Читать книгу Security Engineering - Ross Anderson - Страница 74

On March 13, 1964, a young lady called Kitty Genovese was stabbed to death in the street outside her apartment in Queens, New York. The press reported that thirty-eight separate witnesses had failed to help or even to call the police, although the assault lasted almost half an hour. Although these reports were later found to be exaggerated, the crime led to the nationwide 911 emergency number, and also to research on why bystanders often don't get involved.

John Darley and Bibb Latané reported experiments in 1968 on what factors modulated the probability of a bystander helping someone who appeared to be having an epileptic fit. They found that a lone bystander would help 85% of the time, while someone who thought that four other people could see the victim would help only 31% of the time; group size dominated all other effects. Whether another bystander was male, female or even medically qualified made essentially no difference [513]. The diffusion of responsibility has visible effects in many other contexts. If you want something done, you'll email one person to ask, not three people. Of course, security is usually seen as something that other people deal with.

However, if you ever find yourself in danger, the real question is whether at least one of the bystanders will help, and here the recent research is much more positive. Lasse Liebst, Mark Levine and others have surveyed CCTV footage of a number of public conflicts in several countries over the last ten years, finding that in 9 out of 10 cases, one or more bystanders intervened to de-escalate a fight, and that the more bystanders intervene, the more successful they are [1166]. So it would be wrong to assume that bystanders generally pass by on the other side; so the bystander effect's name is rather misleading.