The more people who witness an emergency, the less likely any one helps

When something goes wrong in a crowd, responsibility gets diluted. The bystander effect describes how the presence of other people lowers any single person’s likelihood of stepping in. The phenomenon was founded in the late 1960s by Bibb Latane and John Darley, whose elegant experiments are still the field’s bedrock. In one, a participant filling out a questionnaire watched smoke seep under the door: alone, most reported it quickly; with two calm-seeming strangers in the room, most sat and did nothing. In another, a participant heard a fellow subject apparently having a seizure over an intercom and was far slower to seek help the more bystanders they believed were also listening.

Those studies isolated three forces, not one. Diffusion of responsibility spreads the obligation thinly across the group. Pluralistic ignorance makes each person read everyone else’s inaction as a signal that nothing is actually wrong. And evaluation apprehension adds a fear of looking foolish by overreacting in front of others.

Its famous origin story is shaky. After the 1964 murder of Kitty Genovese, The New York Times reported that 38 witnesses watched and did nothing. In 2016 the paper acknowledged the account was badly exaggerated. The corrected record shows at least one neighbor shouted at the attacker, scaring him off temporarily; a witness did phone police; and Genovese died in the arms of a friend who came to her side.

The bystander effect is weakened when a situation is clearly dangerous, because real emergencies are recognized faster and prompt more helping.

A 2011 meta-analysis of over 7,700 participants confirmed the lab effect while underscoring that twist. So the cold-city moral drawn from Genovese oversells it. People do help; group size and clarity just shift the odds.