The Stanford prison experiment. · published 2 September 2026
The Stanford prison experiment is remembered as ordinary men spontaneously turning brutal when handed power, but Zimbardo’s own archives show guards were briefed to create an atmosphere of fear and corrected when judged too soft — and the careful replication, the BBC Prison Study by Reicher and Haslam, produced the opposite outcome, with guards failing to cohere and prisoners organising against them.
These are the working notes, unedited. Everything the video
states as fact comes from here, and nothing else may be stated as fact.
The sections on limits and on what the finding does not show are
part of the brief, not an afterthought — see
the method.
Run by Philip Zimbardo at Stanford University in August 1971. Twenty-four male volunteers, screened as psychologically healthy, were randomly assigned to be guards or prisoners in a mock prison built in the basement of the university’s psychology building. It was planned to run two weeks and was stopped after six days.
The popular telling is that ordinary young men, simply given power, turned spontaneously cruel — that the situation overwhelmed their character. This is the version in textbooks, in the 2015 feature film, and in most retellings.
The corrections are substantial and documented.
It was a demonstration, not a controlled experiment. There was no control group and no manipulated variable, so it cannot establish what causes what.
Thibault Le Texier’s “Debunking the Stanford Prison Experiment” (American Psychologist, 2019) drew on Zimbardo’s own archives at Stanford. The records show guards were not left to invent their behaviour: they were briefed to be tough, to create an atmosphere of fear, and were coached and corrected by the research team when they were judged insufficiently harsh. The cruelty that the story treats as spontaneous was substantially instructed.
Recordings and interviews published by Ben Blum in 2018 add to this. The guard whose harshness is most often cited, Dave Eshelman, has said for years that he was consciously performing a character. The prisoner whose breakdown ended his participation, Douglas Korpi, later said he had exaggerated it because he wanted to leave.
The recruitment advertisement asked for volunteers for “a psychological study of prison life”. Carnahan and McFarland (2007) ran an experiment placing that advertisement alongside one identical but for the words “of prison life”, and found the prison wording drew volunteers scoring higher on aggression, authoritarianism, social dominance and narcissism. The sample was not a random slice of ordinary people.
When the study was reconstructed for the BBC Prison Study by Reicher and Haslam (British Journal of Social Psychology, 2006), participants did not slide into their assigned roles. The guards failed to cohere as a group and the prisoners eventually organised and challenged them.
Zimbardo disputed Le Texier’s characterisation and maintained that the study’s central point stands. The disagreement is real and may be mentioned. What may not be said is that the study demonstrated ordinary people spontaneously becoming cruel, because the conditions were shaped by the experimenters.