies. This addresses a common source of delay without relying on fear or coercive crowd control.

Route design should account for familiar paths and social groups. Clear signage, understandable layouts and staff guidance can support occupants making decisions under uncertainty. Capacity and protection must remain adequate for the routes people actually choose.

Crowd management should address physical demand and bottlenecks. Avoiding excessive density and maintaining communication about stopped movement can reduce danger. These measures are relevant whether or not any participant meets the technical definition of panic.

Emergency plans should recognise assistance and group behaviour. Procedures that assume isolated, immediate and uniform movement may miss family gathering or help for slower occupants. Those activities belong in the timeline and staffing arrangements.

The timing of information is particularly relevant. An action that appears unreasonable after the fire's development is known may have seemed reasonable with the limited cues available at the time. Bryan's account emphasises that inappropriate actions often arise from ambiguous or inadequate information.

See also

• Protective action decision model
• Evacuation lift
• Evacuation modelling

References

• Australian Building Codes Board, National Research Council of Canada, International Code Council, & Department of Building and Housing New Zealand. (2005). International fire engineering guidelines (2005 ed.). Australian Building Codes Board.
• Bryan, J. L. (2002). Behavioral response to fire and smoke. In P. J. DiNenno (Ed.), SFPE handbook of fire protection engineering (3rd ed., pp. 3-315–3-341). National Fire Protection Association.
• Proulx, G. (2002). Movement of people: The evacuation timing. In P. J. DiNenno (Ed.), SFPE handbook of fire protection engineering (3rd ed., pp. 3-342–3-366). National Fire Protection Association.