One of the most infamous psychological studies ever conducted was the Stanford Prison Experiment. It's mentioned in almost every intro to psychology textbook. They tend to focus on how unethical it was, and are less critical of its supposed conclusion.

August 14th, 1971. Palo Alto, California. Twelve young men are rounded up from their homes by police, placed under arrest, and brought to a makeshift prison in the basement of Stanford University.

It all begins as a study on the psychology of prison life, led by Stanford psychology professor Dr Philip Zimbardo. 24 volunteers-- 12 guards and 12 prisoners. --have agreed to spend the next two weeks recreating life in a correctional facility.

[guard] The prisoners are booked and stripped nude. They're no longer individuals, forced to wear smocks, stocking caps and shackles. Identified only by their prisoner numbers.

The guards quickly adapt to their new profession. Given anonymity by their mirrored sunglasses, some of them start to control the meager food rations, restrict prisoners' bathroom use. And, as tensions rise, so do their cruel methods.

Within just six days of the planned two-week study, conditions are so bad that the entire operation is shut down. [man] Goddamn it. .

. The study makes international headlines. Zimbardo's fame skyrockets, and his conclusions are taught to students worldwide, used as a defense in criminal trials and are even submitted to Congress to explain the abuses inflicted at Abu Ghraib.

The study brings up a question just as important then as it is today: is evil caused by the environment, or the personalities in it? Zimbardo's shocking conclusion is that when people feel anonymous and have power over depersonalized others, they can easily become evil. And it occurs more often than we'd like to admit.

But while it's true that people were mean to each other during the Stanford Prison Experiment, what if what truly caused that behavior wasn't what we've always been told? The Stanford Prison Experiment has always had its controversies. But a wave of recent revelations have pushed it back into the spotlight 47 years later.

Today, I'm going to speak with journalist Ben Blum, whose recent writings have brought criticism of the experiment to a larger audience than ever before. How did you get involved in the Stanford Prison Experiment in the first place? Well, my involvement was quite personal.

Like everyone, I had kind of absorbed the basic lesson of the experiment through the cultural ether. And then my cousin Alex was arrested for bank robbery. This was a team of mostly military guys with AK-47s.

Alex was the driver. He was a 19-year-old U. S.

Army Ranger. And it was a superior of his on the Rangers that organized and led the bank robbery. Alex thought the whole thing was a training exercise.

He was just so brainwashed in this intense Ranger training that when a superior proposed this bank robbery, he took it as just one more kind of tactical thought experiment. Then Dr Philip Zimbardo participated in his legal defense. Zimbardo submits a letter to the court, advocating leniency in sentencing on the grounds that Alex, my cousin, had been so transformed by the social environment of the Ranger battalion that he participated in the bank robbery without exercising his own free will.

Well, how did that affect Alex's sentencing? He received an extraordinarily lenient sentence of 16 months. So Zimbardo was a family hero.

But over time, Alex, finally he did admit to me, you know what, I knew this was a bank robbery by the end, and I just didn't have the moral courage to back out. Oh, wow. Alex, myself and our whole family came to view the Zimbardo argument as a way to shirk personal culpability, and to put all the blame on the situation.

So you start looking at the Stanford Prison Experiment in particular. You reached out to Dr Zimbardo himself, as well as some of those who participated. What did you learn?

I learned, to my deep surprise, that quite a number of the participants had stories of their experience that completely contradicted the official narrative. Which is, look, these regular people, good people, came together, and because of the situation, became evil. [Ben] Right.

Zimbardo has claimed that the guards were put in the situation, and then the kind of hidden wellspring of sadism that apparently lies in all of us unfolded organically. [Zimbardo] There was an orientation meeting for the guards. They had been told quite explicitly to oppress the prisoners.

That falls under the heading of what psychologists call demand characteristics. Experimental subjects tend to be motivated to give experimenters what they want. [Michael] Demand characteristics occur whenever participants being studied act differently than they normally would because they've guessed what hypothesis is being tested and feel that a certain kind of behavior is being demanded.

There was a recording of explicitly correcting a guard who wasn't being tough enough. So a conclusion you could make from the Stanford Prison Experiment is that when you tell people to be cruel, they'll do it if you tell them it's for a greater good, like science. -Right.

-Who would have thought? I think the study stands still as a fascinating spur to further more careful research as a demonstration that should make anyone curious as to how such extreme behavior could arise in such a short time. The experiment could still be useful, but it might need to be reinterpreted.

Its data might lead to different conclusions than the one that we've been telling for so many decades. Right. The flaws in the experiment that Ben and other critics bring up call into question large portions of the narrative surrounding the study.

So I want to hear from someone who was actually there. Dave Eshelman, the study's most infamous guard, agreed to tell me his side of the story. It's really an honor to meet you.

You're a living, walking piece of psychology history. I'm never recognized in the street or anything like that, although I still get some hate mail. -Are you serious?

-Yeah, absolutely. Well, what do you say to them when they react that way? I say, well, there's probably a lot about that that didn't happen quite the way it's been portrayed.

Well, Dave, before we go too far, I'd like to watch the footage we have here so we can kind of talk about what we see. [Dave] That's me there, by the way. -[Michael] Look at that look.

-[Dave] Mm-hmm. So how did you get involved with a Stanford Prison Experiment? My father was a professor at Stanford, and I was home for summer, looking for a summer job.

So I'm looking through the want ads. $15 a day. You know, in 1971 that wasn't bad.

The way it was introduced to the guards, the whole concept of this experiment, we were never led to believe that we were part of the experiment. We were led to believe that our job was to get results from the prisoners, that they were the ones the researchers are really studying. The researchers were behind the wall.

And we all knew they were filming. And we can often hear the researchers commenting on the action from the other side of the wall. You know, like, "Oh, gosh, did you see that?

Here. Make sure you get a close-up of that. " Okay?

So if they want to show that prison is a bad experience, I'm going to make it bad. But how did you feel doing stuff like that? Didn't you feel bad?

I don't know if this is a revelation to you, but 18-year-old boys are not the most sensitive creatures. -Sure. -My agenda was to be the worst guard I could possibly be.

-And it's pretty serious. -Mm-hmm. This is my favorite part of all the footage we have -from the experiment.

-Mm-hmm. It's you and a prisoner confronting each other after the experiment. I remember the guy saying, "I hate you, man.

" -Yeah. -"I hate you. " Each day I said, well, what can we do to ramp up what we did yesterday?

How can we build on that? Why did you want to ramp things up? Two reasons, I think.

One was because I really believed I was helping the researchers with some better understanding of human behavior. On the other hand, it was personally interesting to me. You know, I cannot say that I did not enjoy what I was doing.

Maybe, you know, having so much power over these poor, defenseless prisoners, you know, maybe you kind of get off on that a little bit. You weren't entirely following a script from a director. Right.

But you also felt like Zimbardo wanted something from you. -Yes. -And you gave that to him.

I believe I did. I think I decided I was going to do a better job than anybody there of delivering what he wanted. But does that excuse me from what I was doing?

Certainly it started out with me playing a role. So the question is, was there a point where I stopped acting and I started living, so to speak? The standard narrative is that Dave Eshelman did what he did because when people are given power, it's easier than we think for abuse to happen.

That may be true, but how predisposed to aggression was Dave? I mean, he signed up to something called a "prison study," after all. Also, his feeling that cruelty was encouraged and helped the experiment, may have affected his behavior.

What I'd like to see is, in the absence of outside influence, can anonymity, power, and depersonalization alone lead to evil? To answer that question, I'd like to design a demonstration of my own. So I'm meeting with Dr Jared Bartels of William Jewell College, a psychologist who has written extensively about the Stanford Prison Experiment and how it is taught.

I would love to do the Stanford Prison Experiment again. You could probably make it more ethical, but still find the same conclusions. That's my hypothesis.

I absolutely think it's worthwhile. It's important. It's interesting.

Probably the best approach is eliminate as best as possible the demand characteristics by eliminating that prisoner/guard dynamic. Why do we even need to call one group "guards" and one "prisoners"? There's a lot of expectations around those roles.

Oh, I'm a guard? -I guess I should act like a guard. -Yeah, you're right.

The cover story is really important, and you want to hide the true purpose of the experiment. Another piece of this is the role of personality and personality traits. So the original ad in the Stanford study asked for participants for a study of prison life.

You know, that's going to draw certain people that were more kind of disposed to aggression. [Michael] Because they saw the word "prison" and thought, -"I want to be a part of that. " -Exactly.

So when you get a group of kind of authoritarian-minded individuals together, not surprisingly they're going to create an authoritarian regime and environment. So, for whatever it is that we're going to do, we should evaluate the personalities of the individuals. Right.

So how do we give people every opportunity to be as evil as they can? I think you have to have those elements that were assumed to be influential in the Stanford study. What are those elements?

You have to have the depersonalization. You have to have anonymity. You have to have some power differences.

Can we elicit some surprising behaviors in just a number of hours? If you kind of come back to the Stanford study, there wasn't anything dramatic that happened -in the first day of the study. -Yeah.

It was the second day of the study when the guards started to assert their authority. That came about because of prisoners testing and challenging the guards' authority. [Michael] Yeah, and that led to fear.

That, like, wait a second, these prisoners need to be -put more in check. -Yeah. Yeah.

So I think you still need that provocation. Yeah. Something that is frustrating.

Something that's going to increase the participants' arousal. Right. All right, so, Jared, would you like to spend some time now brainstorming a new design that peeks into the same questions?

-Absolutely. -Awesome. [Michael] Jared and I sat down with the Mind Field crew to begin the planning process.

Will a person, without any expectations or pushes in a certain direction still be abusive or not? For this demonstration, we want to eliminate all outside variables and really isolate the three core elements of the Stanford Prison Experiment. The first element is anonymity.

Subjects need to believe that no matter how they behave, no one will know it was them. This is where people will be coming in in the morning. This way, everyone's going to be staggered when they come in.

That's important, because we don't want them to ever meet their teammates face-to-face. The original experiment gave guards anonymity by providing mirrored sunglasses and uniforms. But we're taking it much further.

Our study will take place in a room that is pitch-black. [Jared] They'll be taken into this room. [Michael] Ah.

I would love to see how dark this room is going to be tomorrow. [man] Yeah, absolutely. -You ready?

-I'm ready. -Oh, yeah. -[man] Right?

[Michael] This is uncomfortable. Despite the darkness, we will be able to see everything, thanks to infrared cameras. The second element is depersonalization.

From the moment the subjects arrive, they will only be identified by number, not name. [woman] So, come on in. To eliminate the demand characteristics, we don't want our subjects to know what we're studying.

Follow the sound of my voice, if you can. All they'll be told is that we are studying how they solve puzzles in the dark. There is another team in a different location.

-who is also solving a puzzle. -Okay. Because the words "guard" and "prisoner" suggest certain expected behaviors, we've done away with them and will simply give our participants an unseen, distantly located opposing team.

We will measure the cruelty predicted by the standard narrative of the Stanford Prison Experiment by giving our participants a way to exercise the third element: power. What I'm going to show you next is the system by which you can send them a loud noise. -Okay.

-So if you want to. . .

We've armed the teams with a "distractor button" that they can press to blast an extremely loud, jarring noise into the other team's room. Everyone will have a volume dial that ranges from level 1 to 12, and they'll be told that anything below a 7 should be safe for the other team's hearing. And each person has their own control.

Okay. So they can't see what you're doing. -You can't see what they're doing.

-Okay. The intensity level they select, as well as the frequency with which they push the button, will be our indicator of how aggressive the participants become in this situation. Is it-- is it pretty, like, terrible to hear?

Well, I'll give you a demonstration. Hey, Derek, could you play level 3 for me? [loud, discordant horn] So that's a 3.

It's pretty. . .

-it's pretty loud. -Yeah. Perfect.

Participants will be told that when they or a member of their team pushes a distractor button, the volume played in the opponent's room will be determined by the highest level selected on any of their teammates' dials. This is to increase the feeling of diffused responsibility. The question is, will any of these participants take advantage of these factors and act sadistically?

Of course, we would never want anyone to actually be harmed in our experiments, so the other team? They don't exist. Instead, Jared and I will be the ones occasionally blasting the group with noise at a safe level, no higher than a 3.

To see just how powerful the situation can be, we selected participants who would not be predisposed to sadism. We screened our participants using the "Big 5 Personality Scale," "The Personality Assessment Inventory," and picked those who scored the highest in "moral" categories, like honesty and conscientiousness. It looks like, you know, they should be able to see each other.

But it's pitch-dark. There are puzzle pieces on the table in front of you. Thank you, and once I leave the room you may begin.

Okay, here we go. [man 1] [man 2] [man 1] I definitely don't think they're conscious of the control panel at this point. -No.

-They're trying to get focused on the task here. [man 1] [man 2] [man 2] [laughter] [man 2] We picked people who were most likely to have these kinds of personalities. [man 1] [laughs] [woman] -Oh.

-She wants. . .

[woman] All right. [all] [man 1] -[high-pitched squeal] -[woman] Did somebody do it already? -I did.

-Yeah. -Okay. -We should retaliate.

-Yeah, retaliate now. [loud, discordant horn] [all laugh] [horn blares] [laughter] [Michael] Now, they're not retaliating against that most recent buzz. Shall we try again?

[loud, discordant horn] Despite the factors making it easy for them to do so, this team doesn't appear to be turning evil. Now they are, like, just deal with it. Just ignore it and keep working together.

They're not interested in retaliating. [discordant horn blares] Over the course of the two-hour study, we blasted them with noise 23 times. [woman laughs] But they only pushed the button six times, and never above a level 5.

They didn't seem to abuse their power. Puzzle pieces down. What would happen if we introduced demand characteristics that encouraged them to act aggressively?

Your team has been randomly assigned an experimental condition. Although the other team will continue working on a puzzle, your team will not. Your only task is to operate the distractors.

Also, the other team's buttons have been disconnected without their knowledge. You will not hear any sounds if they buzz back at you. We introduce the social roles, where there's a little bit of power differential.

We're kind of mimicking the Stanford-like variables here. [Michael] By now saying that the buzzer is their "task," the participants may feel a greater license to use it liberally. Similar to how instructing prison guards in the original experiment to act tough may have encouraged more use of force.

[man 3] [woman] [man 1] Even though they were given instructions to distract the other team, these participants instead just started chatting with one another. They know that they can be distracting now, but they're not pushing the button. No.

[man 2] Oh. Okay. [woman] A couple of threes.

[high-pitched squeal] Over the course of ten minutes, this group only pushed the button three times. Why do you think they're so uninterested in blasting the other team? Because we have individuals who have been selected, really, with that predisposition, right?

These are individuals who shouldn't be interested in retaliating. It was time to debrief the participants on what we were actually studying. [Michael] I'm going to turn the lights on.

Here I am. I'm Michael, and this is Jared. We're going to debrief you on what was really happening today.

There are no other people. You are the only four here at this moment. There was never another team doing anything.

[man 1] This is a study related to the Stanford Prison Experiment. [man 1] The standard narrative we hear about that experiment is that people just become cruel. So, yeah, we're trying to see if we get the nicest people we can, and we give them complete anonymity and the ability to be cruel, but never encourage them to, will they still do it?

And you guys didn't. Did you have any suspicions about what we were studying or what was going on? Right, but I think that's good.

We just want to make sure you don't think that what we're really looking at is how high you turn your own dial. That's really what we're looking at. It was time to bring in our second group of participants, who, like the first group, were screened to be individuals with high morality characteristics.

Anything up to 7 should be safe. [laughs] Yeah. [woman] So once I leave, you can go ahead and get started.

[woman 1] [laughs] Oh. . .

[high-pitched squeal] Right off the bat she went to 7 and pushed the button. Yeah. [loud, discordant horn] [high-pitched squeal] [Michael] Number two's pushing it at a 3.

[discordant horn blares] [woman 1] Okay, here comes number two. [high-pitched squeal] Number two is still at a volume 3. [Michael] This team seemed more willing to retaliate.

Let's see what will happen if we continue buzzing them. Will they escalate their behaviors? Derek, let's blast them again.

Number 3. [loud horn] Okay, let's. .

. All right, so two just pushed at a 3. But she's not touching the dial.

[Jared] She's not. [loud, discordant horn] [woman 2] It's just annoying. [blaring horn] [high-pitched squeal] [all laugh] It was clear that participant number two was really the only one hitting the distractor button, but it appeared that she only did it in retaliation to our buzzes.

So we decided to see what would happen if we laid off. [man 1] It's been probably four or five minutes, and we have not blasted them with the noise, and they haven't played one either. I have a feeling like if we never played a noise in their room, they would never touch the distractor button.

[Jared] Probably not at this point. In the end, we buzzed this group a total of 44 times, and they buzzed us 38 times, 37 of which came from number two but always in retaliation, and never above a 5. All right, guys.

Puzzle pieces down. The situational factors did not seem to be sufficient to make this group sadistic. It was time for phase 2.

[woman 1] Yeah. -Oh, she. .

. -[high-pitch squeal] It looks like it's at 7. -Wow.

-Yeah, she's-- She's going nuts. At a 7. So number three believes there is no other team.

That might explain why she was just going nuts on the button, because she doesn't feel bad about it. [buttons clicking] Okay, they're all pushing the button a lot more. And they were told this time that it was their only task.

[buttons clicking] [all laugh] What a difference this has made. Just like in the Stanford Prison Experiment. If you tell people that they have a certain task to do, they'll do it, even if it's going to mean that they've been broken.

The thing is, they never hit upon what we really cared about, which is turning the dial into an unsafe level. Yeah. [buttons click] [Michael] Hello, everyone.

I'm going to turn the lights on in this room. [woman 1] Okay. -And slowly.

. . -Ah, it hurts.

. . .

you can look. So, hello. -I'm Michael, and this is Jared.

-Hi. I'll give you time to adjust your eyes. Today, you've been part of a study where all we wanted was to see what would happen when we put people in a room and gave them that feeling of anonymity that comes from, well, if I crank my dial up really high, no one will know it's me.

So you have this opportunity to be cruel. I thought I went nuts. Like, when the other person was pressing-- Sure, but that's-- that's just in-kind retribution.

As it turns out, so far, everyone stays in that "below 7 or under" range. -Yeah. -This final phase was us trying to ramp up the demand characteristics.

And I believe number one, right, you did say at one point, "You've broken me. I did it, fine. " So I loved that phrase, because it says "I didn't want to do this, but I'm doing it because I believe it was expected of me.

" [all] Thank you. Thanks. [Michael] After dismissing our participants, Jared and I sat down to discuss our results.

Really fascinating. We brought in people who had very different personalities than those Zimbardo chose. We put them in a situation that did not demand things from them.

And they behaved according to that personality. I think we have some intriguing support for the idea that it's more than just the situation. We really saw personality kind of shine through.

For the most part, they seemed to be aware -of where that line is. . .

-Yeah. . .

. that they shouldn't cross, and they didn't. None of them did.

It was now time to speak with the man himself, Dr Philip Zimbardo, who I worked with on last season's episode, "How to Make a Hero. " Okay. Lisa, Bear, you guys ready?

For years, Dr Zimbardo has responded to criticisms of his famous study, always maintaining that they aren't valid. I asked him about whether his study is better seen as one on the power of demands from authority, but he wasn't receptive to that idea. I then told him about the study we ran to get his reaction.

I wanted to know what the sufficient conditions might be to make anyone do something evil. And we struggled to get that to happen. We couldn't get anyone to be cruel.

Just giving them anonymity, and a dehumanized other, and the power to hurt that other, they didn't take advantage of it. Well, I mean, maybe the problem was, here's a case where, by picking people who were extremely conscientious, extremely mindful, by selecting people who are high on compassion, high on mindfulness, you broke the power of the situation. In the Stanford Prison Experiment, we had, I presume, a relatively normal distribution.

We gave them six personality scales. And we picked people who, in the scales, who were mostly in the mid-range. In that situation, some people behave cruelly, evilly.

Not everybody, but more of the guards than not. So, again, I think that your study is a demonstration of one way in which personality dominates situation. -Ah.

-Where the personalities are-- so I would say it's a positive result. The personalities are special. Where does this balance lie between the personal, the disposition, the personality, and the situation, the environment?

No, that's the big-- that's the ultimate question. Where is, you know, how much of one and how much of the other. .

. ? Right.

Zimbardo insists that demand characteristics played little role in his subject's behavior. Critics like Ben Blum say they played a big role, that what happened was what was asked for. If that's true, then the Stanford Prison Experiment, like the classic Milgram study, still has an important lesson.

People are quick to be cruel if an authority figure suggests that doing so will serve a greater cause. In our test, we made sure that such influences didn't exist. And not one participant acted maliciously.

Personality rose above the situation. Learning how that happens is vital if we want to improve conditions where power is involved. So it's great that this debate is still ongoing.

And look, questioning methods and interpretations is not a personal attack. It's how we improve our confidence in what we know. And that's how science works.

So stay curious, never stop asking questions, and, as always, thanks for watching. Hey, Mind Field. Michael Stevens here.

There is so much more to satisfy your hunger for psychological knowledge right on this show. Click below to check out more episodes.