Social Psychology
Stanford Prison Experiment
Zimbardo's 1971 study, its sweeping conclusions, and why most psychologists no longer accept them
The Stanford Prison Experiment, conducted by Philip Zimbardo at Stanford University in August 1971, randomly assigned 24 male undergraduates to play guards or prisoners in a basement mock prison. The study was halted after six of the planned fourteen days when guards began abusing prisoners and several prisoners showed acute distress. Zimbardo argued the situation, not the dispositions of the participants, drove the cruelty. Subsequent investigations — most damagingly Le Texier's 2019 archival work and Blum's 2018 Medium piece — revealed coaching of guards, fabricated breakdown footage, and selection bias that undermined the original interpretation. The study now stands as a cautionary tale about ethics and methodology rather than a demonstration of situational power.
- InvestigatorPhilip Zimbardo, Stanford (August 1971)
- Sample24 male college students, randomly assigned
- DurationHalted on day six of fourteen
- Original claimSituation overrides disposition
- Major critiqueLe Texier (2019), Blum (2018)
- StatusFailed direct replication (BBC, 2002)
Interactive visualization
Press play, or step through manually. The visualization is yours to drive — try it before reading on.
Watch the 60-second explainer
A condensed visual walkthrough — narrated, captioned, under a minute.
Why the Stanford prison study matters
- Research ethics. Catalyst for stricter IRB review and informed consent standards.
- Demand characteristics. Showed how strongly experimenter cues shape behavior.
- Replication crisis. Case study in how compelling narratives outlive evidence.
- Cultural reach. Two films, dozens of textbooks, public testimony to Congress.
- Abu Ghraib parallels. Used in real-world legal defense and reform debates.
- Social identity revision. BBC follow-up established cooperative possibilities.
- Methodology training. Frequently used to teach students to evaluate evidence critically.
Common misconceptions
- Guards spontaneously turned cruel. Le Texier's archives show explicit briefing toward harshness.
- Participants were ordinary. Self-selection on the recruiting ad pre-screened for aggressive traits.
- Prisoner #8612 truly broke down. Subsequent testimony suggests he performed to secure release.
- Situation always trumps disposition. The BBC study showed cooperative outcomes are equally possible.
- The findings explain Abu Ghraib. Real prison abuses involve organizational, not laboratory, dynamics.
- The study was rigorous. Zimbardo served as both researcher and superintendent — a fatal dual role.
Frequently asked questions
What happened in the original study?
Zimbardo converted the basement of Stanford's psychology building into a mock prison, hired Palo Alto police to "arrest" the prisoner participants, and assigned guards uniforms, sunglasses, and batons. Within days, guards harassed prisoners with humiliations, exercise punishments, and sleep deprivation. Prisoner #8612 had a breakdown on day two and was released. Five prisoners were released early. Zimbardo's then-girlfriend Christina Maslach, visiting on day six, objected to the conditions, and the study ended.
What was the original interpretation?
Zimbardo argued that ordinary, mentally healthy young men placed in a coercive role transform into abusers without dispositional cruelty — that situation overwhelms personality. He extended the claim to Abu Ghraib in 2004 testimony, framing US military abuses as predictable products of context. The Lucifer Effect (2007) summarizes his case. The view became dogma in introductory psychology curricula for decades.
What did Le Texier reveal?
French researcher Thibault Le Texier (American Psychologist, 2019) examined the Stanford archives and tape recordings. Guards had been explicitly briefed on creating "boredom, frustration, fear" and given specific tactics. Prisoner #8612's "breakdown" was reportedly performed to secure release. Several prisoners reported afterward they had been acting, not breaking. Zimbardo had intervened directly to escalate guard behavior. The Lucifer Effect's situational claim could not survive these findings unchanged.
How does the BBC Prison Study compare?
Reicher and Haslam's 2002 BBC Prison Study attempted a controlled replication broadcast on television. Guards did not spontaneously brutalize; prisoners organized and eventually overturned the hierarchy. The pattern suggested social identity dynamics — when do groups define themselves cooperatively versus oppositionally — explained behavior better than role assignment alone. The BBC results, published in the British Journal of Social Psychology (2006), undercut Zimbardo's universal situational claim.
Were the participants representative?
Carnahan and McFarland (2007) compared volunteers responding to ads for "psychological study of prison life" versus generic "psychological study." The prison ad attracted volunteers higher in aggression, authoritarianism, narcissism, and Machiavellianism. Self-selection — not just situation — produced a sample primed for the observed dynamics. The original study's claim of "ordinary young men" is hard to defend given this sampling.
What ethical issues remain salient?
Many. Insufficient informed consent (participants were not told they could be subjected to humiliation), inadequate prisoner safeguards, principal investigator dual role as research lead and prison superintendent, delayed termination despite visible distress, and shaping of guard behavior. The study contributed to APA ethical guidelines tightening in the 1970s and IRB review reforms. Today the design would not pass any review board.
What should the study be remembered for?
Multiple lessons. (1) Power dynamics in coercive institutions deserve continued empirical attention. (2) Demand characteristics, sampling, and experimenter influence can manufacture seemingly spontaneous behavior. (3) Foundational findings need rigorous replication before generalization. (4) Compelling narratives outlive their evidentiary support. The historical importance is real; the original conclusions are not.