Pluralistic Ignorance: The Room That Filled With Smoke

On the thirteenth of March 1964, Kitty Genovese was murdered outside her apartment building in Queens. The New York Times reported that thirty-eight respectable, law-abiding citizens had watched from their windows and none had called the police. The details of that account have since been substantially corrected — the number was wrong, most witnesses saw fragments and heard rather than watched, and at least one person did call — but the story had already prompted two young psychologists, John Darley and Bibb Latané, to ask a question that turned out to be more important than the case that inspired it.

Their answer, published in 1968, is the reason this chapter exists.

In one study, a student sat in a room filling out a questionnaire when smoke began pouring in through a wall vent. It kept coming until the room was hazy enough to make the paperwork hard to read. Alone, seventy-five percent of subjects left the room to report it, most within two minutes. Seated with two confederates instructed to glance up, shrug, and keep writing, only ten percent reported it. Some sat for the full six minutes rubbing their eyes and waving smoke away from the questionnaire.

In another, subjects heard a fellow participant apparently suffering a seizure over an intercom. When they believed they were the only listener, eighty-five percent went for help. When they believed four others were listening, thirty-one percent did.

Two mechanisms combine, and they are frequently collapsed into one, which makes both harder to counter. Diffusion of responsibility divides the felt obligation to act by the number of people who could act — if six people are present, the duty feels like a sixth of a duty, and the cost of intervening falls entirely on whoever moves. Pluralistic ignorance is the second and stranger one. In an ambiguous situation, each person looks to the others for evidence of how serious it is. Everyone else is doing the same thing. What each person observes is a room full of calm people, and calm is read as data. The group converges on a belief — nothing is wrong here — that no individual in it actually holds.

That last property is what makes pluralistic ignorance different from ordinary cowardice. The information failure is genuine and mutual. Everyone is being misled, and everyone is a source of the misleading.

It also explains a great deal that is otherwise puzzling about institutions. The board that watched a fraud develop and said nothing, in which every member privately had doubts. The department where an abusive manager operated for years and each person assumed the others knew something that made it acceptable. The team that shipped a product it knew was unsafe. In each, the participants are usually not accomplices. They are people who read the room's silence as consent and contributed their own silence to it.

The countermeasure is unglamorous and works. Ambiguity is what sustains the effect, so destroy the ambiguity out loud. Say what you are seeing in plain words — this is smoke, this man needs an ambulance, I think this transaction is fraud — because the moment one person names the situation, the calm of the others stops being evidence. Then remove the diffusion by assigning the duty to a specific person rather than to the room: you, in the blue coat, call 999. Emergency medical training has taught this for decades for exactly this reason.

You do not have to be brave to break it. You have to be first, and specific.