Zimbardo and his team aimed to test the hypothesis that the inherent personality traits of prisoners and guards are the chief cause of abusive behavior in prison. Participants were recruited and told they would participate in a two-week prison simulation. Out of 75 respondents, Zimbardo and his team selected the 24 males whom they deemed to be the most psychologically stable and healthy.[7] These participants were predominantly middle-class.[8] The group was intentionally selected to exclude those with criminal background, psychological impairments or medical problems. They all agreed to participate in a 7–14-day period and received $15 per day (equivalent to $87 in 2015).

The experiment was conducted in the basement of Jordan Hall (Stanford's psychology building). Twelve of the twenty-four participants were assigned the role of prisoner (nine plus three alternates), while the other twelve were assigned the role of guard (also nine plus three alternates). Zimbardo took on the role of the superintendent, and an undergraduate research assistant the role of the warden. Zimbardo designed the experiment in order to induce disorientation, depersonalization and deindividualization in the participants.

The researchers held an orientation session for guards the day before the experiment, during which they instructed them not to physically harm the prisoners. In the footage of the study, Zimbardo can be seen talking to the guards: "You can create in the prisoners feelings of boredom, a sense of fear to some degree, you can create a notion of arbitrariness that their life is totally controlled by us, by the system, you, me, and they'll have no privacy ... We're going to take away their individuality in various ways. In general what all this leads to is a sense of powerlessness. That is, in this situation we'll have all the power and they'll have none."[9]

The researchers provided the guards with wooden batons to establish their status,[10] clothing similar to that of an actual prison guard (khaki shirt and pants from a local military surplus store), and mirrored sunglasses to prevent eye contact. Prisoners wore uncomfortable ill-fitting smocks and stocking caps, as well as a chain around one ankle. Guards were instructed to call prisoners by their assigned numbers, sewn on their uniforms, instead of by name.

The prisoners were "arrested" at their homes and "charged" with armed robbery. The local Palo Alto police department assisted Zimbardo with the arrests and conducted full booking procedures on the prisoners, which included fingerprinting and taking mug shots. They were transported to the mock prison from the police station, where they were strip searched and given their new identities.

The small mock prison cells were set up to hold three prisoners each. There was a small space for the prison yard, solitary confinement, and a bigger room across from the prisoners for the guards and warden. The prisoners were to stay in their cells all day and night until the end of the study. The guards worked in teams of three for eight-hour shifts. The guards did not have to stay on site after their shift.

After a relatively uneventful first day, on the second day the prisoners in Cell 1 blockaded their cell door with their beds and took off their stocking caps, refusing to come out or follow the guards' instructions. Guards from other shifts volunteered to work extra hours to assist in subduing the revolt, and subsequently attacked the prisoners with fire extinguishers without being supervised by the research staff. Finding that handling nine cell mates with only three guards per shift was challenging, one of the guards suggested that they use psychological tactics to control them. They set up a "privilege cell" in which prisoners who were not involved in the riot were treated with special rewards, such as higher quality meals. The "privileged" inmates chose not to eat the meal in order to stay uniform with their fellow prisoners. After only 36 hours, one prisoner began to act "crazy", as Zimbardo described: "#8612 then began to act crazy, to scream, to curse, to go into a rage that seemed out of control. It took quite a while before we became convinced that he was really suffering and that we had to release him."

Guards forced the prisoners to repeat their assigned numbers[11] to reinforce the idea that this was their new identity. Guards soon used these prisoner counts to harass the prisoners, using physical punishment such as protracted exercise for errors in the prisoner count. Sanitary conditions declined rapidly, exacerbated by the guards' refusal to allow some prisoners to urinate or defecate anywhere but in a bucket placed in their cell. As punishment, the guards would not let the prisoners empty the sanitation bucket. Mattresses were a valued item in the prison, so the guards would punish prisoners by removing their mattresses, leaving them to sleep on concrete. Some prisoners were forced to be naked as a method of degradation. Several guards became increasingly cruel as the experiment continued; experimenters reported that approximately one-third of the guards exhibited genuine sadistic tendencies. Most of the guards were upset when the experiment concluded after only six days.

Zimbardo mentions his own absorption in the experiment. On the fourth day, some of the guards stated that they heard a rumor that the released prisoner was going to come back with his friends and free the remaining inmates. Zimbardo and the guards disassembled the prison and moved it onto a different floor of the building. Zimbardo himself waited in the basement, in case the released prisoner showed up, and planned to tell him that the experiment had been terminated. The released prisoner never returned, and the prison was rebuilt in the basement once again.

Zimbardo argued that the prisoners had internalized their roles, since some had stated that they would accept "parole" even if it would mean forfeiting their pay, despite the fact that quitting would have achieved the same result without the delay involved in waiting for their parole requests to be granted or denied.[12] Zimbardo argued they had no reason for continued participation in the experiment after having lost all monetary compensation, yet they did, because they had internalized the prisoner identity.

Prisoner No. 416, a newly admitted stand-by prisoner, expressed concern over the treatment of the other prisoners. The guards responded with more abuse. When he refused to eat his sausages, saying he was on a hunger strike, guards confined him to "solitary confinement", a dark closet: "the guards then instructed the other prisoners to repeatedly punch on the door while shouting at 416."[13] The guards stated that he would be released from solitary confinement only if the prisoners gave up their blankets and slept on their bare mattresses, which all but one refused to do.

Zimbardo aborted the experiment early when Christina Maslach, a graduate student in psychology whom he was dating (and later married),[14] objected to the conditions of the prison after she was introduced to the experiment to conduct interviews. Zimbardo noted that, of more than fifty people who had observed the experiment, Maslach was the only one who questioned its morality. After only six days of a planned two weeks' duration, the Stanford prison experiment was discontinued.[12]