Skip to content

Instantly share code, notes, and snippets.

Embed
What would you like to do?
stanford-prison-cites
Sarah asserts, "I knew there were some sketchy things about the Stanford prison
experiment but whew this piece!"
Dan asserts, "A rare situation where even I don't care about the truth, either
they weren't monsters and we shouldn't say they were, or they were made monsters
and we still shouldn't say they were. Either way, grant them their peace. And we
should still worry about deference to authority."
Sarah asserts, "Not sure you caught the implications — that the basic premise of
the experiment, that there's a monster in all of us, exonerates systems... A
cultural belief that sadism is a spontaneous result of situational placement
rather than, say, carried out at the behest of your direct superiors as part of
a racialized system, seems to matter?"
Thomas assets, "Also the pretty frightening bit about how the experiment
contributed to the “nothing works” attitude that stymied prison rehabilitation
programs for decades."
Dan asks for a citation, but never gets it and seems uninterested in obtaining
it from the article on his own.
So! Citations:
CLAIM 1:
"the SPE's premise is that there's a monster in all of us [waiting to be
"unlocked by a situation], exonerates systems [because it presumes that the
"blame is inherent to the situation and not on the guards or their superiors.]
CITATION 1:
> The appeal of the Stanford prison experiment seems to go deeper than its
scientific validity, perhaps because it tells us a story about ourselves that we
desperately want to believe: that we, as individuals, cannot really be held
accountable for the sometimes reprehensible things we do. As troubling as it
might seem to accept Zimbardo’s fallen vision of human nature, it is also
profoundly liberating. It means we’re off the hook. Our actions are determined
by circumstance. Our fallibility is situational. Just as the Gospel promised to
absolve us of our sins if we would only believe, the SPE offered a form of
redemption tailor-made for a scientific era, and we embraced it.
> Many other studies, such as Soloman Asch’s famous experiment demonstrating
that people will ignore the evidence of their own eyes in conforming to group
judgments about line lengths, illustrate the profound effect our environments
can have on us. The far more methodologically sound — but still controversial —
Milgram experiment demonstrates how prone we are to obedience in certain
settings. What is unique, and uniquely compelling, about Zimbardo’s narrative of
the Stanford prison experiment is its suggestion that all it takes to make us
enthusiastic sadists is a jumpsuit, a billy club, and the green light to
dominate our fellow human beings.
> “You have a vertigo when you look into it,” Le Texier explained. “It’s like,
‘Oh my god, I could be a Nazi myself. I thought I was a good guy, and now I
discover that I could be this monster.’ And in the meantime, it’s quite
reassuring, because if I become a monster, it’s not because deep inside me I am
the devil, it’s because of the situation. I think that’s why the experiment was
so famous in Germany and Eastern Europe. You don’t feel guilty. ‘Oh, okay, it
was the situation. We are all good guys. No problem. It’s just the situation
made us do it.’ So it’s shocking, but at the same time it’s reassuring. I think
these two messages of the experiment made it famous.”
> “What the Stanford Prison Experiment did,” Cullen says, “was to say: prisons
are not reformable. The crux of many prison reforms, especially among academic
criminologists, became that prisons were inherently inhumane, so our agenda had
to be minimizing the use of prisons, emphasizing alternatives to prison,
emphasizing community corrections.”
> In an era of rapidly rising crime, this agenda proved politically untenable.
Instead, conservative politicians who had no qualms about using imprisonment
purely to punish ushered in a decades-long “get tough” era in crime that
disproportionately targeted African Americans. The incarceration rate rose
steadily, standing now at five times higher than in comparable countries; one in
three black men in America today will spend time in prison.
> It would, of course, be unfair to lay mass incarceration at Zimbardo’s door.
It is more accurate to say that, for all its reformist ideals, the Stanford
prison experiment contributed to the polarizing intellectual currents of its
time. According to a 2017 survey conducted by Cullen and his colleagues Teresa
Kulig and Travis Pratt, 95% of the many criminology papers that have cited the
Stanford prison experiment over the years have accepted its basic message that
prisons are inherently inhumane.
> “What struck me later in life was how all of us lost our scientific
skepticism,” Cullen says. “We became as ideological, in our way, as the climate
change deniers. Zimbardo’s and Martinson’s studies made so much intuitive sense
that no one took a step back and said, ‘Well, this could be wrong.’”
> Most criminologists today agree that prisons are not, in fact, as hopeless as
Zimbardo and Martinson made them out to be. Some prison programs do reliably
help inmates better their lives. Though international comparisons are difficult
to make, Norway’s maximum-security Halden prison, where convicted murderers wear
casual clothing, receive extensive job-skill training, share meals with unarmed
guards, and wander at will during daylight hours through a scenic landscape of
pine trees and blueberry bushes, offers a hopeful sign. Norwegians prisoners
seldom get in fights and reoffend at lower rates than anywhere else in the
world. To begin to ameliorate the evils of mass incarceration, Cullen argues,
will require researching what makes some forms of prison management better than
others, rather than, as the Stanford prison experiment did, dismissing them all
as inherently abusive.
CLAIM 2:
"[rather than] sadism spontaneously arising from a situation [as the SPE seems
"[to show], [the body of evidence indicates that sadism tends to arise
"[explicitly from authority figures instead.]"
CITATION 2:
> Despite the Stanford prison experiment’s canonical status in intro psych
classes around the country today, methodological criticism of it was swift and
widespread in the years after it was conducted. Deviating from scientific
protocol, Zimbardo and his students had published their first article about the
experiment not in an academic journal of psychology but in The New York Times
Magazine, sidestepping the usual peer review. Famed psychologist Erich Fromm,
unaware that guards had been explicitly instructed to be “tough,” nonetheless
opined that in light of the obvious pressures to abuse, what was most surprising
about the experiment was how few guards did. “The authors believe it proves that
the situation alone can within a few days transform normal people into abject,
submissive individuals or into ruthless sadists,” Fromm wrote. “It seems to me
that the experiment proves, if anything, rather the contrary.” Some scholars
have argued that it wasn’t an experiment at all. Leon Festinger, the
psychologist who pioneered the concept of cognitive dissonance, dismissed it as
a “happening.”
> A steady trickle of critiques have continued to emerge over the years,
expanding the attack on the experiment to more technical issues around its
methodology, such as demand characteristics, ecological validity, and selection
bias. In 2005, Carlo Prescott, the San Quentin parolee who consulted on the
experiment’s design, published an Op-Ed in The Stanford Daily entitled “The Lie
of the Stanford Prison Experiment,” revealing that many of the guards’
techniques for tormenting prisoners had been taken from his own experience at
San Quentin rather than having been invented by the participants.
> In another blow to the experiment’s scientific credibility, Haslam and
Reicher’s attempted replication, in which guards received no coaching and
prisoners were free to quit at any time, failed to reproduce Zimbardo’s
findings. Far from breaking down under escalating abuse, prisoners banded
together and won extra privileges from guards, who became increasingly passive
and cowed. According to Reicher, Zimbardo did not take it well when they
attempted to publish their findings in the British Journal of Social Psychology.
> A few years later, after deciding to write a book about Alex’s story, I
discovered evidence that he hadn’t told the whole truth about his involvement.
When I confronted him, he confessed to me that his choice to participate in the
bank robbery [which was suggested to him by his superiors in the military] was
freer and more informed than he had ever let on before. Accepting responsibility
was transformative for him. It freed him from the aggrieved victim mindset in
which he had been trapped for years. Zimbardo’s “situational forces” excuse had
once appeared to give my cousin a way to believe in his fundamental goodness
despite his egregious crime, but seeing the personal growth that came with
deeper moral reckoning, I began to wonder if it had really done him a service.
@dakami
Copy link

dakami commented Jun 12, 2018

The fundamental claim of SPE is that normal people acting as prison guards can exhibit a variety of abusive behaviors.

The counterclaim that they may not, that they may instead be reasonable or even cowed by their quarry, does not actually rebutt this.

We have significant evidence of two things:

  1. There exists a wide range of prison cultures, ranging from what is effectively house arrest, through those where guards are seen as legitimate management, to places where just keep your head down and you'll be OK, to...worse.
  2. When things get worse, and there is no corrective force (i.e. a guard gets punished), things continue to get worse along that vector.

If SPE is interpreted as "all it takes to make us sadists is to give us a billy club and we immediately go bad", well, not only is that untrue, but I don't think anyone credible thinks that. The more troubling interpretation is "over time, there exist conditions can make sadistic prison guards out of people who were not originally sadistic". And frankly, there is plenty of evidence for that. It's hard to to define All Children Left Behind as anything other than sadistic, but there appear to be conditions that have over time left made it a position people do not feel free to aggressively reject (or more accurately, do not feel like an expression of their opinion will do anything, so why waste the political capital).

But let's focus on the actual material being cited here:

There is a claim that SPE makes humane prisons not exist, because it makes people think prisons are fundamentally inhumane. That is a causal relationship, that should have causal evidence. I can't see any. 95% of criminology papers may be citing SPE, but that does not mean criminology papers meaningfully influence prison policy. What does influence policy is "get tough on crime" being a winning strategy in elections for district judges, attorneys, and sheriffs. If criminology papers had a causal influence on general culture you might have me there, but I think historical crime rates were a little stronger there. Point being, it's pretty silly to say, "Oh, SPE made us think shitty prisons were a good idea". If this was any more motivated reasoning it'd literally be closing its eyes saying "I think I can I think I can I think I can".

And since Cullen is not an idiot, this is pretty much exactly what he says. SPE "contributed" to a larger trend, in an "era of rapidly increasing crime". He knows as well as I do that there were powerful forces at play pushing towards toxic incarceration policies. He's just annoyed that SPE makes the toxicity seem inevitable. Well, let me tell you something. Prisons don't have to be this bad, but you know? They're bad. Do they make guards abusive? Well, if you don't actively and aggressively push back on the abuse, abuse is what you're gonna get.

===
“What the Stanford Prison Experiment did,” Cullen says, “was to say: prisons
are not reformable. The crux of many prison reforms, especially among academic
criminologists, became that prisons were inherently inhumane, so our agenda had
to be minimizing the use of prisons, emphasizing alternatives to prison,
emphasizing community corrections.”

Here's the thing. Prisons are inherently inhumane. That is a true statement. When they are seemingly not inhumane enough, people complain! They're like, why am I struggling so hard, when that guy is getting $70K/year from the state to live with easy housing, full medical, three meals a day, and no requirement to work? That is a thing that happens and I can easily show you lots of direct evidence if you doubt me. The inhumanity of prisons is a design requirement in a society that is failing to deliver basic services. Placing any of this blame on SPE says more about an obsession with SPE (and it's implications for our own corruptible minds) than it does on trying to build a more just society. Thus, not really caring about little things like minimizing the use of prisons, or emphasizing alternatives, or community corrections. Those are things about other people.

It's a strawman, to say that SPE makes people think all prisons are equally shitshows. I don't think anyone ever argued that. So Cullen's point, that no, we need to figure out what makes prisons not shitshows, is pointlessly linked to SPE. SPE only demonstrates that there's no group of angels you can make prison guards, and be 100% sure will never become corruptable. Because (and I can't believe I have to spell this out) if wealthy white kids can become monsters, anyone can.

Fifty years later that remains true and the wealthy white kids are still pissed.

Let's go to Claim 2.

The methodology of SPE is pretty poor, and it shows in that everyone focuses on the weak part -- that the failure is unlikely to occur instantaneously, and that the initial conditions and promptings of the guards does matter in terms of how long decency holds.

But the reality is that, given time and given arbitrary initial conditions, systems will arrive at a "balance of forces". And that balance of forces will ultimately wear "inherent decency" away just as reliably as Zimbardo claims happens immediately. That's an important thing to recognize. It's how you avoid the slippery slope fallacy in a world where there are in fact slopes that do exist, that are in fact slippery. You can't just say, "Zimbardo was wrong, people don't immediately become monsters, therefore there's no way to know they'll become monsters". Reality is, Zimbardo's right, look around, lots of people say monstrous things now.

We can't run the sort of experiments that would really explain what's happening, because we would end up psychologically damaging people. But we don't need to run the experiments because, uh, we're running natural experiments every day, with millions incarcerated in prisons of various quality. Every moment spent obsessing about this shitty little experiment from half a century ago is a moment spent shoehorning our concerns about ourselves into apparent concerns about other people. This is the lens through which to understand why I take offense to this article. Look how it's so important that people didn't come up with sadistic behaviors themselves, that they were taught them by the experimenter by his experience at San Quentin. See, you and I haven't been to San Quentin, so we'd never fail that horrible horrible way.

Meanwhile, word gets round, prison guards learn the San Quentin methods, life becomes actually more miserable for actual real people.

But you and I, we still get to be good people.

If you're actually concerned about prisons, the Stanford Prison Experiment is literally the least interesting thing in your life. And if it makes people think we should be using prisons less, that's great. I also don't want arguments about how our next round of Japanese Internment camps doesn't actually need to be so bad!

@zyphlar
Copy link
Author

zyphlar commented Jun 13, 2018

@dakami

(1) you say:

The fundamental claim of SPE is that normal people acting as prison guards can exhibit a variety of abusive behaviors.
The counterclaim that they may not, that they may instead be reasonable or even cowed by their quarry, does not actually rebutt this.

However this is an overly-weak interpretation of the counterclaim, and an overly-broad interpretation of the SPE. The SPE's claim under scrutiny is that normal people acting as prison guards naturally exhibit abusive behaviors as a consequence of the guard-prisoner dynamic. And the counterclaim is in fact that normal people acting as prison guards do not do that, but rather more often exhibit abusive behavior when encouraged to by an authority they respect. In other words, Milgram's experiment may be valid but Zimbardo's is not.

All we're arguing about is the bit in italics, not the bit in normal type. Those differences are crucial to understanding our position. We agree on most everything else.

I think this confusion might be central to why you're defending SPE so much: you seem to be assigning it more meaning than it really conveys. We're trying to make a very specific point, but you seem to think we're trying to throw away the entire field of psychology-based prison reform/replacement.

(2) you say:

If SPE is interpreted as "all it takes to make us sadists is to give us a billy club and we immediately go bad", well, not only is that untrue, but I don't think anyone credible thinks that. The more troubling interpretation is "over time, there exist conditions can make sadistic prison guards out of people who were not originally sadistic".

But I absolutely do think that PSYCH101 professors and students -- indeed Zimbardo's own words -- take the SPE to mean that, in the short course of a week, normal mild-mannered college students become abusive guards and psychotic prisoners purely through the facts of their circumstance. That's how the experiment is explained, and that's why the fact that Zimbardo and his colleagues instructed the guards to be "tough" and in various arbitrarily-abusive techniques matters so much. He admits in the article that he was an activist and probably let his desire for prison abolition to color how he handled and reported on the experiment. As evidence for guard-prisoner dynamics on their own -- not authority, not natural abusive tendencies -- spontaneously creating an abusive situation, as SPE purports to, the evidence shows that SPE fails. As evidence that guards can often somehow become sadistic, there is no question. That's not even worth a study, we hear about sadistic guards in the news weekly. The question is why they become sadistic and SPE used to have answers for that: it no longer does. Milgram's experiment is now a better explanation of why.

(3) you say:

 ...that does not mean criminology papers meaningfully influence prison policy...
 and
 "Oh, SPE made us think shitty prisons were a good idea"

I am not here to debate whether SPE did or did not influence prison policy. That's part of the original article but it's not one of Sarah or Thomas' or my assertions. Also nobody is saying "Oh, SPE made us think shitty prisons were a good idea."

(4) you say:

Prisons are inherently inhumane. That is a true statement.
The inhumanity of prisons is a design requirement in a society that is failing to deliver basic services. Placing any of this blame on SPE says more about an obsession with SPE (and it's implications for our own corruptible minds) than it does on trying to build a more just society

However the article cites prisons which are not inhumane, in other societies that do not apparently have the dynamics you describe. So it is not a true statement. It might just be true for most prisons, in most countries, but the focus of SPE was to declare the guard-prisoner dynamic inherently inhumane, and SPE has been falsified, and we see evidence of guard-prisoner dynamics that are not obviously inhumane, so we're back at square one: looking to other experiments like Milgram to see why some prisons are definitely inhumane and others are seemingly quite humane. SPE is unnecessary and misleading baggage on that discussion because the experiment was tainted from the beginning.

(5) you say:

It's a strawman, to say that SPE makes people think all prisons are equally shitshows. I don't think anyone ever argued that.

We agree there. That's definitely a strawman. SPE might make the claim that the guard-prisoner dynamic is inherently evil, but that doesn't purport equality, only badness. There are still degrees of badness. And SPE has been falsified, therefore for all we know some prisons could be good and others could be evil. The question is why.

So Cullen's point, that no, we need to figure out what makes prisons not shitshows, is pointlessly linked to SPE.

I don't think that it's a pointless link. Obviously we want to figure out what to do with murderers in our society, and SPE tried to prove that guards are inherently immoral, but failed, so we still need to figure out what to do with murderers in our society... ideally humanely. If we're not trying to reform/revolutionize prisons, SPE is a pointless exercise.

SPE only demonstrates that there's no group of angels you can make prison guards, and be 100% sure will never become corruptable. Because (and I can't believe I have to spell this out) if wealthy white kids can become monsters, anyone can.

We can't be 100% sure that any group of people will be uncorruptable. That's a tautology, a useless statement, and a weak/broad summary of the SPE. Again, the specific bit of SPE under critique is whether guards will naturally become corrupt as a consequence of the position itself. And SPE is false, so we're left wondering why "wealthy white kids can become monsters." And other experiments, like Milgram (and indeed the meta-narrative of SPE itself), show that the answer is not that they naturally did so, but that they did so because they were told it was the right thing to do. You seem very hung up on the overarching themes that SPE touches on, but the article and my/Sarah/Thomas's conclusion isn't that the whole theme is bunk, but that the answer to the question of "why do good people do bad things" is less "because of their situation" and more "because they were convinced that being bad is good." That's it. SPE was absolutely used as proof of situational corruption, but it's bunk, and it turns out that coercion is one of a few much better explanations. That's all we're doing here. That's all we wanted you to read, was that bit, and now that you read it, we're finally debating apples and apples. That was literally the only goal of this whole long megathread.

(6) you say:

The methodology of SPE is pretty poor, and it shows in that everyone focuses on the weak part -- that the failure is unlikely to occur instantaneously, and that the initial conditions and promptings of the guards does matter in terms of how long decency holds.
But the reality is that, given time and given arbitrary initial conditions, systems will arrive at a "balance of forces". And that balance of forces will ultimately wear "inherent decency" away just as reliably as Zimbardo claims happens immediately. 

Because the SPE was set up with scientists prompting the guards to be sadistic, the whole SPE is just bunk. All it shows is that some people who were told to act shitty did as they were told. (Which bolsters Milgram, leaving Zimbardo hanging out to dry.) It does not test or show how quickly things degrade, because the test was set up in a degraded state. It does indeed show that the initial conditions and promptings do matter (Milgram.) It does not test or prove anything about systems arriving at a balance: that's an assertion you're making on your own. Whether it's true or not is good material for another experiment: how long can good guards given good instructions maintain goodness in degraded conditions? That's what SPE purports to test, but it does not, because they were given bad instructions immediately.

(7) you say:

You can't just say "Zimbardo was wrong, people don't immediately become monsters, therefore there's no way to know they'll become monsters". Reality is, Zimbardo's right, look around, lots of people say monstrous things now.

Nobody is saying that. Science doesn't prove negatives, "there's no way to know _____" is an inherently false statement. Zimbardo purported to prove that people spontaneously become monsters within a week of acting as a prison guard. That is false. We know that some monstrous prison guards obviously exist. Other experiments besides Zimbardo's flawed experiment explain the data more fully and reliably than Zimbardo did. The existence of monstrous people does not prove Zimbardo right, because Zimbardo did only one flawed experiment and made a specific flawed claim. Monstrous people do exist, and if we believe that they were once good people, other experiments explain that transition better than Zimbardo. Specifically, experiments like Milgram where good people are told by a trusted authority that doing bad things is actually good. I'm sure other experiments expand on this and test other aspects. We will arrive at the truth of this question if we reject flawed experimental evidence and seek/accept new, stronger experimental evidence.

(8) you say:

We can't run the sort of experiments that would really explain what's happening, because we would end up psychologically damaging people. But we don't need to run the experiments because, uh, we're running natural experiments every day, with millions incarcerated in prisons of various quality.

We agree that it'd be real hard to re-run Milgram or Zimbardo properly and ethically, and that in-situ studies of actual prisons would be a good idea. I'm not a scientist or science funder, but if I were I would advance that cause.

(9) you say:

Every moment spent obsessing about this shitty little experiment from half a century ago is a moment spent shoehorning our concerns about ourselves into apparent concerns about other people. 

I feel like if we want science and criminal justice to be good and trustable, it's important to refute bad experiments and find/create/promote good experiments. This is part of that. I don't know about you, but this has nothing to do with my concern about myself, and I don't even care who Zimbardo or Milgram are. I do care that my understanding of human psychology and criminal justice policy is scientifically sound, because I do write my senators about public policy and accurate foundational knowledge is essential to advocate good policy. I think we have time to do both: learn good science and advocate for real-world improvements in policy affecting others. Has nothing to do with concern about myself. Not a waste of time.

If you're actually concerned about prisons, the Stanford Prison Experiment is literally the least interesting thing in your life.

I am actually concerned about prisons. But I'm interested in psychology. I'm a pretty good multitasker, I can do both. You are setting up a lot of emotional false dichotomies towards the end here.

And if it makes people think we should be using prisons less, that's great.

The means of perpetuating a flawed experiment that tangentially absolves people of culpability because they can blame it on their situation is not justified by the ends of making people think that prisons are bad. In fact I'd say that's the thesis of the article, though I don't care to argue about that.

I also don't want arguments about how our next round of Japanese Internment camps doesn't actually need to be so bad!

The reason racial internment camps are bad isn't because they have abusive guards, it's because people are deprived of liberty and property based on race instead of due process for a crime they individually committed. So this is a non sequitur. Other experiments, data, and morality explain why racial internment camps are bad, and it was not a thing tested by the SPE. Both Milgram and SPE are predicated on the idea that those being punished deserve the punishment, and it's just a question of specifically how guards lose their functional empathy for prisoners. If you presume that those being punished do not deserve it, or want to explain the social forces giving rise to racial internment, you need a whole other set of experiments. This is science, you can't just grab the nearest study and dress it up until it makes the point you want it to make: all the scientific method shows us is that, when done right, a well-crafted hypothesis explains observed responses to stimuli better than the null hypothesis ("there is no relationship between stimulus and response.") It doesn't even prove the hypothesis, it just indicates that the hypothesis is so far a better answer than nothing. To get "proof" the experiment needs to be repeatable by any experimenter, and that is where SPE falls woefully short. SPE proves nothing. It tried to say something about prison guards and failed. It says nothing about internment camps. If we want guidance about internment camps, we have to look elsewhere. Don't try and inflate SPE to try and prove that point, because it'll blow up in your face: there is other, better evidence for the question of guards and internment than SPE. That's all we're saying. We know guards often become sadistic. Milgram, and the new reading of SPE provided by this article, suggests that it's more often the guards' belief that they are doing what's expected of them, either explicitly or implicitly, than the mere fact that they are "guards" assigned to guard "prisoners." Prison sadism is not a foregone conclusion, and there is plenty of evidence for that.

@dakami
Copy link

dakami commented Jun 13, 2018

The SPE's claim under scrutiny is that normal people acting as prison guards naturally exhibit abusive behaviors as a consequence of the guard-prisoner dynamic. And the counterclaim is in fact that normal people acting as prison guards do not do that, but rather more often exhibit abusive behavior when encouraged to by an authority they respect"*

You're proposing a viral model of monstrousness, where without guidance from an authority, a sort of "original decency" prevails.

That is a bold claim, and not I think a true one. SPE's obvious failing is that it suggests the collapse is near-instantaneous. That's as false as all the people who think, in the event of a disaster, their neighbors are going to bash down their door and try to kill them. The data does not reflect that. Cooperation dramatically rises, to a point.

The Iron Law of Social Science is that hungry people riot. Nobody tells them to. No trusted authority pushes the riot button. Pain is its own authority, and hunger is a source of pain. You can't just model ideas as being transferred from one person to another. At some point, circumstances create concepts -- explanations, predictions, actions. It is terrifying to believe that, under the right circumstances, you will start forming certain ideas and concepts that you presently consider noxious. Alternatively, it's an excuse people use, oh anyone would have become a Nazi in this situation, you can't blame me!

What I see of value in SPE, is: Don't let those circumstances occur. You're not covered because you hired "the right people". Upper class white kids, in these circumstances, with these pressures, with those incentives, became monsters. You want to argue that is a thing that does not happen, and I will tell you, it does. There's increasingly distressing evidence all around us. It's super uncomfortable, and so we doth protest too much.

He admits in the article that he was an activist and probably let his desire for prison abolition to color how he handled and reported on the experiment.

There's a context you're missing, which is that at the time this work was done, there were two prison systems. One, the "justice" system. The other, mental institutions. It was a thing, a constant background radiation threat, to just pluck people out of the population. This other system was draped in Science, and had no problem grabbing White People.

That's a system as coldly rational, as intelligently designed as you can imagine. It's also the only Prison system I'm aware of that was actually dismantled. SPE sought to show, you design a system according to certain rules, you're going to see certain behaviors emerge, no matter who's involved. That's not wrong, though I wish it was. It's not people in a system have to become corrupt, but if a system does not manage it's corruption, it will be consumed by it.

Oddly, that's your complaint about SPE, that it "cheated" by having respected authorities suggest bad behaviors. Yes, that certainly accelerates collapse. But, like, there's some physics here. If they'd suggested, I dunno, not wearing socks -- probably not as easy uptake.

Also nobody is saying "Oh, SPE made us think shitty prisons were a good idea.

“What the Stanford Prison Experiment did,” Cullen says, “was to say: prisons are not reformable.

That...is in fact what Cullen is saying here. And it's the "legitimizing hook", if you will. It's how you make this, not about how we feel about ourselves:

“You have a vertigo when you look into it,” Le Texier explained. “It’s like, ‘Oh my god, I could be a Nazi myself. I thought I was a good guy, and now I discover that I could be this monster.’

Yeah, that feels terrible. And so there's this giant pile of motivated reasoning that says, make this experiment invalid.

However the article cites prisons which are not inhumane, in other societies that do not apparently have the dynamics you describe.

Look closer, I directly addressed this: "The inhumanity of prisons is a design requirement in a society that is failing to deliver basic services." The delivery of basic services is critical! If good actors cannot make end's meet, those who are bad actors must be worse off. Otherwise, why be good? Just shoplift something and get taken care of, which is actually a thing that is happening with increasing frequency (particularly in Japan).

Did you ever think of that? That how we manage free society places bounds on the quality of life in prisons? That prison health care becomes problematic if free people can get better care by trading in their freedom? The flip side of that is when prisons are too good, free society demands they be made worse. So that effect in this article, where the claim is somebody would only become magically evil if they were ordered to behave like those San Quentinites, is not a random failure. It's systemically inevitable.

If we're not trying to reform/revolutionize prisons, SPE is a pointless exercise.

SPE demonstrates that ugly systems will generate ugly people. You don't want ugly people? Don't run ugly systems.

The experiment is obviously faulty because it accelerates the process. But the process obviously, inexorably occurs. People really want SPE to be false, and thus they're incorruptible. I wish it were true.

SPE was absolutely used as proof of situational corruption, but it's bunk, and it turns out that coercion is one of a few much better explanations.

Situational corruption exists, because situations cause coercion causes situations causes coercion ad nauseum. That you draw this false distinction is really the most interesting element of this entire thread, and makes me consider this time extremely well spent. I had no idea there was a huge debate about whether circumstances or people influence behavior. I am trying to thing of a more false dichotomy, and I can't. Other people are engines of circumstance that actively but non-randomly alter their behaviors in a variety of responsive ways.

The problem is the concept that, if we are corruptible by situation, then how can we be responsible for our behavior? And the answer is, yes, we are expected to resist corruption. Even if it is difficult, even if the corruption is incredibly tempting, to a point.

Practically nobody would shoot a baby. But there is in fact some level of torture that would create that outcome. We're moderately understanding of these sorts of acute corruptions. It's corruption that takes longer, but not that much longer, that we don't understand well.

SPE completely fails to capture the temporal component, and I think that's a major flaw in the experimental design, and even in the conceptual form. So when you say:

Zimbardo purported to prove that people spontaneously become monsters within a week of acting as a prison guard.

I'm willing to concede that this is uniquely terrifying, and somewhat unfairly so. It probably is the case that many people, not under significant distress, not explicitly given direct instructions to misbehave, don't "go dark".

But some do. And if nobody punishes those defectors, more, and more, and more will, because as the rule goes, 10% will do right, 10% will do wrong, 80% are seeing who wins. You have to be incredibly careful about the systems you build. You can't just say, oh, we hired a bad guy to be a torturer. If it was a good guy torturer, real stand up family man, everything would have been ok.

However imperfect SPE is, the purported gains from rebutting SPE are ten kinds of ugly. Whenever a system fails, it's only because the wrong person was hired. We got rid of the monster so we can start from scratch, who knows what the new guy will do.

I feel like if we want science and criminal justice to be good and trustable, it's important to refute bad experiments and find/create/promote good experiments. This is part of that.

This article isn't about finding, creating, or promoting good experiments. It's about loose claims that SPE made prisons bad, and about how those guards didn't really go bad anyway. Now, I could sit here and say, those people are lying, they're making up excuses for why they weren't monsters. That's as valid for me to say, as it is for them to say nuh uh we weren't. It's also as helpful, by which I mean, I don't care. If saying this now, fifty years later, makes them feel better -- good. Nobody should have to live with scientific proof that they could be turned into a monster, nobody should have to live with scientific claims that they could be turned into a monster. I don't want us to do anything but agree, those people experienced an abusive experiment that revealed nothing about them, personally.

But seriously, let's not forget the lessons of SPE with regards to the systems we do or do not build. You build a monster engine, you're gonna get monsters. It ain't random.

@dakami
Copy link

dakami commented Jun 13, 2018

"We just gotta stop hiring monsters for prison guards, that's the problem. It's our pipeline."

@zyphlar
Copy link
Author

zyphlar commented Jun 13, 2018

You're proposing a viral model of monstrousness, where without guidance from an authority, a sort of "original decency" prevails.

I didn't say that, if anything I said the opposite. I said "people exhibit abusive behavior when encouraged to by an authority they respect".

Now, is monstrousness viral? I think studies would show that yes, values/morals/behaviors are transmitted socially, like a meme.

Does monstrousness happen without guidance from an authority? SPE says yes, but it's false so it's no help there. Milgram says that one way that it happens is explicitly under guidance from an authority. We would need tests to know whether decency prevails in a vacuum or whether monstrousness organically arises. We've been debating whether Lord of the Flies is true for a long time now, and I don't think we have a good answer. We have definitely seen examples of both immediate monstrousness in vacuums and remarkable cooperation and humanity in vacuums. I would wager that "put humans in a vacuum" is not sufficient predictor of monstrosity, and we have to look at other metrics and details. It's just hard to design experiments because keeping innocent random people in cages against their will for long periods of time is against our moral code. In short, we don't know. But I am not making a claim or proposition to that effect. I'm just saying SPE is bunk.

SPE's obvious failing is that it suggests the collapse is near-instantaneous. That's as false...

So we agree that SPE's main assertion is false! My job here is done.

At some point, circumstances create concepts -- explanations, predictions, actions. It is terrifying to believe that, under the right circumstances, you will start forming certain ideas and concepts that you presently consider noxious.

I don't disagree. I and the article in question just disagree that SPE shows that. We would need other studies on the interactions between circumstances and morality (of which I'm sure there are a few already.)

What I see of value in SPE, is: Don't let those circumstances occur. You're not covered because you hired "the right people". Upper class white kids, in these circumstances, with these pressures, with those incentives, became monsters.

Sure but "if you hire 'good' guards and tell them to act brutally for a good cause, some of them will do it," is not a unique experiment: it's already been done by Milgram, it's already obvious. The whole reason SPE exists, and the whole reason this article exists and the whole reason Sarah and Thomas and I are arguing with you, is because the SPE tried to show that any guard-prisoner situation will inherently and organically create abuses. And that's just false. We have counter-evidence, SPE is a lie, it's just false. Zimbardo doesn't get a spot on national TV for a groundbreaking study saying that "if you fill a good person with bad ideas about being a prison guard, they'll be brutal." We don't need a psychology professor to tell us that that's true. He tried to show that prisons are by definition abusive and he failed. That's it. Everybody with a brain already knows that "a few bad apples" in your criminal justice system will poison the whole thing. We simply do not need SPE, SPE is fake news, and the value you see in it is not worth the harm in believing falsified data. If you want to prove this point that you've just tried to make, just use Milgram's study to make the exact same point and you'll get way farther with it with way less argument.

The flip side of that is when prisons are too good, free society demands they be made worse. ... It's systemically inevitable.

This is not shown by SPE and is falsified by the fact that good prisons exist in other societies. You're using an anecdote about contemporary American society to make a fallacious point about all possible societies.

SPE demonstrates that ugly systems will generate ugly people. You don't want ugly people? Don't run ugly systems.
The experiment is obviously faulty because it accelerates the process. But the process obviously, inexorably occurs. People really want SPE to be false, and thus they're incorruptible.

The truth we now see about SPE demonstrates only that guards who are told to be evil will be evil. It says nothing about systems except for the preceding sentence, it does not say that it obviously or inexorably occurs, and falsifying SPE does not prove that people are incorruptible. Milgram is still true, people are still corruptible, an ugly system (i.e. one where supervisors encourage/tolerate bad guards) will still be ugly. SPE just isn't necessary to prove any of that.

Situational corruption exists, because situations cause coercion causes situations causes coercion ad nauseum.

I don't personally dispute that. SPE just doesn't show that because it's a lie. Milgram shows coercion, and it's obvious that situations and coercion could have some feedback loops, so ditch SPE. Use Milgram.

 I had no idea there was a huge debate about whether circumstances or people influence behavior. 

I personally believe that both do. However SPE does not prove that circumstances influence behavior, because it started out with the experimenters explicitly dictating behavior. Other tests would be needed. Ditch SPE. Use Milgram. Seek out other situational studies which surely exist.

It probably is the case that many people, not under significant distress, not explicitly given direct instructions to misbehave, don't "go dark".

Then you agree that SPE is probably bunk and more experiments are needed. Because Zimbardo would not agree with you here. He was trying to prove that "many people, not under significant distress, not explicitly given direct instructions to misbehave, DO INDEED QUICKLY AND SOLELY BY THEIR CIRCUMSTANCE 'go dark'." SPE is bunk. Look at Milgram.

But some do. And if nobody punishes those defectors, more, and more, and more will, because as the rule goes, 10% will do right, 10% will do wrong, 80% are seeing who wins. You have to be incredibly careful about the systems you build.

Neither Milgram nor SPE tests the case where guards are allowed free reign, some eventually go bad (an untested presumption as far as I know), and those who go bad are either punished or not punished. That would be a really great experiment! Neither Milgram nor SPE did that experiment however. And SPE is bunk. You have a very compelling anecdote that I personally agree with, but I know of no scientific study that backs it up. I would like to see such a study.

You can't just say, oh, we hired a bad guy to be a torturer. If it was a good guy torturer, real stand up family man, everything would have been ok.

SPE tried to test this, but it failed. I would absolutely like to see the results of an honest, reproducible SPE to see what happens when good guards endure the life of being a guard, with good supervisors for a long time. Obviously if they are instructed to torture, then we know the outcome: that's what Milgram tests. There are no "good guy torturers." The question is, is it possible to have good prison guards? The existence of idyllic Norwegian prisons suggests it is possible. Further research is needed.

However imperfect SPE is, the purported gains from rebutting SPE are ten kinds of ugly.

False. Truth and reproducibility in science are the only thing that tells us anything about our world. If we kept, and supported, flawed science based on false premises, we'd literally still think that the sun revolves around the Earth. (I mean look at it, it's obvious to anyone with eyes!) -- sure you might cause a Crusade over the truth that the Earth revolves around the sun, but the truth will let your species set foot on the moon within six generations. The truth is worth it. And the downsides are far less severe than you're insinuating. Milgram already proves what you seem to be trying to prove, letting go of SPE won't hurt. Just open your eyes, the ground is right there, you won't fall. Most politicians probably don't even know what the Stanford Prison Experiment is.

[this article is] about loose claims that SPE made prisons bad, and about how those guards didn't really go bad anyway.

That's one element in the article which you'll notice is not the bit quoted above. The bit Sarah, Thomas and I wanted you to read is this bit in this Gist, and now that you've read it we're finally debating apples and apples instead of Chapter Two Oranges versus Chapter Three Apples. Are there loose claims that SPE made prisons bad? Yes. But that's not the part that interested us. The interesting and useful part was the two claims we made that you have been resisting: "the SPE's premise is that there's a monster in all of us [waiting to be unlocked by a situation], exonerates systems [because it presumes that the blame is inherent to the situation and not on the guards or their superiors.]" and "[rather than] sadism spontaneously arising from a situation [as the SPE seems [to show], [the body of evidence indicates that sadism tends to arise [explicitly from authority figures instead.]" Those two points indicate that clinging to SPE is simply unnecessary, possibly harmful, and other studies more-accurately show what SPE actually showed now that we know the truth. What SPE purported to show was tainted, is not useful, is bad science, and much much more research is needed to figure out the truth of. (Perhaps long-term studies of prisons that control for the goodness or badness of guard supervisors? Just not one-week studies that were tainted by obviously-evil supervisors.)

Nobody should have to live with scientific proof that they could be turned into a monster, nobody should have to live with scientific claims that they could be turned into a monster.

Unfortunately, Milgram (and the true SPE) shows that you can become a monster if you're told by someone you respect that monstrous actions are justified. Whoopsie, we have an individual duty to disobey immoral orders even when they fit with our biases, and we can't blame it on "the job."

I don't want us to do anything but agree, those people experienced an abusive experiment that revealed nothing about them, personally.

We definitely agree the guards experienced an abusive experiment in both SPE and Milgram. Unfortunately, both showed that they could be convinced to behave like monsters. Whoopsie.

But seriously, let's not forget the lessons of SPE with regards to the systems we do or do not build. You build a monster engine, you're gonna get monsters. It ain't random.

We both agree that if you build a monster-engine, you get monsters. The SPE just doesn't tell us much about what elements of an engine make it monstrous. Is it the fact that there are people in cages? Is it the fact that the people outside cages are called "guards" with all the social presumptions associated? We don't know. Maybe. Maybe not. Is it the fact that guards can be convinced to act terribly? Yes, SPE and Milgram both show that guards can be convinced to act terribly. An engine which asks its participants to be monsters will most likely become a monster engine. SPE is not required to show that.

"We just gotta stop hiring monsters for prison guards, that's the problem. It's our pipeline."

I don't think either SPE or Milgram make this assertion, and I don't think absence of SPE will justify anyone making this assertion. I think this is what you're worried will happen if SPE is abandoned, but I don't think it's likely. Milgram and common sense already prove that anyone can be convinced to act terribly.

The question is not "how do we make lazy disprovable excuses for our prison problems," the question is "what elements of problem prisons make them problems, and what elements of good, rehabilitative criminal justice keep it good?" When people re-run SPE, they have not yet found that guards naturally become evil. When people re-run Milgram, they often find (and we can see in daily life) that good people can be coerced to do bad things. So with those bits of information, we know that a "good criminal justice system" could include guards and "prisons," but must train those guards to do good things and never do bad things. Anecdotal evidence suggests that getting rid of the cages and poor living conditions and negative attitude from guards may be a good idea, we should probably test stuff that scientifically.

SPE is not necessary, and in its original presentation may actually hinder "good prisons."

I think at this point my original goal has been satisifed: literally all I wanted you to do was read the paragraphs Sarah and Thomas and I were trying to get you to read. You've now read and digested them. We win, whether you want to reject or uphold SPE at this point is up to you but we at least finally got you to RTFM and it only took like three dozen tweets and two days.

@dakami
Copy link

dakami commented Jun 14, 2018

This is silly. You have entirely missed the point that the societies with humane prisons are themselves much more humane, and that the humanity of the society creates a bound for the humanity on the prison. You seem to be desperately holding onto "even if it's true, this is a bad experiment", which has validity but not as much as you desire. I'm not going to go point by point on this, you're going to believe what you want to believe, because you consider the outcome of losing this argument to be more noxious (basically, that you can't blame prison guards for being evil).

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment