The Human Factor in Evacuation

Engineers sometimes model a crowd leaving a building as a fluid flowing out through pipes. It is a useful calculation and a dangerous mental picture, because real people do not behave like water. Water starts moving the instant the tap opens. People do not start moving when the alarm sounds - and the gap between the alarm and the first step is where a great many fire deaths are decided.

Decades of investigation into real fires have overturned the comfortable myth of instant, orderly, rational evacuation. People are social, busy, sceptical and attached to their routines, and under the ambiguity of an early fire they tend to carry on, seek information, and delay rather than drop everything and run. This is not stupidity or panic - in fact unreasoning panic is far rarer than films suggest; the more common and more lethal behaviour is the opposite, under-reaction. People ignore alarms they assume are false, finish a transaction, gather belongings, look for family, or wait to see what others do before they accept that anything is wrong.

> The dangerous human response in a fire is usually not panic but delay - people carry on, seek information and wait, losing the minutes that decide survival.

For a designer this is both liberating and sobering. It means escape time is not fixed by corridor widths and travel distances alone; a large, variable part of it is human, set by how quickly people accept there is a fire and decide to act. And it means design and management can shorten that human time - or, done badly, lengthen it fatally. The rest of this lesson is about that hidden clock and how to design for the people who really exist, not the obedient evacuees we wish they were. Fire plans drawn for ideal occupants fail real ones; plans drawn for real ones protect everybody. The research matters here precisely because intuition misleads: left to common sense, designers and managers assume people will simply leave when told, and build alarms, routes and procedures on that false premise - which is why so many escape strategies look sound on paper and stall in the first two minutes of a real event.

Fire engineers split the total time people need to escape - the Required Safe Escape Time, RSET - into two parts, and the split is the single most important idea in this lesson:

RSET (time people NEED)  =  pre-movement time  +  travel time
  pre-movement  =  recognition (notice + believe the cue)
                +  response    (decide + act: finish, gather, seek info)
  travel        =  actually walking the route to safety

SAFE  when  RSET  <  ASET (time until conditions become untenable)

Architects instinctively design for travel time - the walking part, governed by distances, widths and stairs. But investigations repeatedly find that pre-movement time is often the larger and far more variable part, sometimes dwarfing the walk itself. The seconds or minutes people spend deciding whether this is real and what to do can exceed the time it then takes them to leave the building.

Pre-movement has two stages. Recognition: the occupant has to notice a cue - an alarm, smoke, a warning - and believe it; a bland, cried-wolf alarm that has sounded falsely many times before is widely ignored. Response: having accepted there is a fire, the person decides and acts - and here they finish tasks, collect belongings, search for children or colleagues, make a phone call, or simply wait for an authority figure to tell them to go.

Management choices feed this too: an alarm regime riddled with false alarms teaches a whole building to disbelieve the warning, so the reliability of the warning is itself a life-safety design concern. The design consequence is direct. If you can shorten pre-movement - with an unambiguous alarm, ideally a voice message that tells people plainly what is happening and what to do, with clear information and with trained staff who start the movement - you buy time that is often cheaper and larger than anything you could win by shaving a few metres off a travel distance. You cannot design the walk well and ignore the wait.

Beyond delay, people move in patterns that design must anticipate. The strongest is the familiarity bias: in an emergency, people overwhelmingly head for the route they know - usually the door they came in by - and ignore nearer, unfamiliar exits they have never used. A mall with ample emergency exits can still funnel a crowd back toward the main entrance, because that is the way people remember. Exits that are hidden, that look like staff-only doors, or that are never used in normal life are, in the moment, nearly invisible.

People are also intensely social in emergencies. They engage in herding - following the crowd and copying what others do - which means a confident first mover can lead many to safety, while a hesitant majority can hold everyone in place. They wait for authority and for permission to act. They tend to evacuate as groups (families, colleagues) and will not leave without one another, which slows the flow. And affiliation draws them toward the familiar and toward their companions rather than toward the nearest exit sign.

> In a fire, people go the way they know and do what others do. Design and signage must fight the pull of the entrance and make the right exit the obvious one.

This is why a pre-event walk of the exits, or even a single prior use of an alternative door, measurably speeds later evacuation: the once-strange exit becomes a known one. Buildings that only ever move people in and out through a single entrance are quietly training their occupants to ignore every other way out. None of this behaviour is irrational; it is deeply human, and it is predictable enough to design for. The lesson is not to wish people behaved differently but to build for how they actually behave - make escape routes visible, lit and legible, give exits an everyday presence so they are not strangers in the emergency, and provide the clear cues and human prompts that turn herding from a trap into a current flowing the right way.

Put the behaviour together and a design agenda falls out - a set of moves that shorten pre-movement and steer real occupants to safety. Treat these as principles to apply with the code and, on complex buildings, a fire engineer.

DESIGN FOR REAL BEHAVIOUR
  RECOGNISE faster   clear, distinct alarm; voice messaging where apt;
                     detection that warns early (Module 5)
  DECIDE faster      plain instructions; trained staff / wardens to initiate;
                     remove ambiguity (say plainly: this is an evacuation)
  MOVE the right way  visible, lit, legible exits + signage; exits used in
                     daily life; counter the pull of the entrance
  GIVE OPTIONS       two remote routes, so a known door being blocked is not fatal
  PROTECT the slow   refuge + accessible egress for those who cannot rush
                     (Module 3.4; Universal and Accessible Design)

Several of these cost little and repay hugely. A voice alarm that says what is happening cuts the recognition-and-denial delay that a bell cannot touch. Way-finding - signage, exit visibility, emergency lighting - fights the familiarity bias. Staff and wardens are not something you draw, but a good strategy assumes them and the building supports them; trained people who start moving break the herd's hesitation. And giving more than one protected route respects the fact that people will try the door they know first, so the strategy must survive their trying the 'wrong' one.

The honest limits apply here too. Human behaviour and evacuation timing are a specialist field; for crowded, tall, phased-evacuation or otherwise complex buildings, the behavioural assumptions and the RSET-versus-ASET analysis belong to a fire engineer, sometimes supported by evacuation modelling. Your contribution as architect or interior designer is to refuse the fantasy of the instant, orderly evacuation and to design, from the first sketch, for the hesitating, social, familiarity-bound human who will actually be there - because the building that protects the real person protects everyone.