Subtitles section Play video
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.