In 1971, a team of psychology researchers led by Philip Zimbardo studying how people respond to authority conducted what has become known as the Stanford Prison Experiment.
This book is Zimbardo’s account of and reflections on that experiment almost forty years later, and his attempt to make sense of it and relate it to more contemporary phenomena, such as the atrocities at Abu Ghraib.
In the Stanford Prison Experiment, volunteer college students were randomly split into groups of “prisoners” and “guards” and placed in a college building that had been converted into a makeshift prison. The guards were not armed and were disallowed from using direct physical violence on the prisoners, but they were given instructions to keep the “prison” running in an orderly fashion. Zimbardo and his team got them started with some procedures designed to disorient the prisoners and minimize their individuality (dressing them in identical garb, replacing their names with numbers, etc.), but beyond that left it up to the guards to keep the prisoners in line by whatever rules and whatever enforcement they deemed appropriate.
The experiment was intended to run for two weeks but was scrapped after only six days, due to the organizers’ alarm that the role playing had become entirely too real, and that emotional damage was being done to the increasingly sadistic “guards” and increasingly humiliated and submissive “prisoners.” Indeed, Zimbardo confesses that he was wrong not to stop the experiment even sooner, and contends that he himself had come under the influence of the circumstances just as the mock guards had, and had similarly become inured to the suffering of the mock prisoners.
Zimbardo and others contend that the experiment demonstrates how powerful situational forces are, and how evil tends not to be a matter of bad people spontaneously doing bad things, but (almost) all people in certain perverting situations doing bad things. Critics disagree over whether this is the most plausible explanation of the behavior exhibited in the experiment, and the degree to which the experiment can be generalized to real life.
I find the experiment, and this book about it by Zimbardo, to be highly thought-provoking. I’ll structure the remainder of my discussion around a (very partial) list of the questions that reflecting on the book raised in me.
1. Did Zimbardo make the right call in stopping the experiment early?
Had it been up to me, I don’t think I would have stopped the experiment when Zimbardo did.
But if I had, I believe it would have been on different grounds.
Zimbardo was most focused on the suffering, the harm. I would be more concerned with the consent.
Sometimes people choose to go through something difficult, painful, risky, traumatic, etc., for a variety of reasons. It may be to challenge themselves, to toughen themselves up for future experiences, to bring about a certain mental and emotional intensity that makes them feel vital, to impress someone, to bring about certain consequences that are important enough to them to override the sacrifice, or even something as simple as their receiving something in exchange for the ordeal–e.g., money–that they feel makes it a worthy trade.
I don’t believe in forcing such challenges on people. But if it’s something they’re doing by choice, then I’m inclined to leave them be.
Not that I would never intervene or put a stop to things if I could. But like I say, for me it’s far more a matter of examining if there is something dubious about the consent, rather than ascertaining the degree to which the person is suffering. Whereas for Zimbardo, the opposite seemed to be the case.
I would further say that all else being equal, the easier it is to revoke consent, the less inclined I am to paternalistically intervene because the person is being harmed.
If you hire some drill sergeant style personal coach to berate you constantly to work out and diet and do all the things that are good for you that you have struggled with due to a lack of self-discipline, then even if you complain about it and seem to be unhappy with his treatment of you, it’s not for me to somehow prevent him from continuing to provide this service to you. As an autonomous human being, you get to decide if undergoing the unpleasantness of dealing with this guy is worth whatever good you think it’s doing. And if and when you decide it isn’t, you–not me–can fire him at will.
Whereas irreversible consent is something I’m not so comfortable with. The whole “can your present self bind your future self?” raises all kinds of questions about autonomy and coercion and related matters.
Consider the difference between these two cases, for example: The officer training program in An Officer and a Gentleman with Richard Gere, versus a soldier in the current volunteer military going through some form of boot camp or basic training and being sent off to war. Both are hellish in their way. I wouldn’t consent to either one. I would have some doubts about anyone who would consent to either one. I’m not convinced of the ultimate benefits for society as a whole of either of these things.
However, though they have many things in common, one very important difference is that in the first case the person’s consent to undergo the suffering and abuse and such is always revocable. A large part of the drama of the movie is the whole question of whether the participants will succeed in sticking it out, since they always have the choice of walking away. Whereas in the second example, participants are not allowed to change their mind. Their initial consent is treated as binding on their later self, and if they attempt to cease participating, the powers that be feel justified in punishing them severely.
So while it’s certainly possible to approve of both or neither, the second has a major additional element of concern for anyone who assigns disvalue to imposing suffering on someone contrary to their (present) will.
But in the experiment, that complication isn’t present. It’s not as if the participants bound themselves for two weeks and we are now having to decide whether to give them a break and change the rules in the middle of the experiment. They have always had the freedom to change their mind and walk away. To intervene now and “protect” them by ending the experiment prematurely is paternalism. Perhaps a case can be made that it is justified paternalism, but at the very least I would give the participants’ consent to continue the experiment a great deal of weight.
So I look at it now with two questions in mind: Is their continued consent tainted in some way, such that it’s not really consent or it’s only sort of consent, and thus it deserves no weight or less weight than full consent? And, are other factors–that is, the harm Zimbardo perceives and is most influenced by–of sufficient weight to override the consent factor?
As to the first matter, the most troubling thing by far to me is that it appears some of the participants became convinced that they were not free to quit after all, as a result of the one unbalanced prisoner telling them–why is unclear–that he’d tried to quit and not been allowed to. So even though it remained objectively the case that they always were free to leave, subjectively they thought that option had been removed.
That undercuts consent significantly. For me to be free to do something, I not only have to have the opportunity to do it, but I have to know I do, and know how to go about doing it.
That the experimenters didn’t bring about these false beliefs matters little. Once the participants think they can’t quit, then in an important sense their continued participation is non-consensual.
I believe once the experimenters became aware of this misimpression, they were under an obligation to unambiguously re-inform the participants that they were still fully free to quit at any time. That may well have adversely affected the “illusion” that was making the experiment so valuable, but so be it. I would have chosen to risk weakening the experiment in that manner, over continuing it in a way that was arguably non-consensual or putting an end to it entirely.
Another consent-based objection that I’m considerably less sympathetic toward is that some of the participants may have still been aware they could quit, yet due to the inherent pressures of the situation made an unwise choice to continue, and that that did not really constitute continued consent because in a sense they were of diminished capacity.
Pride was the likely culprit in most such cases. For example, one prisoner reluctantly continued despite his mother’s protests because he didn’t want to shame himself and feel like less of a man in his father’s eyes (and his own). Some would see that case as especially troubling because Zimbardo manipulated the situation to encourage the prisoner to feel that way.
No doubt people make questionable decisions under powerful and distorting influences all the time, but I think we have to tread very carefully in labeling those decisions as somehow non-genuine or non-autonomous.
People have all kinds of reasons for the choices they make, influenced by all kinds of values and character traits. I’m prepared, as noted above, to say that continued participation in the experiment is non-consensual–or at least significantly less consensual–if it is based on the false belief that “If I announce a decision to quit, it will be turned down and I will be laughed at and punished for even trying to quit.” But I’m not prepared to say that if the decision (regardless of whether I agree or disagree with it) is based on a person’s desire to challenge and test himself, his need to feel like a man, his desire to practice for when he may be arrested and imprisoned for real for anti-war or other protest activities, his desire for money, his belief that this experiment is providing benefits to science and humanity, his competitive urge to outlast his fellows, etc.
I would make the call on a case-by-case basis (which is basically what they were doing before the decision to halt the experiment entirely). If an individual is so out of it as to be basing his decision to continue on delusions, then yes, his consent is becoming less and less meaningful. Otherwise, have some respect for autonomy and their desire to continue.
But the other question is, are the participants being so severely damaged by the experiment that their consent to continue should be set aside and the experiment ended?
Clearly this is what Zimbardo believes, and he thinks it’s obvious and that he was only blinded to it at the time by being too close to the experiment and having his thinking distorted.
I think you can make at least as good a case that his being so close to it is what caused him to wimp out.
When we’re not talking about coercion or a violation of autonomy, but instead about something’s beneficial and harmful consequences, then for me I don’t think the consequences for the direct participants deserve any more or less weight than the consequences for anyone else. (Sort of. It depends on if we’re talking moral ontology or moral epistemology. I would say, as a rule of thumb direct consequences are much more ascertainable and measurable, while the more indirect the consequences the more speculative and uncertain they are, therefore it’s generally correct to focus much more on the pros and cons of the immediate situation. But not because those costs and benefits are somehow more important than others.)
But I don’t know that even if we focused exclusively on the harms and benefits to the participants that I’d be inclined to have halted the experiment prematurely. And once we add in the indirect impact on other people down the line, I think things would come out even more strongly in favor of continuing it.
Of course the participants were disliking much of what was going on, even to some extent being traumatized by it. But I’ll bet the harm was mostly short term, while some of the benefit had the potential to be more lasting.
Think about what these individuals are learning about themselves, both good and bad, that they can put to use in the future. They’re gaining a much better sense of their strengths and weaknesses, where they need to improve, what they need to guard against. They’re being given reasons for being humble, perhaps learning to be more understanding and compassionate toward others who behave in less-than-heroic fashion in trying circumstances. They’re earning the sense of accomplishment of taking part in an experiment that will ultimately be famous and regarded as very beneficial in advancing our understanding of certain psychological facts. They’re going to be able to say “No matter how much I hated it, I endured.”
And remember, I’m not saying “I know better than they what’s good for them, and I think we should force them to continue no matter how much they want out.” They want to continue. It’s the opposing side that would be saying, “We know better than they what’s good for them, and we think we should force them to quit even though they want to continue.”
Who knows what people can endure, if they choose to tough it out. I still have a videotape from 1982 of Julie Moss participating in a triathlon, which remains one of the most amazing sports performances I’ve ever seen. Having led the eleven hour race most of the way, she was staggering to the finish almost completely spent. Once she got within two hundred yards or so, she was alternating walking and running. A little closer and what little control she’d had left of her mind and body faded away. Disoriented, she had to be pointed in the right direction. She collapsed, got up, collapsed, got up, collapsed. She soiled herself. She collapsed a final time just a few feet from the finish line. Unable to get up, she crawled from there. She finished. (In second place, as another woman passed her while she was crawling.)
Why didn’t someone stop it? Why didn’t one of the organizers of the event realize the error of their ways, pity this human being who was so thoroughly beaten mentally and physically, and put an end to this avoidable human suffering by preventing her further participation?
Because in almost all cases it’s best to let the individual decide how much is too much. Let the person himself or herself determine if they’re benefiting enough to outweigh the suffering of what they’re going through.
I don’t know if some of the participants in this experiment will be traumatized in a way that adversely affects them the rest of their life, or if they’ll look back on it as one of the most valuable experiences of their life to learn and grow from, or both, or neither. But in order to take the decision out of their hands and force them to quit, the evidence better be very strong that this is damaging them in a significant and permanent way that they are not in a position to assess for themselves.
Then there’s still the matter of the consequences for other people besides the participants. This I think adds more weight in favor of continuing the experiment.
It’s not like this is some neutral activity they’re doing to pass the time. It’s something they’re doing because they have every expectation it will further the understanding of some things that are of importance to human psychology. And by all appearances it’s done just that. People are still talking about it decades later. Zimbardo is writing a book about it decades later and applying the lessons learned to contemporary real life situations.
None of those benefits would have come about had he been “enlightened” enough back then to recognize the supposed evil of the experiment and decided against conducting it. On the other hand, had the experiment been allowed to run its course for the full two weeks, one would have to think considerably more would have been learned.
To summarize, autonomy trumps the cost/benefit analysis of the consequences, so the main thing I’d have done in designing the experiment (or redesigning it on the fly as it was going on and the flaws were exposed) is to better safeguard consent. That includes making sure everyone was always aware that they were allowed to quit at any time, and what they had to do or say to quit.
(Humble suggestion: Let’s learn a little something from sexual deviants. In the S&M community, they have the concept of “safe words.” A “safe word” is a predetermined word that acts as a signal that one of the participants wants out, and upon the uttering of which the proceedings must immediately and completely stop. So for instance if the word is “fiddlesticks,” then no matter how much the submissive is crying, complaining, begging for mercy, screaming in pain, etc. while being whipped, beaten, kept in bondage, verbally abused, sexually abused, and so on, then it is assumed everyone is still “in character” and the submissive is getting what he at some level, as a masochist, wants. But as soon as he says “fiddlesticks,” then it is understood that he genuinely wants things to stop, and his wishes must be immediately respected. Something like that could have been used in this experiment to better distinguish when the participants were basically play acting or “method acting” that they were prisoners who wanted out of their confinement and suffering, and when they were signaling that they were experimental subjects genuinely calling for an immediate end to their participation in the experiment.)
Furthermore, individuals should have been closely monitored to see that their mental and emotional state had not deteriorated so much that they were delusional and no longer able to meaningfully consent, and if that point were reached, the decision could be taken out of their hands and their participation ended if doing so were determined to be in their self-interest.
It’s entirely possible these changes–mainly making sure they were always conscious of the option to quit–would have made it harder for the participants to stay “in character” and made the experiment even less equivalent to real imprisonment and thus less revealing of the psychological dynamics of such situations and less valuable overall. Or for that matter it might have resulted in them all quitting early. In that case, so be it, because, again, respecting the autonomy of the subjects trumps the consequences.
But if meaningful consent from enough of the participants to keep the experiment going happened to persist, then the experiment should have continued to the completion of the full two weeks. That is, unless the damage to the participants (or, less likely, to other people later or to society as a whole) was clear and significant and irreversible, with the standard for that being quite high.
And I don’t think at the time the experiment was prematurely halted, it had reached that level of severe harm that would have outweighed the benefits of its continuation.
So the experiment should have been allowed to continue, not to satisfy the cruel impulses of the experimenters, but to respect the decision of the participants to endure, and to bring about the benefits that greater knowledge of these psychological phenomena can lead to.
I fault Zimbardo for not better safeguarding consent. He let the prisoners operate under apparent false beliefs because doing so was making their experience more like that of real prisoners and thereby was making the experiment more valuable, and then he lost his nerve and stopped the experiment early. He should have done neither. He should have made sure that the subjects’ continued participation was always meaningfully consensual, and if it was he should have let the experiment run its two week course.
2. Beyond issues of non-consensuality or harm to the participants, what else was flawed about the experiment or should give one pause in interpreting its conclusions? To what extent did it succeed/fail in creating something sufficiently like prison conditions to warrant believing that its lessons can be applied to real life prison and prison-like situations?
Obviously the main difference is that the participants knew this was going to last only two weeks, and for that matter sort of knew it could last even less then that since they could quit at any time. It’s a hugely different situation psychologically to be incarcerated for real with no option to leave.
No doubt the experiment would have been a more accurate reproduction of the prison experience if the participants had been tricked into thinking they’d been incarcerated for real, but that’s not a morally acceptable option.
But it’s interesting that this difference didn’t seem to have the impact one would expect. I share Zimbardo’s surprise that the prisoners became as distraught, hopeless, and defeated as they did, given that they knew it was all just a brief role playing game.
One thing that stood out to me, just in comparing what I was reading with what I learned about prison from the time I spent as a prison volunteer, is that the experiment was quite a bit worse than real prison as far as the constant petty humiliations of the “count” and such. As bad as prisons are, my understanding is they’re typically not lining people up and making them recite rules and sing like some fraternity hazing baloney (though I wouldn’t put it past that psychotic “Sheriff Joe” in Arizona).
That style of breaking a person down through constant humiliation and reminders of his powerlessness seemed less like prison to me and more like the boot camp in Full Metal Jacket, or some brainwashing type re-education facility in a totalitarian Communist state, or one of those expensive torture facilities American teenagers who smoke pot are sent to to cure them. Most prisons I think are more about, “Let’s stick them somewhere and make sure they don’t escape.” I don’t think they put so much effort into concocting diabolical ways to destroy people psychologically.
I think there are just too many prisoners who would tell them to go [bleep] themselves because they’d prefer the punishment over cooperating with such humiliation. So there’d be constant conflict. I don’t see it as conducive to a smooth running facility. I mean, yes, if they could do it so thoroughly and so consistently as to be dealing solely with utterly defeated, obedient zombies, then I suppose that would be an easy to run facility. But I think reaching that goal is a lot harder with real life prisoners than with these experimental subjects.
And in this experiment you had a roughly one-to-one ratio of staff to prisoners to conduct the petty, demoralizing games. What prison could afford to hire that many guards?
Maybe more of this goes on in real prisons than I realize and I’m too influenced by my experiences, which were all at a prison that was unusually pro-rehabilitation and respectful of its inmates. But even when the prisoners described in detail their time at less enlightened facilities, such as in IMU (Intensive Management Units–where they throw the “worst,” most recalcitrant prisoners), as awful and dehumanizing and all the rest as the conditions were, I don’t recall it sounding like this kind of juvenile sadistic hazing, with people chanting and doing jumping jacks and all that crap.
(And for what it’s worth, I think this kind of brainwashing/humiliation/hammering home the message of powerlessness form of confinement would be even more nightmarish than regular prison to me, just due to my personality, values, phobias, strengths and weaknesses, whatever. If forced to make a choice, I would go to prison before I would go in the military or one of those cult-like drug rehabilitation things. Actually I might choose death over any of the above, but assuming that’s not an option, I would feel slightly less hopeless about retaining some minute amount of human dignity in prison compared to these other places.)
So even though it’s the Stanford “Prison” Experiment, and they added various trappings like visitors day and parole hearings and the like to make it more prison-like, I think what they put together was actually more like certain other horrific coercive institutions than like real prisons, and so the lessons learned might apply even more reliably to those non-prison environments.
Another difference is in real life you don’t have the guards and prisoners starting from scratch and making it up as they go along. The overwhelming majority of prisoners have already been “hardened” by earlier negative experiences, including earlier forms of incarceration such as in juvenile facilities. They’ve developed a whole elaborate set of habits, defense mechanisms, strategies, knowledge of what they can get away with and what they can’t, hustles, ways of competing with or forming alliances with their fellows, etc. Similarly, a new guard learns the ropes from veterans who’ve developed ways of doing things that “work” in some sense. It’s not like they’re all new at the same time; the new ones can mostly just follow along until they get comfortable.
I don’t want to overstate that. No matter how deeply you’re enmeshed in a criminal subculture, no matter how many times you’ve been through the revolving door, there’s always going to be an element of shock, of disorientation, of depression, when you’re suddenly deprived of your freedom. But I think that factor is stronger and more likely to leave you floundering if you have zero experience with it, and there isn’t even a pre-existing culture around you of people who’ve been through it all and learned the ropes that you can attempt to connect with or learn from.
As far as this factor goes, maybe the experiment is less like a typical prison, and more analogous to some Eastern European country quickly overrun by Nazis, and now all the Jews (largely naive civilians with no experience dealing with a criminal justice system) are tricked into gathering in the city square for some bogus “registration” and then whisked away to a concentration camp.
So on top of the physical conditions themselves–the confinement, the crappy food, the menacing guards, etc.–the experimental subjects had to contend with the additional debilitating and disorienting factor of unfamiliarity. They’re dealing with all this in a vacuum, with less in the way of internal resources than real prisoners would typically have.
So while I take it that one of the lessons is supposed to be the surprise that even “good” people like ordinary college students and squares can show sadistic qualities, turn on each other, display character flaws, and in general behave decidedly non-heroically in a prison environment, I would contend that it’s actually less surprising that we’d see that out of these folks than out of “real” guards and prisoners. I’d think panicky squares thrown into a totally new world with challenges importantly different from what they’re used to and where they have no relevant bearings, would be especially prone to make stupid, weak, or even cruel moves.
Another difference that comes to mind again has to do with the experimental subjects’ awareness that they’re in an experiment. I already mentioned their awareness that they can quit any time they want, but I think it goes beyond that.
I sensed that the participants often were unsure what the “rules” were. Not the rules of the “prison,” but the rules of the experiment. Compounding the problem was that one of the rules they were clear on was that they weren’t supposed to step out of character and refer in any way to this being an experiment.
So for instance, they knew they weren’t supposed to say, “I need to pause here for a second and ask a question. These letters you’re having us write. Are we supposed to write fictitious letters to fictitious people, like asking my non-existent wife to remain faithful to me while I’m in the joint the next ten years, or are we supposed to write letters to our real friends and family asking them to visit us here at Stanford and such?”
In a real prison you at least know the issues you’re dealing with are real; you aren’t in this gray area where you have to play make believe and try to act out how you think you would behave. The participants in the experiment are dealing with an ambiguous situation that’s sort of a prison and sort of not.
Zimbardo seems puzzled or alarmed at times the way the prisoners play along with made up facts–such as accusations thrown at them by make believe parole board members–or the way that one prisoner looks for convoluted technicalities about his contract that might give him a right to leave, rather than just announcing he’s quitting the experiment. He sees this as an indication that the pressures of incarceration are somehow damaging them psychologically, making them lose touch with their own identity as they internalize their roles.
But I think they’re just trying to guess at the ambiguous rules for their roles, and trying to abide by them for as long as they’re in the experiment. It’s not necessarily the case that the circumstances have driven them to some delusional state where they think they really did commit crimes, or really do need a lawyer, or really do need to convince a parole board they’re rehabilitated; more likely they’re intentionally playing make believe because they think that’s how the experiment’s supposed to work. (Though in other respects they did seem to be passing into some sort of delusional or disturbed or depressed state, which as noted several paragraphs above, is interesting and surprising.)
Some might say the willingness to play along in itself is what’s disturbing though, that the participants did the things they did and endured the things they endured just because they guessed that’s what the authority figures in the situation–the people running the experiment–required of them. But that’s a quite different phenomenon. That wouldn’t be “Wow, look how willing people are to obey their captors when they’re wrongfully imprisoned;” it would be, “Wow, look how willing people are to try to abide by the rules of a psychology experiment they’re voluntarily participating in.”
Which is quite a bit less disturbing. They’re going along with abusive behavior and such in a situation that’s a murky combination of make believe and reality, but ultimately make believe. I don’t think that’s enough evidence to jump to a more general conclusion about people being too willing to go along with authority. Just as I don’t think my willingness to play by the rules at Monopoly means I’m of a submissive and obedient personality type that will lead to my collaborating with evil when the government turns dictatorial.
A game’s a game. The experiment’s participants know it’s a game (mostly). Reality is reality. Real guards and prisoners know it’s reality.
Something I think is maybe less of a difference between the experiment and a real prison is that the guards in the experiment were supposedly limited to non-physical punishment and control devices, so they had to be inventive in coming up with verbal and psychological humiliations and such.
I say “supposedly,” because a lot of what they were doing was intended to and did cause physical suffering, not just emotional suffering. When people are confined in small rooms, or even smaller closet-type rooms (the “Hole”), when they are bound, when they are sprayed from a fire extinguisher, when they are forced to sleep on the floor without a blanket, when they have food crammed into their mouth and face in an effort to force them to eat, and so on, this is not non-physical abuse.
Granted, they weren’t beaten up with fists and night sticks, but that kind of thing is probably less prevalent in real prisons than people think. Yes, occasionally guards will “goon” a guy, but it’s not like that’s the routine way of controlling prisoners. It’s there as a last resort–which it wasn’t for the guards in the experiment–but prison staff nearly always gets its way without having to resort to it. So I don’t know that this was a huge difference between the experiment and reality.
Anyway, not surprisingly, the experiment only duplicated the psychological reality of a prison to a very limited degree. In numerous respects there were dynamics present in the experiment that would have been absent in a real prison, as well as dynamics absent in the experiment that would have been present in a real prison. So one must be cautious about generalizing too much from the experiment.
In spite of that, my feeling is the experiment is at least somewhat informative about real life. That’s just my gut; I’d need to do a lot of research about psychology experiments in general, and what the literature says about their reliability in mirroring the real world versions of what they purport to be studying. But I’m inclined to believe an examination of this experiment can indeed help us to understand some things about human nature, and how people react to certain situations, including being prisoners or imprisoners.
Well, there are many, many other aspects of this book worthy of analysis and discussion, but as this is monstrously long already, I’ll arbitrarily cut it off here.
Needless to say, I recommend the book, and recommend giving the matters it raises serious thought.