In 1971, a psychologist made a fake prison at Stanford University. Some college students became guards, and others became prisoners. The guards soon started to treat the prisoners badly.

The study was planned for two weeks, but it stopped after six days. For many years, people said that a role can make good people cruel. But later, researchers found old recordings. The staff told the guards to be tough.

The Stanford Prison Experiment is one of the most famous studies in psychology. In 1971, Philip Zimbardo randomly divided 24 healthy male students into guards and prisoners in a fake prison. The study was supposed to last two weeks, but it was stopped on the sixth day because the guards behaved badly and the prisoners became upset. For decades, it was used as proof that ordinary people become cruel simply because of their roles.

However, in 2019 a French researcher examined recordings and notes kept at Stanford. He pointed out that the staff had encouraged the guards to act tough and that many of the rules had been prepared by the staff. Some participants also said they had been acting.

The study was planned to last two weeks. It was stopped on the sixth day. In 1971, Philip Zimbardo randomly assigned 24 male students to be either guards or prisoners in a mock prison, and for half a century the Stanford Prison Experiment was cited as evidence that ordinary people readily become abusive once they are given a powerful role.

The design, though, had serious weaknesses. There was no control condition, and Zimbardo himself acted as prison superintendent. In 2019, Thibault Le Texier analyzed archival recordings and documents. He argued that guards had been coached to be tough, that the staff had set many of the rules, and that some participants had been performing what they believed the researchers wanted.

Haslam and colleagues read the same records differently. In their reading, cruelty arose when leaders fostered a shared group identity, and roles alone did not produce it. A 2007 study raised a separate problem.

People who answered an advertisement for a "prison life" study scored significantly higher on aggression. Zimbardo disputed these criticisms, and opinions are divided. Even so, the simple claim that people spontaneously turn cruel without guidance can no longer be taken at face value.