Freedom to Learn

The roles of play and curiosity as foundations for learning.

Why Zimbardo’s Prison Experiment Isn’t in My Textbook

One of the questions I'm often asked by professors who teach from my introductory psychology textbook is this: "Why don't you include Zimbardo's classic Stanford Prison Experiment in your book, like all other introductory psychology textbook authors do?" Here's why. Read More

Not Entirely Invalid

While I agree with you that the results of the experiment was inevitable given the permission Zimbardo and others gave the 'guards' to mistreat the 'inmates,' I do think the experiment does have enormous value. I think the experiment just proves something they weren't actually testing for. The experiment shows just how horribly things can go wrong when the societal norms are changed. That there is something truly disturbing in the way humans feel such a need to fit in, that they will commit acts of brutality, merely because they are told to, or because 'everyone else is doing it.' There is a reason why it's usually mentioned in the same context as the Milgram Experiment.

Sure as an example of what goes on in real prisons, it's meaningless, but I think it puts forth evidence of something far more important - the genuine need that we need to encourage social norms that encourage altruistic behaviour over the selfish and cruel. This is why talk of prisons has, to my knowledge, left his avenue of work and focused more on his Heroic Imagination Project.

Isn't it True that Zimbardo

Isn't it true that Zimbardo only discontinued the experiment at the request of his wife who at a certain time, realized that someone needed to intervene, and her husband was not going to on his own accord? She witnessed behaviors in her own husband that she was able to convey to him and influenced the controller of the experiment (her husband) to bring it to an early end. We know what the possibilities of killing ( and other reptilian brain activities ) were likely to have appeared, and yet a more objective witness was necessary. What that means to me is all of these types of experiments should require a third entity, as over site insurance, yes? We need innovative ideas played out to move the discussion and learning forward, so not to discount this amazing willingness to explore this hypothesis, or idea. Zimbardo was brave.

and Milgram?

Thanks for this interesting discussion. I wonder if your textbook included the equally famous Milgram experiments, which were sold as confirming many of the same Hobbesian characteristics supposedly innate in human beings? Or did you drop it as well, due to similar methodological issues? For more on the critique of Milgram, see this interview, for example:


Hi Christopher, I do discuss the Milgram experiments to some extent in my textbook. Milgram was explicitly studying obedience, and the experiments were well designed to do so. As I describe in my textbook, I do think there was value in these experiments. However, I also explain why I don't think the results explain much about such real-world atrocities as the Nazi concentration camps. For one thing, Milgram's subjects, unlike Nazi torturers, didn't have time to think about what they were doing. -Peter

A different reason why the Zimbardo experiments are interesting

Maybe the experiment is interesting but not for the reasons we normally think. As you say, we normally think that the Stanford experiment reveals that any people will act horribly toward each other when put into natural conditions. As you say though, these were not natural conditions but rather, the researchers told particpants how to act. And so maybe, like the Milgram experiments, maybe this experiment mostly showed that the experimental participants would be obedient and agree at all to treat prisoners badly. Isn't that finding interesting in and of itself?

Seems right...

This was what I thought reading your piece. Seems like you're right in a lot of ways, and in fact the conclusion should be that Zimbardo's study demonstrates effectively the same dynamic as Milgram's: people (esp. middle-class Caucasian males) are willing to do nasty things to innocent others when they think that someone in a position of knowledgeable authority wants them to. Still interesting, but the real problem you're highlighting is when that part of the dynamic gets glossed over and this is taken as showing something about the 'natural' dynamics of interaction (ie, that anyone will intrinsically tend toward nasty brutality simply by virtue of themselves being placed in a position of punitive authority).

You make good points about the complexities

You make good points about the complexities involved in such experiments and of the "demand characteristics" playing a substantial role in the outcome. It seems appropriate to acknowledge such questions when discussing the Prison Experiment. Yet despite the concerns, the Prison Experiment may still have something to show us.

Ordinary_People mentions the experiment may have proven something they weren't actually testing for. As with many experiments, there may be something to be gained other than the original intent. Tainted or not, the Prison Experiment has offered me insight into social and interpersonal dynamics witnessed over the last few years; people's unconscious, dysfunctional attitudes and behaviors toward others in situations "outside the norm", or vulnerable.

There are surely other experiments overlapping the social behaviors but the Prison Experiment provides an interesting lens to view certain behaviors when someone has "power over" another. The level of contempt some direct toward assault victims, trauma survivors, persons with disabilities, or even someone simply "grieving" - appears, in many instances, to echo elements of Zimbardo's experiment. As many flaws as it may have, the seemingly extreme behaviors may very well provide insight into dynamics taking place when one person or group considers themselves superior to another. How different is the Prison Experiment from recent high profile cases of young men assuming an unconscious young girl is "fair game", or for that matter - the bystander's support of the aggression?

The last few years I've observed people violating personal boundaries, even in misguided attempts to be helpful. On rare occasions, being "helpful" is a smoke screen for someone's Covert Aggression but often people seem to loose their ability to view someone in a vulnerable situation with dignity and respect. Seeing someone vulnerable seems to trigger something in many people, allowing for a sense of superiority. Maybe it's fear or denial we can (and will) all be vulnerable at some point. Maybe it's a reinforcement of Social Valuation of winning at all costs (to the victor go the spoils), or Just World Hypothesis. The underlying reasons may vary among people, as dramatically as the behaviors. But there seems to be a dynamic compelling some people to reinforce assumptions, biases, and most importantly - their sense of Personal Power, at the expense of another. Social norms are protected, often through punitive social and interpersonal courts of judgment. That's how I see the Prison Experiment being of potential value.

Kudos Dr. Gray for not just

Kudos Dr. Gray for not just following the crowd, and checking each study carefully before putting it in your textbook.

I do think that one of the reason this study has become so iconic is because so many people use the study to come to the conclusion that humans are basically cruel, and will given the right circumstances that cruelty will inevitably come out (thus they can justify their own cruelty).

According to wikipedia “Several guards became increasingly cruel as the experiment continued; experimenters reported that approximately one-third of the guards exhibited genuine sadistic tendencies.”

Does that mean that one-third of people are sadistic if given the chance? Again according to Wikipedia

“Also, it has been argued that selection bias may have played a role in the results. Researchers from Western Kentucky University (Thomas J. Carnahan, PhD and Sam McFarland, PhD) recruited students for a study using an advertisement similar to the one used in the Stanford Prison Experiment, with some ads saying "a psychological study" (the control group), and some with the words "prison life" as originally worded in Dr. Zimbardo's Stanford Prison Experiment. It was found that students who responded to the classified advertisement for the "prison study" were higher in traits such as social dominance, aggression, authoritarianism, etc. and were lower in traits related to empathy and altruism when statistically compared to the control group participants.”

So it seems like the Stanford Prison study provided the opportunity for the students who were already cruel to act cruelly towards the subjects who were unlucky enough to end up as prisoners.

Your point about the students acting out “their stereotyped views of what prisoners and guards do” is also very important. People modeling their behavior based on what they see others doing is very powerful. For example (again from Wikipedia) “Marco Reus said that the study placed undue emphasis on the cruelty of the guards, such as one who was nicknamed "John Wayne", and who said that he caused the escalation of events between guards and prisoners after he began to emulate a character from the Paul Newman film Cool Hand Luke. He further intensified his actions because he was nicknamed "John Wayne", even though he was trying to mimic actor Strother Martin, who had played the role of the sadistic Captain in the movie.”

Obviously the film didn't make this guy cruel, but it did provide a model of a cruel, but entertaining, prison guard. Unfortunately for cruel prison guards and violent prisoners there are plenty of cruel prison guards and violent prisoners stereotypes to emulate.

Another example is many third-world countries have radio or TV soap operas where the appealing characters have few children, and the loser characters have many children. It has been well studied and show that these soap operas lower fertility rates much quicker and cheaper than educating girls.

Could a prison drama where the appealing guards are decent the loser guards are cruel, and where the appealing prisoners are non-violent and working to prepare themselves for their release, while the loser prisoners are violent and cruel, have a positive effect in prisons?

The conclusion

The conclusion isn't really supposed to be that people are basically cruel; rather, the conclusion is supposed to be that, all other things being equal, if people do not have any kind of diagnosable psychopathology, they could act inhumanely toward others due to the circumstances of the situation.

My understanding is that two

My understanding is that two thirds of the prison guards in the Stanford Prison Study did treat the prisoners humanly (and one third inhumanly), so I don't see how the study supports the idea that the circumstances of the study forced the prison guards to be cruel. The Stanford Prison Experiment gave the guards the freedom to mistreat the prisoners, but did not force them to so.

On the other hand, all the humane prison guards did stand by while the inhumane prison guards mistreated the prisoners. They could have quit or put a stop to everything at any time by going to the authorities, but they didn't and that is an important lesson of the study.

In a way the study parallels what is happening in schools with the problem of bullying. They've found that what does work to reduce bullying is if the non-bullying classmates step in to protect the victim. Anti-bullying programs not often include teaching kids how to stop in to stop bullying. Namely role-playing and modeling how to stand up for the victim effectively.

Perhaps if the human prison guards had has this type of training or modeling they might have been more likely to understand the bullying nature of the inhumane prison guards, and also the inhumane nature of the whole experiment, and been more likely to help the abused prisoners.

Only a few

You're right, Katie, there were only a few who were abusive qua guards. The fact remains, if the inhumane treatment did not come about because of essential characteristics or pathologies of the individuals, it must have arisen as a result of situational factors.


Well the fact that only one third of the prison guards acted sadistically points to the cause being “essential characteristics or pathologies”, and not situational, otherwise all the prison guards would have acted sadistically.

The only thing the Stanford Prison Study demonstrated was that if you give people with sadistic tendencies the opportunity and encouragement to act sadistically, they will take that opportunity, and that non-sadistic people are too easily cowed into silence.

I see your point

I see your point. And yes, I suppose you're right. I'll make a further distinction, though. A person could have certain essential characteristics without having a pathology; it's just that the person's personality traits could manifest themselves negatively given a specific situation. For instance, maybe the guy who called himself John Wayne had a natural penchant for aggressiveness. Well, a situation like the SPE could be fertile soil for that aggressive tendency to grow into something else. Basically, I'm agreeing with you, at least in part. It's really a matter of this complex interplay with certain personality traits and an environment in which those personality traits can interact in particular ways.

Yes we are pretty much in

Yes we are pretty much in agreement. But any personality trait taken to an extreme is going to be a pathology. For example if someone is extremely aggressive (in psychology aggressiveness is usually used to mean a willingness to harm others) and extremely low in empathy, then they would be considered a psychopath.

Probably the sadistic prison guards would not be to the extreme where they would be considered psychopaths. But perhaps they were high in the “Dark Triad personality traits of narcissism, psychopathy, and Machiavellism”

If Zimbardo had bothered to do personality tests on the subjects beforehand, the study would have much more value.

The irony

I saw Zimbardo talk (I think at APS) a few years back. His talk was designed to provoke an emotional response, I suppose. It was after his Lucifer book came out and he made sweeping statements (about torture and the Bush administration) without saying much about science at all. My personal reaction was that I found his talk repellent, although I am no fan of Bush or torture. Zimbardo's prior beliefs were primary, evidence was secondary, and contradictory evidence didn't exist.

When Zimbardo's time was up, he made comments to the effect that his message was too important to be constrained to the allotted 50 minute time slot and continued to talk. (Kahneman, who had proceeded him, had stuck to the time limit.) When someone stood up to leave, Zimbardo called him out. I don't remember the details, but I do remember that Zimbardo tried to shame him, and created a situation of immense social pressure on the man to sit back down, with a large crowd watching.

I thought the moment was supremely ironic. Here was Zimbardo, talking about the coercive power of social pressure, and how it should not be abused. And he was using that power to try to make people stay as his talk went over time.

In the Stanford Prison Study. Zimbardo wanted to investigate the power of the situation. So he created a situation specifically designed to make the "guards" into abusive monsters. It is supremely ironic that, in a study about the power of the situation, it was exactly that power that Zimbardo used to create demand characteristics and get the results he wanted.

Irony withstanding

Irony withstanding, do you think that the circumstances of Zimbardo's experiment made people act inhumanely? If so, do you think the conclusion follows, that, put in the right(/wrong) situation, some people will act inhumanely?

Even if there were demand

Even if there were demand characteristics, it's still amazing that the guards were willing to make the inmates' lives so terrible. So in a sense, the study still shows the power of the situation--just not the situation that Zimbardo was talking about. It shows the power of a situation where an experimenter tells subjects to be mean. In that sense it's sort of similar to the Milgram experiments. But Milgram was up front about that fact, whereas Zimbardo was not. Furthermore, given the revelations that have come out since, it's not necessarily safe to assume we know exactly how much the guards were pushed to behave badly.

SPE - Early Version of "Reality" TV

Good article. Valid points.

The claims of ex con, Carlo Prescott, are the most striking to me. He seems to have a bias agenda as far as what he wanted to study to demonstrate; nevertheless, if what he says is true, it definitely brings into question how much manipulation was being performed by the researchers. It suggests what we saw in the experiment was similar to the manipulation we see today on "reality" television.

With that said, I think the study still has value and can be used to introduce the idea of context and environment influencing behavior. This can still be done if you reveal all the pre-experiment/methodical details mentioned in this article.

The pedagogical case for including this study in your textbook

Dr. Gray,

You rightly point out serious limitations of the research, and it is difficult to see how someone could walk away from your article without gaining some insight into how to appropriately interpret research. You end the article by saying that "If I had included Zimbardo’s experiment in my textbook, it would have been to make" the point that "[t]oo often, in our psychological research labs, we trivialize reality." I find that compelling enough of a case to include it.

I am reminded of an article you published in the APS observer in 2008 ( in which you said that

instead of lecturing on Freud, or on Freud’s beliefs, or on definitions of all of the defense mechanisms, I choose one of Freud’s most interesting and still-relevant ideas and lecture on that, presenting the best understanding we have today concerning that idea. In the lecture, I describe Freud’s evidence for this idea, but I would also describe current research that tends to support, refute, or delimit the idea.

It seems to me that including the Zimbardo study along with an abbreviated version of the arguments in this article would do more to support students' development as critical thinkers than omitting it. Given your good arguments in that 2008 article that a central function of PSY 101 should be to promote critical thinking, I am surprised that you wouldn't demonstrate that with the Zimbardo study. I say that I am surprised primarily because of the impact this study has had on public discourse about prisons. If this study has really had the kind of misleading effect that you argue for in this article, getting students to think critically about this study seems to be even more important than getting them to think critically about Freud.

This article is leading me to think that a guided class discussion about this study would be beneficial for my own PSY 101 students. I am specifically considering whether or not it would still be important if it was interpreted as demonstrating the "power of suggestion" rather than the "power of the situation." On the one hand, if Zimbardo had intentionally provided those not-so-subtle suggestions to the guards, it seems that it would mirror the not-so-subtle suggestions the Abu Ghraib guards received. One could make the argument that the study is still informative even without the argument that the guards came up with the abusive tactics. On the other hand, the selection bias issues and other differences with the reality of prisons are still important limitations that would not be weakened by the change in interpretation. I am fairly certain that students would emerge from this discussion with the same kind of skepticism of the Zimbardo study you advocate, but would also gain an improved ability to think critically about one of the most famous studies in the history of psychology.

Critical thinking in Psych 101

Eric, thank you for this thoughtful comment. In fact, when I was teaching the introductory psychology course I did generally present the Zimbardo study in lecture and asked students to critique it from a methodological point of view. I found that many were able to come up with the role-playing critique on their own. Sometimes I did this as part of the discussion of methodology, early in the course; sometimes I did later, when we were into social psychology. I don't remember for sure, but I think I even suggested this to others who use my book, in the Instructors' Manual. -Peter

hmm i find that i agree with

hmm i find that i agree with the author's critique... but i still think that Zimbardo's Prison Experiment still has it's meaning too...

i believe i experienced a similar prison experiment...

at first i tried to stand my ground and defend my identity...

but after around a year of drills, i noticed that i succumbed to their will...

every statement, every decisions i made will be directly denied deviantly, with no reasons, no explanations...

and the next step was, a set of games will be set up directly where the statement, etc. will be broken... the process will be just like the description in the brainwash method...

at first it's nothing, i can hold my ground, as the manipulations are still easy to identify... but after a whole year of such procedure repeated again and again... you just gets tired of the whole childish games... and things gets blurred... you don't really know what is right and wrong anymore... after all everything you said has been denied consistently...

that it's better to fulfill their wishes from the start on and act and think just like how they want you to be... so you can move on with your life... but that is also the time where you get attacked... this is where your responsibility will be questioned...

at the moment, i am thinking about, if Chinese culture revolution can be seen as Zimbardo's Prison Experiment in mass scale...

if we alter/tone down the concept of "prison" a bit... a relationship can be experienced as a prison too, just as a family, a school, an institution, and even the whole society... etc...

depends in which "prison" we are, we are supposed to act in certain way that conform with certain set of norm...

this is also where things gets abused... especially when private and public matters, official and inofficial matters, and all sort of things gets mixed together...

Zimbardo's prison experiment

I agree Peter that the experimenter and his subjects were playing a role. With the exception of those who pulled out, they had become deluded when they mistook play for reality. The problem seems to be related to Phenomenology; is "A philosophy or method of inquiry based on the premise that reality consists of objects and events as they are perceived or understood in human consciousness and not of anything independent of human consciousness." I am no philosopher but I smell a rat here. Reality is based on reality and just because we can observe a phenomenon does not make it real. In play we are observing an illusion and that by definition means that it is not real.

Experiments in social psychology

Reading your article made me think about experiments in social psychology generally. The most famous experiments are Zimbardo's Stanford Prison Experiment (SPE) and the Milgram Shock Experiment (MSE). The experiments ostensibly provide putative scientific support for intuitions that most people have. With regard SPE, the conclusion is that, all other things being equal, some people will act violently, inhumanely, or commit acts of evil given certain kinds of bad conditions. Perhaps Zimbardo's experiment doesn't prove that, per se, but maybe that's because it is not even clear what it would like like for that hypothesis to be wrong.

You wrote that Zimbardo basically told those students in the role of guards what to do--minimally, he implied that they qua guards should use coercive measures, demeaning language, and a dusting of 'light' torture techniques, which you identified as demand characteristics. And in effect, it appears as though from some students qua guards and some students qua behaviors he got the sorts of reactions he had asked for or implied he was looking for. Taking the demand characteristics into account, wouldn't he have gotten the conclusion he was looking for, anyway, because the conclusion is imprecisely stated? That is, if the form in which I stated it above is essentially the conclusion that Zimbardo was looking for, what would count as verification could be situations in which authority figures express demand characteristics for outcomes and some people comply, and therefore act inhumanely toward other people. Couldn't Zimbardo argue, "Yes, I expressed what I was looking for from the study, but these relatively normal people complied, and isn't that shocking, giving evidence to the fact that relatively normal people will comply with the wishes of authority figures under the right circumstances, no matter how immoral the actions might be"? If Zimbardo were looking to express that conclusion, just a slightly modified one, would the study be so objectionable? Again, perhaps it is from a lack of imagination, but it doesn't even seem that SPE or MSE or any social psychological experiment could "prove" exactly that normal or good people can act badly/poorly/etc. under the right conditions, because either that sentiment is trivially true or so open to interpretation as to be meaningless in terms of how one go about designing an experiment to prove it.

Incidentally, what do you think of Milgram's experiment in comparison?

Zimbardo compared to Milgram

Good thoughts, Billie. Milgram was explicitly studying the effects of demand. He wanted to see how far his experimental subjects would go in carrying out his "demands" to perform cruel actions. He did many such experiments and showed that there are some conditions in which people would give what they believed to be powerful shocks to the victims and other conditions in which they wouldn't. Overall, it was a study of the conditions that increased or decreased people's willingness to obey authority or rebel against it. I discuss Milgram's experiments extensively in my textbook and think there was real value in that work, though I also discuss the limitations of applying it as an explanation of real-world atrocities. In contrast, Zimbardo did not claim to be studying the effects of his demands, or to be studying obedience at all. He claimed he was modeling a prison and studying the effects of a prison-like situation on the behavior of guards and prisoners. We can interpret Zimbardo's experiment as a study of demand; but unlike Milgram's studies, it was not a good study of demand, because that was not its original purpose and, unlike Milgram, he did not manipulate conditions to see when people would or would not obey him. -Peter

It seems to me that Milgram's

It seems to me that Milgram's experiment falls into the same category of delusion on the part of the experimenter, caused by confusing that which is in play, with reality.

I hang out with academic hypnotists who make the same mistake studying MPD, only theirs is a double delusion, namely the delusion of hypnosis and the delusion of Multiple Personality Disorder.

I beg to differ

I disagree with Dr. Gray's decision not to include one of the most famous experiments in psychology in his introductory psychology text. His blog, a response to my own (The Rarely Told True Story of Zimbardo's Prison Experiment), explains what he found to be lacking in the study but did not cover, as I did, the hidden gems in the original article. Would it not make more sense to describe the study in the text and then discuss its shortcomings? I try to teach my students to read experiments, famous and otherwise, with a critical eye. This study could provide a could vehicle for such an analysis. Furthermore, Dr. Gray does the students whose professors assign his book a disservice. His text does not give them information considered so basic to psychological knowledge that it's in every single social psych course, not to mention the exam needed to get into graduate school (the GRE). I only hope the instructors who use his text make up for this deficit by teaching about the prison experiment in the classroom- covering both its weaknesses and its strengths.

I think this is a valid

I think this is a valid point, and obviously many textbook authors agree with it. But I think Dr. Gray is right. There is a place to talk about a study and its shortcomings, and a social psychology textbook would be a good place to start. But in an Intro Psych textbook, generally speaking, the goal is, and should be, to communicate the basics. I could see Zimbardo's study going in a chapter (or a box) on "how to do research right" as a cautionary tale, but I'd prefer to leave it for the next level classes. They can be ready for the GRE without encountering it in their first class.

Dr. Gray disagrees

As I noted in my comment above, Dr. Gray has published an article in the Observer in which he argued that critical thinking skills are at the heart of the course. He wrote:

From a liberal arts perspective, it is hard to imagine any objective more fundamental than that of improving students’ critical thinking. Indeed, one definition of liberal arts education is that it is education that liberates the mind from the bondage of habit and custom.

Students' preparation for the GRE (or the MCAT, given the recent changes to that test) shouldn't be our concern in the course. A very small fraction of our students ever go on to graduate work. Most of them will forget most of the details of our courses. But modeling and reinforcing habitual critical thinking is a potentially transformative goal, and as Gray noted and Nisbett has found, studying psychology supports those goals. We can't teach habitual critical thinking if we tell our students that it's not important enough to cover in intro.

Thanks, Susan

Susan, thanks for your thoughts here. For more on why I decided not to include the study in the book, see my responses to some of the other comments here--especially the one (below) that I titled "Responsibility." Different textbook authors have different view of what is responsible or not responsible in textbook writing. -- I wish you continued success with your blog! -Peter


Interesting article and I always thought this experiment was somewhat silly. I, myself, did four year in prison and I can say most gang member committing crime look at prison as a stepping stone to climb the criminal hierarchy. Meaning, interaction amongst guards and prisoners is at a minimum of possible conflict, although it does happen. Prison, at least in California, is as I said a projection of the gang life on the streets. It does not stop in prison, rather, it is even more so as group identity is all important for survival. So, again, this experiment is silly.
I am getting ready to transfer to a 4 year college, then plan to do my own research on such matters. The difference, I have first hand knowledge of how it really is. Thus, in my opinion, we need a fresh look at prison and gangs. Here, is where evolutionary psychology has much to offer.

Post new comment

The content of this field is kept private and will not be shown publicly.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Allowed HTML tags: <a> <em> <strong> <cite> <code> <ul> <ol> <li> <dl> <dt> <dd>
  • Lines and paragraphs break automatically.
  • You may quote other posts using [quote] tags.

More information about formatting options

Peter Gray, Ph.D., research professor at Boston College, is author of the newly published book Free to Learn (Basic Books) and Psychology (a textbook now in its 6th edition).


Subscribe to Freedom to Learn

Current Issue

Let It Go!

It can take a radical reboot to get past old hurts and injustices.