Crowd Science

Module 3, lesson 3 · 35 minutes

Collective resilience and crowd communication

Shared danger can turn a crowd of strangers into a group that looks after itself. Communication that treats the crowd as a partner strengthens that; communication that assumes panic undermines it.

You’ll be able to

  • define collective resilience, and explain how a shared identity can emerge from common fate;
  • describe the evidence from Drury, Novelli and Stott’s study of a music event;
  • derive practical rules for communicating with a crowd, and connect them to the “Information” in Fruin’s FIST and Still’s DIM-ICE.

Introduction

A 2015 paper by Drury, Novelli and Stott has a title that tells its story: “Managing to avert disaster: Explaining collective resilience at an outdoor music event”. The people they surveyed and interviewed had been in a crowd that could have gone badly wrong. It did not. Asked why, participants said that “self-organization in the crowd prevented disaster”.

Self-organisation by whom? Not by stewards or police, in the participants’ account, but by the crowd itself. Lesson 1 showed that people in emergencies help each other far more than the panic story predicts. This lesson asks where that capacity comes from, and how the people managing a crowd can strengthen it rather than get in its way.

Collective resilience

Collective resilience is the capacity of a crowd to coordinate and support its own members in an emergency, which arises when its members come to share an identity.

Drury’s 2018 review pulls the evidence together into a three-step process. A group of strangers faces a danger together: common fate. That shared experience produces an emergent shared identity: people begin to think of themselves as “us”. And a shared identity produces support: people expect help from others, give it, and coordinate.

The second step is the surprising one. The crowd does not need to have been a group beforehand. Commuters on a train, strangers at a concert, shoppers in a mall: their shared identity can be created by the emergency itself. That is what Drury, Cocking and Reicher found among the London bombing survivors: “common fate entails a redefinition of self (from ‘me’ to ‘us’)”.

Top: the process in Drury’s review. Bottom: the relationships found at a music event, where identification with the crowd predicted feeling safe both directly and through expecting help and trusting others.

The evidence from a music event

Drury, Novelli and Stott combined a survey of 48 people with interviews of 20 who had been in the crowd. The survey tested how identification with the crowd related to how safe people felt. Identification predicted feeling safe in two ways: directly, and through two routes that the theory predicts, expecting help from others and trusting them.

Read the bottom of the figure as a story. If I feel part of this crowd, I expect that people around me will help if something goes wrong, and I trust them. Both of those make me feel safer, and I also feel safer just from belonging. In a crowd where people feel safe and expect help, you would expect them to make space, pass on information and help someone who falls.

It is one event, with a survey of 48 people and 20 interviews, and the claim that self-organisation prevented disaster is the participants’ own account. On its own it would be suggestive. Its weight comes from fitting a larger body of work: the survivor studies of lesson 1 and Drury’s review of mass emergencies all point the same way.

Information: the “I” in two frameworks

Two of the field’s risk frameworks give information a letter of its own. In Fruin’s FIST, from Module 2, the I is “INFORMATION (I) upon which the crowd acts or reacts, real or perceived, true or false”. In Keith Still’s DIM-ICE model, which Module 7 covers in full, it is one of three influences on crowd behaviour:

“Understanding the design (D) limitations (capacity, throughput), the information (social media, signage, PA announcements, news) elements and the management (M) systems (processes and procedures) are key to understanding how to influence crowd behaviour in places of public assembly.”

Keith Still, DIM-ICE

Fruin’s phrase “real or perceived, true or false” is a warning. A crowd acts on what it believes, and if the managers do not supply information, rumour will. The resilience research adds a second point: information is not only a set of instructions, it is also part of how a crowd understands what kind of group it is and how it is being treated.

How to talk to a crowd

Put the evidence of this module together and a set of practical rules follows. These are our synthesis, not rules from a single study.

  1. Say what is happening. People delay when they lack information (lesson 1). A clear account of the situation is the fastest way to get them moving.
  2. Say what to do, specifically. “Please leave by the exits on your left, beside the bar” beats “please make your way to the nearest exit”.
  3. Treat the crowd as a partner. Address people as a group that can help: “we need everyone to…”. It invites the shared identity that underpins resilience.
  4. Do not assume panic. “Don’t panic” implies that panic is expected, and euphemisms (“a technical problem”) invite people to wait and find out more.
  5. Keep it consistent and current. The PA, the screens, the signs, the staff and social media should say the same thing, and be updated as things change.
  6. Be legitimate. Explain the reason for a restriction. Lesson 2 showed how treatment that feels illegitimate turns a crowd against the people managing it.

Worked example: one announcement, rewritten

A concourse exit has to close because of an incident outside. The first draft of the announcement reads: “Ladies and gentlemen, due to an operational issue, gate B is closed. Please do not panic.”

A version that follows the rules: “This is an announcement for everyone near gate B. Gate B is closed because of an incident in the street outside; nobody inside is in danger. Please use gate C, to your right as you face the pitch, or gate A, behind the north stand. Stewards in orange will show you the way. We’ll update you in five minutes.”

The second version says what, why, where and when, gives two alternatives so the load spreads, names the people who will help, and promises an update. It never mentions panic.

A festival’s plan says that in an emergency, the screens beside the stage should show a generic “Please remain calm” message until the safety officer decides what to do. What would you change, and why?

“Please remain calm” gives no information about what is happening or what to do, and implies the organisers expect panic. People will delay, seek information from each other and rumour will fill the gap. Prepare specific messages in advance for likely scenarios (which area, which route, what to do), use them immediately, and update them. Coordinate the screens with the PA and the stewards so the crowd hears one consistent account, and phrase it so the crowd is a partner in the response.
  • “Only crowds that were already groups can support each other.” A shared identity can emerge from the emergency itself, through common fate.
  • “Withholding information prevents panic.” Panic is rare; delay is common, and a lack of information causes delay.
  • “Communication is only about giving instructions.” It also tells a crowd how it is seen and treated, which shapes the identity that resilience depends on.