The Stanford Prison Experiment

1. Historical and Intellectual Context

By 1971 social psychology had spent two decades wrestling with the legacy of the Second World War. Hannah Arendt's reporting on the Eichmann trial had popularized the idea of the "banality of evil." Stanley Milgram's Yale obedience studies, conducted a decade earlier, had shocked the discipline by showing that ordinary adults would deliver what they believed were dangerous electric shocks at the instruction of an unremarkable experimenter. Solomon Asch's conformity experiments had already established that group pressure could override perceptual judgment.

Philip Zimbardo, then a young Stanford professor who had been a high-school classmate of Milgram in the Bronx, wanted to take the situationist argument one step further. Instead of focusing on a single moment of compliance, he wanted to ask what would happen when ordinary people inhabited a coercive institutional role over time. The American prison system was a natural object of attention: a 1971 prison riot at San Quentin and the Attica uprising in New York occurred in the same period, and prison reform was a live political issue.

The U.S. Navy and Marine Corps, concerned with conflict between guards and inmates in military brigs, had funded the research through the Office of Naval Research. Zimbardo also drew on a body of work showing that anonymity and deindividuation — wearing masks, uniforms, or being part of a crowd — increased aggressive behavior in laboratory tasks. He wanted to test whether the architecture of a prison, plus the role asymmetry it imposed, could be enough to produce institutional cruelty without selecting for particularly cruel individuals.

It is important to remember the intellectual climate. Personality psychology had been wounded by Walter Mischel's 1968 critique arguing that trait measures were poor predictors of behavior across situations. The pendulum was swinging hard toward situationism. The Stanford Prison Experiment was designed in that climate, and its findings, as Zimbardo reported them, fit the climate perfectly. That fit is part of why the study became so influential — and part of why later critics have pressed so hard on whether the data justify the story.

2. Research Questions

The official research question was deceptively simple: would normal, psychologically healthy young men, randomly assigned to play the role of guard or prisoner in a simulated prison, begin to behave in ways characteristic of real guards and real prisoners? Embedded in that question were several more specific claims that Zimbardo wanted to test.

• Whether institutional roles themselves can override individual personality and prior values
• Whether deindividuation (uniforms, mirrored sunglasses, numbers in place of names) accelerates aggressive behavior
• Whether the prison environment generates conflict regardless of who occupies the roles
• Whether "ordinary" people contain the latent capacity for the kinds of abuses observed in real correctional systems

Critics later pointed out that none of these questions had a clean operational test. There was no control group, no comparison condition with different role labels, and no pre-registered prediction about what counted as "becoming" a guard or prisoner. The study was framed more as a demonstration than as a hypothesis-testing experiment, and that framing affected what kinds of evidence it could yield.

3. Method and Procedure

Recruitment

An advertisement in the Palo Alto Times offered $15 a day (roughly $115 in 2026 dollars) for participation in "a psychological study of prison life." About 75 men responded; 24 were selected after interviews and psychological screening intended to exclude anyone with a history of psychological problems, criminal convictions, or substance use. The final pool was assigned at random to the guard and prisoner conditions, with a small number held as alternates.

The Mock Prison

The basement corridor of Jordan Hall was converted into a "prison." Three small offices became cells with bars on the doors. A closet was designated as solitary confinement, nicknamed "the Hole." A larger room served as the guards' quarters. Cameras and a hidden microphone allowed continuous observation; Zimbardo himself acted as "Prison Superintendent," with a graduate student, David Jaffe, as "Warden." This dual role of researcher and prison authority would later become one of the most criticized features of the design.

Arrest and Intake

To heighten realism, the Palo Alto police agreed to "arrest" prisoners at their homes on a Sunday morning. They were handcuffed, read their rights, taken to the station, fingerprinted and booked, then transported blindfolded to the basement. On arrival they were stripped, deloused, and dressed in smocks (without underwear) and a stocking cap to simulate shaved heads. A chain was placed around one ankle. They were addressed only by their assigned number.

Guard Briefing

Guards were given khaki uniforms, mirrored sunglasses to prevent eye contact, whistles, and wooden batons. According to Zimbardo's later writing, they were told not to use physical violence but were otherwise instructed to maintain order in the prison. Archival audio of the actual briefing, recovered decades later by Thibault Le Texier and others, has shown that the guards received more specific instructions about creating frustration, fear, and a sense of arbitrary power than the published accounts suggest. This is one of the central modern critiques.

Daily Schedule

Prisoners were subjected to counts at all hours, including the middle of the night, push-up punishments, restricted bathroom access, and varying privileges. Guards worked eight-hour shifts in teams of three. There were no formal rules at the outset; rules were generated on the fly, often punitively. Observation continued around the clock.

4. Participants and Setting

The 24 selected participants were almost all white, almost all middle-class, and all male college students or recent graduates. They were strangers to one another at the start. Pre-experimental psychological screening using standard inventories of the era suggested they were within normal ranges on personality and pathology measures. Subsequent research has questioned whether such screening could have detected the relevant individual differences in dispositions like authoritarianism or social dominance.

A widely cited 2007 study by Carnahan and McFarland recruited volunteers using two versions of an advertisement — one identical to Zimbardo's 1971 wording mentioning "prison life," and one omitting that phrase. Those responding to the prison-themed advert scored significantly higher on measures of aggressiveness, authoritarianism, social dominance, and Machiavellianism, and lower on empathy and altruism. This suggests that self-selection alone may have biased the sample in directions that would make harsher behavior more likely, independent of the situation.

The physical setting itself was small, hot, and windowless. There were no clocks. Lights were left on. The constant surveillance and the absence of a clear distinction between research and theater are essential to understanding what unfolded. Participants were not simply behaving in an environment; they were performing in an environment they had been told was a prison, watched by researchers who had a stated interest in seeing the prison work.

5. Results

The Standard Narrative

In Zimbardo's account, the simulation deteriorated rapidly. By the second day a prisoner rebellion broke out; guards crushed it with fire extinguishers and increasingly punitive measures. Prisoner #8612 began screaming and showing signs of acute distress within 36 hours and was released. Other prisoners followed, often with stress symptoms — uncontrolled weeping, disorganized thinking, somatic complaints. By day six, when Zimbardo's then-girlfriend (later wife) Christina Maslach visited and objected to what she saw, the study was terminated.

Guards, in this account, fell into recognizable types. A subset — most famously a participant nicknamed "John Wayne" — became actively cruel. Others were "tough but fair." A third group did the minimum and tried to be kind to prisoners. The lasting image, replayed in countless textbooks and documentaries, is of guards who arrived as ordinary college students and within days were marching prisoners with bags over their heads.

What the Numbers Actually Show

The "results" of the Stanford Prison Experiment are difficult to summarize because it was largely an observational case study rather than an experiment with measurable outcome variables. Zimbardo and colleagues never published the study in a peer-reviewed empirical journal; the main publication was a chapter in Naval Research Reviews in 1973 and later writings and films. Subsequent analyses have noted that only about a third of the guards engaged in clearly abusive behavior, another third followed orders neutrally, and a third actively tried to support prisoners. The dramatic textbook version compresses this distribution into a single arc of escalating cruelty.

Selected Behavioral Observations

• Push-up punishments and forced exercise increased over the six days
• Sleep was interrupted by counts conducted at 2:30 a.m. and other irregular hours
• Several prisoners were placed in the "Hole" closet for periods of time
• Sexualized humiliation tasks emerged later in the study
• Five prisoners were released early because of acute psychological distress
• One prisoner staged a hunger strike and was placed in isolation

6. The Researchers' Interpretation

Zimbardo's reading of the data was emphatic. He argued that the experiment demonstrated the awesome power of social roles and institutional structures to override individual character. In his framing, the guards were not sadists; they were ordinary people placed in a sadistic role. Likewise, the prisoners were not weak; they were ordinary people whose autonomy had been stripped by the architecture of the institution.

This situationist reading carried obvious political weight. If prison brutality was structural rather than the product of bad apples, then reform must focus on institutions rather than individual screening. Zimbardo developed this position over the following decades, culminating in his 2007 book The Lucifer Effect, which extended the analysis to the U.S. military abuses at Abu Ghraib. Zimbardo testified for the defense of a guard charged in the Abu Ghraib case, arguing that situational forces were largely responsible for the abuses photographed there.

The interpretation also drew on Zimbardo's prior work on deindividuation. Uniforms, sunglasses, numbers in place of names, and group cohesion were said to weaken individual moral restraint. Combined with diffusion of responsibility and the institutional authority granted to the guard role, the conditions, in his view, predictably produced the observed cruelty.

This account became one of the most widely repeated narratives in undergraduate psychology. Generations of students learned it as a parable about themselves: that they too could become a guard or a prisoner if placed in the wrong institutional structure. The argument was rhetorically powerful and morally serious. The question raised by later researchers was whether the data could actually bear the weight of that argument.

7. Modern Reanalyses and Criticisms

Demand Characteristics

The first wave of critique came from social psychologists who suspected demand characteristics. Participants knew they were in a study about prison life, run by a researcher who had a clear interest in observing prison-like behavior. The advertisement had primed them. The setting was theatrical. Cameras were rolling. From this perspective, what the experiment captured was not the emergence of brutality from neutral conditions but the willingness of college students to play the roles they had been recruited to play.

The Le Texier Archival Work

French historian and researcher Thibault Le Texier published the most thorough archival reanalysis to date in his 2018 book Histoire d'un Mensonge ("History of a Lie"), drawing on documents and audio recordings held at the Stanford University Archives. Le Texier showed that the guards were not simply released into a role with minimal instruction. They received an explicit briefing in which Jaffe and Zimbardo described the desired psychological effects — disorientation, dependency, fear, helplessness — and discussed concrete tactics for producing them. The guards were, in effect, given a script.

The Korpi and Eshelman Tapes

Two participants central to the textbook story have given strikingly revisionist accounts in recent decades. Prisoner Douglas Korpi, whose on-camera breakdown is one of the iconic images of the study, later said he was performing — that he wanted out of the study because he was missing study time for his graduate-school exams, and that he believed an emotional outburst would secure release. Dave Eshelman, the guard known as "John Wayne," has said he consciously developed a Strother-Martin-from-Cool Hand Luke persona because he understood that the researchers wanted dramatic behavior and that he was contributing to important research. These are not the only participant accounts, and other participants describe the distress as genuine, but they significantly complicate the canonical narrative.

Reicher and Haslam's BBC Prison Study

In 2002, Stephen Reicher and Alex Haslam ran a partial re-enactment in collaboration with the BBC. With ethical safeguards, they randomly assigned participants to guard or prisoner groups in a simulated environment. The results were nearly the opposite of Stanford. Guards in the BBC study were uncomfortable with authority and reluctant to impose order. Prisoners cohered as a group and successfully challenged the regime. Reicher and Haslam argued that what determined behavior was not the role label but whether participants identified with their assigned group and saw its actions as legitimate. They proposed an "engaged followership" model in which people behave cruelly when they actively endorse the goals of an authority they identify with — not because situations automatically deform them.

Methodological Failings

• No control or comparison condition
• Researcher (Zimbardo) was simultaneously a participant (Superintendent), violating observer independence
• Self-selection of participants likely biased the sample (Carnahan and McFarland, 2007)
• Guards were explicitly coached rather than left to define their own roles
• Only a minority of guards behaved abusively; the modal behavior is rarely emphasized
• The study was never published as a peer-reviewed empirical article
• Outcome measures were largely qualitative observations selected after the fact

2018 Reappraisals

In 2018, journalists and scholars including Ben Blum and Le Texier published widely read pieces compiling the archival, audio, and participant evidence. The cumulative effect was a substantial dent in the experiment's standing. Many textbooks have since added caveats; some have removed the study entirely. Zimbardo himself has responded to the critiques in writing and in interviews, defending the broad situationist conclusion while acknowledging that the original framing was sometimes oversimplified.

8. Ethical Considerations

The Stanford Prison Experiment would not pass any contemporary Institutional Review Board. Participants experienced acute psychological distress that was foreseeable and, after the first day, observed in real time without the study being stopped. Informed consent, as we now understand it, was inadequate: participants agreed to participate in a study of "prison life," but they were not told that they could be arrested at home, strip-searched, or placed in solitary confinement, nor were they given a meaningful way to leave once the simulation began.

The simultaneous role of Zimbardo as researcher and Prison Superintendent created an irreconcilable conflict of interest. As Superintendent, he had reasons to keep the simulation running; as researcher, he had a duty to protect participants from harm. Christina Maslach, then a recent Stanford PhD, has described visiting the basement on the evening of day five, seeing prisoners being marched to the bathroom with bags over their heads, and confronting Zimbardo. Her intervention — what she has called "you have to stop this" — is widely credited as the trigger for terminating the study.

By modern standards, multiple violations are evident. There was no independent monitor empowered to halt the study. The right to withdraw was practically obstructed; one prisoner who tried to leave was told he could not. Debriefing occurred but was retrospective; there was no ongoing safeguard during the simulation. Documentation by APA Ethics Code standards, which themselves were revised in the 1970s partly in response to studies like this one, would now require continuous risk assessment, distinct roles, and clearer exit options.

The study's ethical legacy is two-edged. It is part of the cluster of studies — alongside Milgram's obedience research and the unethical Tuskegee syphilis study — that prompted reform of human-subjects research in the United States, leading to the National Research Act of 1974 and the Belmont Report. In that sense, the harm it caused has been partially redeemed by institutional reforms. But the harm was real, and several participants have described lasting effects from the experience.

9. Influence on Psychology

Whatever its scientific merits, the Stanford Prison Experiment has had an outsized cultural and disciplinary footprint. For five decades it has appeared in almost every introductory psychology textbook in English, often paired with Milgram's obedience research as the canonical demonstration of situationist social psychology. It has been the subject of feature films, documentaries, plays, and novels. The phrase "Stanford Prison Experiment" has entered common usage as shorthand for institutional corruption.

Within psychology, the study contributed to the consolidation of social-role theory and to a renewed interest in deindividuation, conformity, and obedience. It was cited approvingly in influential analyses of cult dynamics, military abuses, workplace harassment, and the dynamics of bystander silence. Zimbardo's later work on heroism — the Heroic Imagination Project — emerged in part from his attempt to find positive counterparts to the situationist account of evil.

The experiment also influenced policy debates around solitary confinement, prison conditions, and police training. Reform advocates cited it as evidence that prison environments must be designed with the recognition that ordinary people will behave badly inside them. Whether or not the study scientifically established that claim, it played a documentable role in moving public discussion in that direction.

Within social psychology specifically, the long contest over the study has produced something arguably more useful than the original results: a more sophisticated framework for thinking about how groups, roles, identification, and authority interact. The engaged-followership model developed by Haslam, Reicher, and colleagues is in part a direct descendant of the Stanford study, even though it largely contradicts the simple version of Zimbardo's conclusion. The field is in some ways stronger for having had to argue about the Stanford basement for fifty years.

10. What the Experiment Means Today

So what should a student, a teacher, or a reader make of the Stanford Prison Experiment in 2026? The honest answer is that the study is best understood as a complicated historical event rather than a clean piece of evidence. It does not, on its own, prove that situations alone produce cruelty. It does not show that random assignment to a role is sufficient to transform a person. It does not provide a controlled test of any specific psychological mechanism.

What it does show, with the help of subsequent scholarship, is that institutional contexts, role expectations, instructional framing, group identification, and self-selection can interact to produce serious harm — and that researchers who care about a finding can unintentionally shape conditions to produce it. Both lessons are important. The first is the situationist insight, properly chastened. The second is the methodological insight that has come to characterize the credibility revolution in psychology more generally.

The events at Abu Ghraib in 2003 reanimated the original questions. Photographs of U.S. soldiers humiliating Iraqi detainees bore a disturbing resemblance to images from the Stanford basement. Zimbardo's expert testimony in the case argued that situational forces — leadership vacuum, isolation, dehumanizing language, ambiguity of rules — explained much of what occurred. Critics responded that engaged followership, in which soldiers actively endorsed the goals of an in-group and saw detainees as legitimate targets of cruelty, was a better fit to the evidence. Both readings drew on the Stanford experience without depending on its scientific status.

The contemporary teaching of the Stanford Prison Experiment increasingly reflects this layered view. Rather than presenting it as a knockdown demonstration, instructors use it to illustrate the entanglement of design, demand characteristics, ethics, and theory. Many courses now pair the original 1971 account with Le Texier's archival material, the BBC Prison Study, and the Carnahan and McFarland self-selection work. The story is no longer "ordinary people become guards." The story is "here is how one researcher in 1971 tried to ask that question, here is what really happened, and here is what we have since learned."

For the general reader, perhaps the most useful takeaway is intellectual humility. Vivid demonstrations do not automatically make for valid science. Famous studies deserve the same scrutiny as obscure ones. And the human tendency to behave badly in coercive institutions is real — but the path from that observation to a controlled experimental claim is harder than the most famous social-psychology study of the twentieth century made it look.