The Stanford Prison Experiment Explained

In August 1971, Stanford psychologist Philip Zimbardo turned a basement hallway into a mock prison, randomly assigned 24 male college students to be "guards" or "prisoners," and planned to watch for two weeks. He shut it down after six days, when guards had escalated to psychological abuse and at least one prisoner suffered an acute breakdown. For decades the study was taught as proof that ordinary people turn cruel once given a role and a uniform. A 2018 investigation, drawing on Zimbardo's own archived audio, found that the guards were actively coached toward cruelty, not simply an obedient consequence of the situation itself.

What the Stanford Prison Experiment actually tested

The Stanford Prison Experiment (SPE) is a study Philip Zimbardo ran at Stanford University in August 1971 to test whether the brutality documented in American prisons came from the personalities of guards and inmates, or from the roles and institutional structure those people were placed into (Haney, Banks & Zimbardo, 1973). Zimbardo recruited 70 respondents through a newspaper ad offering $15 a day for a "psychological study of prison life," screened them for psychological stability and no history of crime or drug use, and randomly assigned 24 of the healthiest, most stable applicants to play either "guard" or "prisoner" by a coin flip (Haney, Banks & Zimbardo, 1973).

The setup was deliberately theatrical. Zimbardo converted the basement of Stanford's psychology building into a makeshift jail, complete with barred cells, a "solitary confinement" closet, and a numbering system that replaced each prisoner's name with an ID number stenciled on a smock. Prisoners were arrested at their homes by real Palo Alto police officers as part of the simulation, fingerprinted, stripped, deloused, and issued a chain around one ankle. Guards wore mirrored sunglasses, carried wooden batons, and were given no formal training beyond an instruction from Zimbardo, who also served as the "prison superintendent," that they should maintain "law and order" without physical violence (Haney, Banks & Zimbardo, 1973).

Zimbardo's stated hypothesis, later summarized in his popular writing as the power of the situation, held that institutional roles alone, independent of any guard's or prisoner's individual disposition, would push ordinary, psychologically healthy young men toward domination on one side and helplessness on the other (Zimbardo, 2007). That claim, more than any specific data point from the study, is what made the SPE one of the most cited findings in introductory psychology for the next 40 years.

Zimbardo planned the Stanford Prison Experiment to run for two weeks. He ended it after six days, on day six of a planned fourteen, after his graduate student and future wife, Christina Maslach, objected to what she saw on a walk-through of the mock prison (Zimbardo, 2007).

How the six-day simulation unfolded

The timeline matters because most popular retellings compress six days of escalating events into a single dramatic image, guards abusing prisoners, without showing how quickly the situation moved or what specifically happened at each stage. The table below reconstructs the documented sequence from Zimbardo's own published account and subsequent historical review.

Day	Key events
Day 1 (Sunday)	"Prisoners" arrested at home by real police, booked, strip-searched, deloused, and issued numbered smocks; guards issued uniforms, sunglasses, and batons with minimal instructions
Day 2 (Monday)	Prisoners stage a rebellion, barricading their cell doors with beds; guards break the rebellion using fire extinguishers and begin isolating instigators in solitary confinement
Day 3–4	Guards escalate control tactics: forced exercise, sleep disruption, degrading tasks; one prisoner (identified as #8612) has an acute stress reaction, crying and screaming, and is released
Day 5	A rumor of a planned mass escape triggers extra guard shifts and heightened surveillance; a prisoner develops a psychosomatic rash across his body and is released
Day 6	Christina Maslach, brought in to conduct interviews, objects to the conditions; Zimbardo ends the study roughly eight days ahead of its planned two-week run

What was supposed to be a two-week study of institutional roles ended after six days, once a researcher outside the "prison" frame objected to what the simulation had become.

By Zimbardo's own account, five of the ten prisoners were released early due to what he described as extreme emotional distress, and the guards, working in three eight-hour shifts, grew progressively harsher as the days went on, with the most aggressive behavior concentrated in the night shift, whose guard the other participants nicknamed "John Wayne" (Haney, Banks & Zimbardo, 1973; Zimbardo, 2007). Zimbardo has since acknowledged that his own dual role, as both principal investigator and the prison's "superintendent" character, compromised his objectivity while the study was running, since he was simultaneously observing the situation and actively shaping it (Zimbardo, 2007).

The breakdown that ended the study

The single most quoted event from the SPE is the psychological collapse of prisoner #8612 on day 2. According to Zimbardo's account, #8612 began experiencing what the research team described as acute anxiety, uncontrollable crying, and disorganized thinking after roughly 36 hours in the mock prison, and asked to be released (Haney, Banks & Zimbardo, 1973). Zimbardo's team's initial response was to treat the request skeptically, worried he was faking distress to be let out, an interpretation Zimbardo later called one of the study's most troubling moments, since it showed how completely he and his staff had absorbed the guard-versus-prisoner framing themselves (Zimbardo, 2007).

Christina Maslach's intervention on day 6 is the event most consistently credited with ending the study. Maslach, then a recent Stanford PhD and Zimbardo's romantic partner, was brought in to help conduct interviews with participants. Watching guards march a line of hooded, shackled prisoners to the bathroom, she reportedly told Zimbardo she was disgusted by what the study had become, one of roughly 50 outside visitors to see the mock prison in six days, and the first to object (Zimbardo, 2007). Zimbardo has written that her reaction forced him to confront how deeply he had lost the outside perspective a researcher needs to maintain, and he ended the study within a day of that conversation, at day 6 of a planned 14 (Zimbardo, 2007).

The Stanford Prison Experiment has never been formally replicated with Zimbardo's exact design. Modern institutional review boards would not approve a study involving simulated arrest, deception about duration, and documented psychological distress without a clear, unpressured right to withdraw.

Could a researcher run the Stanford Prison Experiment today?

No. The study predates the modern research-ethics review system that grew out of cases like this one and Milgram's obedience studies. A design combining simulated arrest, sustained psychological distress, and a dual role for the investigator as both scientist and prison authority would not clear a contemporary Institutional Review Board (IRB) in the United States.

The guard coaching controversy: what a 2018 investigation found

For roughly four decades, the SPE's dominant interpretation, that ordinary people turned cruel simply because a prison-like situation and an assigned role pushed them there, went largely unchallenged in psychology textbooks. That changed substantially after journalist Ben Blum published a widely read investigation in Medium in June 2018, drawing on archived audio recordings and interviews with former participants held at Stanford's Special Collections library (Blum, 2018).

Blum's reporting centered on the study's "warden," undergraduate research assistant David Jaffe, whose voice appears on archived tape explicitly directing a guard to be tougher. In one recording Blum obtained and published, Jaffe tells a guard identified as Vandy that the guards need to be firmer because the study "isn't working" the way the team wanted, coaching that undercuts the idea that guard cruelty emerged spontaneously from the situation alone (Blum, 2018). Blum also interviewed Douglas Korpi, the participant known as prisoner #8612 whose breakdown became the study's signature moment, who told Blum in 2018 that his distress was substantially exaggerated for effect and driven by wanting out to prepare for a graduate school exam, not by a spontaneous psychological collapse purely caused by the prison conditions (Blum, 2018).

Ben Blum's 2018 investigation found archived audio of undergraduate "warden" David Jaffe coaching a guard to act tougher, and quoted "prisoner #8612" Douglas Korpi describing his famous on-camera breakdown as substantially performed rather than a spontaneous psychological collapse (Blum, 2018).

These revelations reframed a core question about the study: if guards were directed toward cruelty by the research team rather than discovering it spontaneously through the power of the role, the SPE demonstrates something closer to demand characteristics, participants behaving as they believe the experimenter wants, than a pure, unprompted "power of the situation" effect (Blum, 2018; Le Texier, 2019). Zimbardo has publicly disputed elements of Blum's account, arguing that Jaffe's coaching represented a single research assistant's initiative rather than a scripted directive from Zimbardo himself, and that the broader pattern of guard behavior across six days went well beyond anything one assistant could have manufactured (prisonexp.org, Zimbardo's response, n.d.).

The peer-reviewed case against the Stanford Prison Experiment

Blum's journalism was followed the next year by a more formal scholarly challenge. French historian of science Thibault Le Texier spent years examining Zimbardo's original archives, including his unpublished daily logs, guard shift notes, and recordings, for a 2018 book later summarized in a peer-reviewed 2019 article in American Psychologist (Le Texier, 2019). Le Texier's central argument went beyond guard coaching: he documented that Zimbardo's own contemporaneous records showed most guards did not behave cruelly, some actively tried to help prisoners, and the sharpest abuses were concentrated in a small number of guards on one shift, a pattern closer to ordinary variation in how people respond to a role than a uniform demonstration that the situation alone corrupts nearly everyone placed in it (Le Texier, 2019).

Le Texier also raised a methodological point that is easy to miss in popular retellings: the SPE had no control group, no random assignment to different experimental conditions beyond the guard/prisoner coin flip, and only one instance of the "prison" ever being run, meaning its central claim was built on a single, uncontrolled case study rather than a comparison across independently repeated conditions (Le Texier, 2019). That is a different standard of evidence than most findings taught alongside it in a psychology curriculum, and it is a substantial part of why professional bodies quietly reduced the SPE's prominence in newer textbook editions after 2018 and 2019.

Does this mean the Stanford Prison Experiment is worthless?

Not entirely, but its evidentiary weight is far lower than its fame suggests. The SPE still documents a real, troubling instance of institutional roles enabling abusive behavior in a specific six-day case, and it remains historically important for the ethics reforms it helped inspire. It should not be read as proof that most ordinary people spontaneously become cruel once given power over others, a stronger claim than the study's own uncontrolled, single-run design and the 2018 coaching revelations can support.

The BBC Prison Study: what happened when the coaching was removed

One of the most useful pieces of evidence for evaluating the SPE's central claim comes from a partial attempt to test it under different conditions. In 2002, psychologists Alex Haslam and Stephen Reicher, working with the BBC, ran the BBC Prison Study, a televised, ethically reviewed reconstruction that randomly assigned participants to guard and prisoner roles inside a purpose-built facility, but deliberately avoided giving guards leadership training or directive coaching toward harsh behavior (Reicher & Haslam, 2006).

The result diverged sharply from the SPE. Guards in the BBC study did not spontaneously develop a shared, aggressive group identity, and several actively resisted using their authority forcefully; it was the prisoners, not the guards, who eventually organized an effective coordinated group identity and challenge to the guards' authority, the opposite dynamic from Zimbardo's account (Reicher & Haslam, 2006). Haslam and Reicher's own interpretation, which they call the social identity model of tyranny, argues that group behavior in role-based situations depends heavily on whether participants are actively encouraged to identify with and act on a group identity, not on the mere existence of a "guard" role and a uniform (Reicher & Haslam, 2006). Under this reading, the SPE's outcome looks less like an automatic consequence of any prison-like situation and more like a product of the specific, directive way Zimbardo's research team ran that particular six-day study.

Why the study still gets taught despite the controversy

The SPE remains in nearly every introductory psychology course for reasons that have little to do with whether its central causal claim holds up under current scrutiny. It functions as a case study in three separate lessons that are each independently valid regardless of the coaching controversy: how research-ethics review came to exist, how role and authority structures can create real institutional harm even in a short, artificial setting, and how a scientist's dual role as both observer and active participant, as Zimbardo was as prison "superintendent," can compromise a study's objectivity in ways that are hard to see from inside the study itself.

The SPE, together with Stanley Milgram's obedience experiments conducted a decade earlier, became one of the standard teaching examples cited when the U.S. government built its modern human-subjects protection system, including the Belmont Report and the requirement that federally funded institutions maintain an Institutional Review Board (U.S. Department of Health and Human Services, 1979). Both studies exposed participants to real psychological distress under a research design that offered limited, socially pressured ability to withdraw, and both are now taught as much for what they reveal about the limits of unregulated human-subjects research as for their original psychological claims.

How does the Stanford Prison Experiment differ from the Milgram experiment?

Milgram tested obedience to a direct order from a present authority figure, delivered moment to moment through scripted verbal prods. The SPE tested how an assigned social role, guard or prisoner, shapes behavior over an extended, less directly supervised period, though the 2018 and 2019 revelations show Zimbardo's team gave guards more explicit direction than the original "power of the situation" framing acknowledged. For a full breakdown of Milgram's design and its own replication history, see the Milgram experiment explained.

A related but distinct line of research examines what happens when no authority figure or assigned role is present at all. The bystander effect describes how the mere presence of other witnesses, without any guard, prisoner, or experimenter role assigned to anyone, can lower the odds that any one person intervenes in an emergency. Reading the SPE, Milgram's obedience studies, and bystander research together shows three different mechanisms, coached role behavior, direct authority, and diffused responsibility among peers, that each independently shape whether ordinary people act to help or to harm.

Is the Stanford Prison Experiment considered a real scientific experiment?

Critics including Le Texier (2019) argue it functions more like an uncontrolled demonstration or case study than a rigorous experiment, since it had no control condition, ran only once, and involved a research team, per the 2018 audio evidence, actively directing some of the behavior it claimed to be passively observing. Zimbardo and defenders of the original interpretation maintain that the coaching was limited and does not erase the broader pattern of guard escalation documented across six days.

What this means

The Stanford Prison Experiment is best understood today as two separate stories layered on top of each other. The first is Zimbardo's original claim that institutional roles alone can push psychologically healthy people toward cruelty or collapse within days, a claim that real events during those six days in August 1971 partly support. The second, added by Ben Blum's 2018 reporting and Thibault Le Texier's peer-reviewed 2019 analysis, is that Zimbardo's research team actively coached guards toward the harsh behavior the study became famous for, and that the study's most quoted moment of prisoner distress was later described by the participant himself as substantially performed. Both stories can be true at once: a real, troubling six-day event that happened inside an artificial prison, and a research design too compromised, uncontrolled, and directive to support the sweeping claim about human nature it has been used to illustrate for 50 years. Read the study for what it documents about role, authority, and the limits of unregulated human research, and treat its "ordinary people spontaneously turn cruel" headline with the same skepticism the 2018 archive review earned it.

FAQ

Q: How long was the Stanford Prison Experiment supposed to last? Zimbardo planned a two-week study. He ended it after six days, on day 6, after his colleague Christina Maslach objected to the conditions she observed during a walk-through interview session (Zimbardo, 2007).

Q: Were the guards in the Stanford Prison Experiment told to be cruel? Archived audio published by journalist Ben Blum in 2018 shows undergraduate "warden" David Jaffe coaching a guard to act tougher because the study "wasn't working." This undercuts the original claim that guard cruelty emerged spontaneously from the prison role alone, though Zimbardo has disputed the extent of that coaching's influence (Blum, 2018; prisonexp.org, n.d.).

Q: Was prisoner #8612's breakdown real? Zimbardo's team documented visible acute distress and released him after roughly 36 hours. In a 2018 interview with Ben Blum, the participant himself, Douglas Korpi, said his distress was substantially exaggerated to secure his release rather than a purely spontaneous psychological collapse (Blum, 2018).

Q: Has the Stanford Prison Experiment ever been replicated? Not with Zimbardo's exact design; current ethical standards would not permit it. The closest comparable attempt, the 2002 BBC Prison Study by Alex Haslam and Stephen Reicher, deliberately avoided directive guard coaching and found guards did not spontaneously develop the same aggressive group dynamic Zimbardo reported (Reicher & Haslam, 2006).

Q: Why do textbooks still teach the Stanford Prison Experiment if it's controversial? It remains a standard case study in the history of research ethics, alongside Milgram's obedience studies, because both exposed participants to real psychological distress and directly influenced the creation of Institutional Review Boards and the Belmont Report (U.S. Department of Health and Human Services, 1979). Many newer editions now teach it alongside the 2018–2019 criticisms rather than as settled evidence about human nature.

Q: What is the main scientific criticism of the Stanford Prison Experiment? The strongest peer-reviewed critique, from historian Thibault Le Texier's 2019 American Psychologist article, is that the study had no control group, ran only once, and relied on a research team that actively directed some of the guard behavior it claimed to be neutrally observing, making its evidentiary weight far lower than a single uncontrolled six-day case study can support (Le Texier, 2019).

Sources: Zimbardo's response to recent criticisms, Ben Blum, "The Lifespan of a Lie," Medium, 2018

---

Read next:

• The Milgram Experiment Explained
• The Asch Conformity Experiment Explained
• The Bystander Effect Explained