The Stanford Prison Experiment: What It Really Showed (and What It Didn't)

The corridor was in the basement of Stanford's psychology building, and for six days in August 1971 it was a prison. Doors were relabelled as cells. A cupboard became solitary confinement, nicknamed the hole. Nine young men who had answered a newspaper advertisement were arrested at home by real police as a favour to the researchers, blindfolded, stripped, deloused and given smocks and numbers. Nine others were given khaki, mirrored sunglasses and wooden batons and told they were guards. Philip Zimbardo, the psychologist running it, appointed himself prison superintendent.

What followed has been retold in textbooks, documentaries and at least two feature films. Guards became domineering, inventing punishments and humiliations. Prisoners became listless or rebellious, one apparently suffering acute distress within thirty-six hours. A planned two-week study was stopped after six days. The conclusion that travelled was blunt and unforgettable: put good people in a bad place and the place wins.

Why it became so famous

Timing had a great deal to do with it. The study landed in a decade already arguing about institutional violence, prison reform and the moral responsibility of ordinary participants in systems. It arrived shortly after Stanley Milgram's obedience research, which had made a similar point about situations overwhelming individual conscience. Read together, the two studies seemed to form a single verdict about human nature, and Zimbardo was an exceptionally effective communicator of that verdict.

It also had the shape of a story rather than a dataset. Cells, sunglasses, batons, a superintendent who was also the scientist. You can summarise it in a sentence and picture it immediately, which is precisely what makes a finding spread and precisely what makes it hard to correct later. The obedience research it is usually paired with is covered in the Milgram explainer, and the third member of the classic trio is the Asch conformity study.

The critiques that changed how it is taught

For decades the study was presented as settled. In the 2010s a series of re-examinations, including archival work on Zimbardo's own tapes and notes at Stanford and interviews with surviving participants, made that presentation untenable. The criticisms are worth stating precisely, because the loose version, that it was all fake, is as wrong as the textbook version.

• Guards were briefed, not simply released into a role. Recordings show the research team encouraging guards to be tough and expressing disappointment when they were not. That is guidance toward a result, which undercuts the claim that cruelty emerged spontaneously from the situation.
• Some behaviour was performed. At least one guard has said he consciously played a character modelled on a film role because he thought that was what the study needed. Participants who believe they know the expected outcome tend to supply it, a problem known as demand characteristics.
• The famous breakdown is disputed. The prisoner whose distress helped end the study later said his reaction was substantially exaggerated, partly because he wanted to leave.
• There was no control group and the sample was tiny. Around two dozen young American men, self-selected by an advertisement mentioning a prison study, is not a basis for a general claim about human nature. Later work found that such advertisements attract people scoring higher on traits including aggression and social dominance.
• The experimenter was inside the experiment. Running the prison as superintendent while studying it removed any distance between observation and management, and made the researcher part of the situation he was measuring.
• The ethics would not pass today. Deception, humiliation, unclear routes to withdrawal and genuine psychological harm are exactly the concerns that modern review boards exist to prevent.

What the study still teaches

It would be a mistake to file the whole thing as a hoax and move on. Something real did happen in that basement, and the more careful modern reading is arguably more interesting than the original slogan.

The revised lesson is about roles, permission and identification rather than automatic transformation. People did not become brutal simply because they were handed a uniform. They became brutal when an authority they respected framed cruelty as necessary for a legitimate purpose, gave them cover, and provided no visible cost for going along. A later BBC study by Alexander Haslam and Stephen Reicher found groups did not slide passively into their assigned roles at all, and that leadership and shared identity determined what happened. That is a more demanding conclusion, because it puts the emphasis back on the specific things institutions do: who is told what, who is watched, whose objection counts.

How to read a famous study

The Stanford basement is a useful lesson in something beyond social psychology: how a memorable narrative can outrun its evidence for fifty years. The features that made it spread, a clean plot, a shocking outcome, a quotable moral, are unrelated to whether it was well designed. That is worth remembering every time a single striking study arrives with a ready-made conclusion attached.

The habit that helps is asking three plain questions of any famous finding: how many people, compared with what, and who decided what counted as the result. Applied here, all three answers are uncomfortable. Applied to Milgram, the answers are better but not spotless. Applied to Asch, better again, though the effect turns out to be smaller and more variable than the summary suggests.

Disclosure: MindSnap is our app. We built The Obedience Files as a narrative collection precisely because studies like these are usually taught as a punchline and are far more instructive as history, including the parts where the research itself went wrong. You can see the other narrative collections on the collections page, and our Turning Points scenarios put you inside a decision before revealing what actually happened, which is a fairer test of intuition than reading a verdict someone else reached.

The basement prison ran for six days. The story it produced has run for fifty-five years, which is arguably the more remarkable finding.