A shared universe › Blog › Why we don’t help when there are more people around · Español
A man feels dizzy on an underground platform and sits down on the floor. There are twenty people nearby. Almost all of them see him, and almost all of them glance at the others to work out whether it is serious. A few very long seconds go by before someone steps forward.
In 1968, the psychologists John Darley and Bibb Latané published an experiment that still appears in social psychology textbooks. They invited students at New York University to a conversation about university life. Each one sat in a separate booth and spoke to the others over an intercom, in turns.
In fact there was nobody else: the other voices were recordings. At one point, one of the supposed fellow students began to speak with difficulty and seemed to be having an epileptic seizure. The only thing that changed from one run to the next was how many people the participant believed were hearing the same thing.
When they believed only they could hear it, 85% left the booth to raise the alarm before the seizure ended. When they believed one other person was listening, 62% did. When they believed four more were, 31%. And the more people they thought were there, the longer those who did react took to do so.
The most striking thing was what they did not find. Personality tests were no use for knowing who would help and who would not. The authors concluded that the failure to react was not explained by coldness or lack of concern for the victim, but by something in the situation: knowing that others were there.
They called it diffusion of responsibility. If you are on your own facing someone in trouble, the responsibility is entirely yours. If there are five of you, each person feels they have a small share, and also assumes that someone else will already have done something.
That same year, Latané and Darley published another study in which a room began to fill with smoke. Of those who were alone, 75% went out to report it. Of those who were with two people who, at the researchers’ request, stayed perfectly calm, only 10% did.
That is where the other trap appears. When something is odd and we don’t know whether it is serious, we look at others to decide. If everyone seems calm, each person concludes that nothing is wrong, even though inside everyone is wondering.
In 1970, Latané and Darley brought their work together in a book and described helping as a chain of steps. If one fails, the help doesn’t arrive:
The presence of others can break the chain at several of those points: it distracts, makes the urgent seem normal and spreads responsibility so thin that it is almost nothing.
These studies came about, in part, from an event in New York in 1964: the murder of a young woman, Kitty Genovese. A newspaper reported at the time that dozens of neighbours had witnessed it and done nothing, and that version was repeated for decades in books and classrooms.
In 2007, the psychologists Rachel Manning, Mark Levine and Alan Collins reviewed the trial and police records, and concluded that there was no evidence that there had been thirty-eight witnesses, or that they saw the crime, or that they did nothing: at least one neighbour shouted at the attacker, and it is possible that someone called the police. In 2016, the newspaper itself acknowledged that its account had greatly exaggerated the number of witnesses and what they had actually perceived.
The bystander effect exists and has been confirmed many times in the laboratory. But the picture of a crowd looking the other way is bleaker than the data say.
In 2019, a team led by Richard Philpot, of Lancaster University, analysed 219 real arguments and fights recorded by CCTV cameras in Amsterdam, Lancaster and Cape Town. In 91% of cases, at least one person intervened: they asked for calm, stepped in between or comforted the victim. And the more people there were around, the more likely it was that someone would help.
This doesn’t contradict Darley and Latané. Each individual, taken separately, may feel less called on to act when there are more people; but among many, it is easier for someone to take the step. In addition, a review of studies led by Peter Fischer in 2011, with data from more than 7,700 participants, found that the effect weakens when the situation is seen as dangerous, and that in some dangerous emergencies it disappears altogether.
Knowing it exists already helps you not to fall into it. And there are simple gestures that undo the sharing-out of responsibility:
And if you are the one who needs help in the street, the same applies: pick someone out with your eyes and ask that person.
The bystander effect doesn’t happen only in emergencies. It happens in a group chat when someone says they are struggling and everyone assumes someone else will write to them privately. It happens in families, with “I’m sure her sister has already called her”. It happens with the neighbour nobody has seen for days.
Asking once more costs little. Assuming someone else will do it can leave a person with nobody. Sometimes the kindness of a stranger arrives just when nobody else had felt addressed.
In a sky full of stars it is also easy to think that others will already have stopped by each one. That is why there is “Keep someone company”, which takes you to the ones that have had the least company. Your touch may be the first.