Criticism of the Stanford Prison Experiment.

[RandFan]

Having referred to the experiment numerous times I feel it appropriate to point out the criticisms that have come to light in recent times.

The following are pod casts. The first provides documented arguments and points. The second also deals with the Millgram experiments in addition to the Standford (Philip Zimbardo) which I have also referenced many times.

http://skeptoid.com/episodes/4102

Psychology is complicated, and there will probably never be a perfect theory explaining all human behavior; so people should never assign too much significance to the results of any given experiment like the Stanford Prison Experiment. And, when an experiment receives a large amount of scholarly criticism from mainstream science, as this one did, you have very good reason to look past its portrayal in the popular media and, instead, be skeptical.

http://www.pointofinquiry.org/ (http://www.pointofinquiry.org/)

In this interview with D.J. Grothe, Barbara Oakley shares her criticisms of the research of influential social scientists such as Philip Zimbardo and Stanley Milgram, and explains why the biological sciences should be brought to bear on research about human evil. She addresses how her thesis in Evil Genes might be used as an excuse by some people in our society to do bad things, and details specifics from the life of her sister that serve as a window into her exploration of human evil. She also addresses the implications of her thesis for organized religion, arguing contra Christopher Hitchens that religion is not evil per se but that it might attract evil people to its institutions.

Yes, I should have been skeptical and it's clear I was too ready to accept the conclusions.

Damn! :(

[GreNME]

I don't see why the conclusions aren't still useful to a degree. Giving them too much weight, I agree, would be a mistake, but in terms of psychology and its social implications they are still quite illuminating to how humans operate in a perceived hierarchy.

[RandFan]

I don't see why the conclusions aren't still useful to a degree. Giving them too much weight, I agree, would be a mistake, but in terms of psychology and its social implications they are still quite illuminating to how humans operate in a perceived hierarchy.They're interesting but flawed. The former in particular.

[GreNME]

I don't know if flawed is the proper term. Not entirely representative is a better descriptor, I think.

[Travis]

Slight derail, didn't they make a foreign language film based on the Stanford Prison Experiment?

[Smackety]

Some people just assign far too much importance to experimental findings - they prove what is possible but they do not provide much in the way of predictive power because every situation is unique. Even if we attempted to recreate the experiment today there would be so many differences from the original that the results could be completely different. But they are not invalid results: the subjects really did adopt a different persona in short order.
I think it was a useful experiment - at least it provides a possible mechanism to explain findings like this (http://www.haworthpress.com/store/ArticleAbstract.asp?sid=DHFHXRN31CDS9K5LG9G9BMBHQQ CK5NF0&ID=34286) (a reduction in recividism through trancendental meditation) that otherwise might look like pure woo.

[Jeff Corey]

Technically, Milgram's 1963 "A behavioral study of obedience" was not an experiment. There was no independent variable.

[ImaginalDisc]

They're interesting but flawed. The former in particular.

I'm a little confused. I thought the most cutting criticism of the Stanford Prison Experiment was ethical, not procedural.

[Smackety]

I'm a little confused. I thought the most cutting criticism of the Stanford Prison Experiment was ethical, not procedural.

The procedure to place subjects into prisoner or guard roles was not properly double-blinded (among other things), so it is possible they self-selected into 'comfortable' roles. It calls the common (and mistaken) conclusion that ANYONE who is placed in such a situation will fall prey to groupthink into question, but does not invalidate the actual finding that it is possible. Remember when the 'guards' started treating the 'prisoners' as if they were guilty of something? It was astonishing! The ways we resolve cognitive dissonance are amazing sometimes.

[Elaedith]

The procedure to place subjects into prisoner or guard roles was not properly double-blinded (among other things), so it is possible they self-selected into 'comfortable' roles. It calls the common (and mistaken) conclusion that ANYONE who is placed in such a situation will fall prey to groupthink into question, but does not invalidate the actual finding that it is possible. Remember when the 'guards' started treating the 'prisoners' as if they were guilty of something? It was astonishing! The ways we resolve cognitive dissonance are amazing sometimes.

I don't see any way that assignment to roles could possibly be blinded or double-blinded. Participants obviously knew what role they were in, as did any observers.

Self-selection would be a violation of random assignment to conditions which is a different issue from blinding. The study was reported to have used random assignment (toss of a coin) to assign roles.

[Jeff Corey]

I don't see any way that assignment to roles could possibly be blinded or double-blinded. Participants obviously knew what role they were in, as did any observers.

Self-selection would be a violation of random assignment to conditions which is a different issue from blinding. The study was reported to have used random assignment (toss of a coin) to assign roles.

True. Zimbardo's active participation as the "prison superintendent" raises the issues of experimenter and confirmation bias, however.

[volatile]

The experiment has been repeated, with similar results each time (as far as I understand). What exactly is your [procedural / evidential] criticism?

[Jeff Corey]

The experiment has been repeated, with similar results each time (as far as I understand). What exactly is your [procedural / evidential] criticism?
I am not familiar with any published replications, and given that Zimbardo felt he had to terminate it early, would be surprized if there were any. The criticism appear in the first link in the OP.

[Professor Yaffle]

Here's one replication/reality TV experiment:

http://crimepsychblog.com/?p=987

[Jeff Corey]

And apparently the results were not "similar". By the way, such studies would probably not be published in the US now. APA and APS journals require that one documents the approval of the university (or other entity) Institutional Review Board to ensure conformity with current ethical treatment of human participants standards.

[RandFan]

I'm a little confused. I thought the most cutting criticism of the Stanford Prison Experiment was ethical, not procedural.Did you listen to either podcast? The first is pretty short and makes excellent argument. I'm not one to argue via link so if you want me to I will summarize. It's just that he does the job better than I could and it is not a big investment of time.

Let me know.

[RandFan]

I don't see any way that assignment to roles could possibly be blinded or double-blinded. Participants obviously knew what role they were in, as did any observers.

Self-selection would be a violation of random assignment to conditions which is a different issue from blinding. The study was reported to have used random assignment (toss of a coin) to assign roles.

The problems are legion.

There are numerous methodological problems including the selection of the participants through the advertisements. A separate study has been conducted showing that such an ad is likely to get participants with given predispositions.

Also, the conclusions are clearly confirmation biased as evidence that is counter to the investigators theory is ignored.

[RandFan]

I don't see any way that assignment to roles could possibly be blinded or double-blinded. I think you've hit on the most serious problem of the study. That an experiment can't possibly meet rigor is hardly reason to elevate the experiment. You are making the same argument that Homeopaths and Psychics use.

If the experiment isn't ameniable to scientific testing then what good is it?

[Elaedith]

I think you've hit on the most serious problem of the study. That an experiment can't possibly meet rigor is hardly reason to elevate the experiment. You are making the same argument that Homeopaths and Psychics use.

If the experiment isn't ameniable to scientific testing then what good is it?

I'm not making an 'argument'. I'm responding to an apparent confusion between blinding and sampling issues.

These problems are true of any social psychology study that requires participants to take on defined roles, uses volunteers, and does not use extensive deception. They are not specific to this study.

Any study that uses volunteers has self-selected participants. Participants can't ethically be compelled to participate in any study and must give informed consent, meaning they have to be told something of what the study is about.

The alternatives would be to use naturalistic observation only (and sacrifice any experimental control such as random allocation to conditions), violate ethical guidelines, or not conduct any such research.

[RandFan]

Any study that uses volunteers has self-selected participants. But clearly there are better methods for gathering self-selected participants as was pointed out in the second pod-cast.

The alternatives would be to use naturalistic observation only (and sacrifice any experimental control such as random allocation to conditions), violate ethical guidelines, or not conduct any such research. The study is flawed for the reasons stated in the podcasts. The flaws are numerous. There are ways to reduce error in sociological studies.

Our problem is one of imagination. That you and I can't conceive of ways around the problems of these two studies hardly proves that there are no such ways.

[Smackety]

There are many many other, less controversial studies, that prove the same point one step at a time. I suppose I will be asked to cite some of them and I guess I could put in the effort if someone is truly interested. The whole Zombardo thing is really only used in introductory psychology to get students interested, you don't really see it after that.

[Jeff Corey]

You could mention those studies and what point was made. Apparently I missed them.
And, believe this, the Zimbardo and Milgram studies are covered in advanced experimental psych courses required of psych majors as examples of flawed studies that would be unethical today.

[Smackety]

You do not believe that groupthink or cognitive dissonance have been adequately shown to exist by other studies?

[Jeff Corey]

Those studies have little, if anything at all, to do with groupthink or cognitive dissonance. Zimbardo supposedly demonstrated that assignment of roles can drastically change behavior and Milgram studied conformity to authority.

[Smackety]

Why do you think that assignment of roles can drastically change behavior? You do not think it might have something to do with cognitive dissonance?

[Phaedrus74]

Slight derail, didn't they make a foreign language film based on the Stanford Prison Experiment?

I think you are referring to Das Experiment (http://www.imdb.com/title/tt0250258/)

[Jeff Corey]

Why do you think that assignment of roles can drastically change behavior? You do not think it might have something to do with cognitive dissonance?

I never said that I thought assignment of roles can drastically change behavior. But in any case I don't see how the hypothetical motivational state called cognitive dissonance is relevant.

[Smackety]

It is relevant because provides a possible mechanism through which assignment of roles can drastically change behavior. I am surprised you use the term "hypothetical" - is there an alternate theory that predicts similar behavior?

[Jeff Corey]

It is relevant because provides a possible mechanism through which assignment of roles can drastically change behavior. I am surprised you use the term "hypothetical" - is there an alternate theory that predicts similar behavior?
I use the term hypothetical in the sense that it is an inferred, assumed or theoretical mechanism whereby two unobservable cognitions are purported to be in conflict and generate an unobservable aversive emotional state, which theoretically prompts a person to somehow modify one or both of the unobservable cognitions to be more congruent.
How this applies to Zimbardo's controversial interpretation of his questionable results escapes me.

[Mercutio]

Jeff--

I gotta check my copy of Obedience (the book, not the movie), but I thought it did indeed have an independent variable, or variables (they varied the salience of the learner and of the teacher, and varied whether the participant was alone or part of a group, and when part of a group, which role he or she played).

On the other hand, the Zimbardo experiment had no independent variable, and was horrific design. Not the first bad design for Zimbardo, actually.

The Milgram experiment, for all its notoriety, really was the beginning of good experimental ethical controls. Milgram, if I recall, went to great pains to debrief his experimental participants and to follow up on their reaction to the experiment.

Although one could presumably work Cognitive Dissonance into these experiments (I have seen it done with Milgram, at least), there are many (CD, self-perception, role theory, prolly a few more) potential explanations, once we start allowing hypothetical intervening processes. Or we could just stick to what was manipulated and what was measured, but what would be the fun in that?

[Jeff Corey]

Merc,
Not in the original study. There was no independent variable, all subjects were treated the same, except some who showed relucance to continue were give more "prods" to continue pressing the higher and higher shock button.
Later on, he varied factors like gender (not an easily manipulated IV) and proximitry to the "learner".
A real shocker was Sheridan & King (1972), who thought that Milgram's subjects might penetrated the ruse and knew the "learner" was not being shocked, got undergrads to give real shocks to puppies who whimpered, cried and shrieked in real pain.
Social psychology, oxymoron of kings.

[technoextreme]

I don't know if flawed is the proper term. Not entirely representative is a better descriptor, I think.
Yeah it is. Any experiment that has a basic underlying assumption that people are morons is a stupid experiment in my book. Even shocking a dog still has that underlying assumption that people are morons.
real shocker was Sheridan & King (1972), who thought that Milgram's subjects might penetrated the ruse and knew the "learner" was not being shocked, got undergrads to give real shocks to puppies who whimpered, cried and shrieked in real pain.
You do realize it was still a ruse though and any reasonable educated person would easily see through it. In fact I know of a special place where people watch people risk getting hit with a couple thousand volts. Scary? Yes. Painful? Probably. Harmless? Yes.

[Jeff Corey]

Yeah it is. Any experiment that has a basic underlying assumption that people are morons is a stupid experiment in my book. Even shocking a dog still has that underlying assumption that people are morons.
*Indignant electrical engineer here*

I think the assumption was that "People just follow orders". (Godwin, 1990).

[technoextreme]

I think the assumption was that "People just follow orders". (Godwin, 1990).
Ehhhhh.... No. There is no assumption because the problem is that they idiotically fell for the same trap that they felt the original experiment had. Get back to me when the pain inflicted is on the level of blunt force trauma like breaking someone's finger with a hammer.

[Jeff Corey]

Huh?

[technoextreme]

Huh?
You said that the repetition of the Milgram experiment was done with a dog because people thought it was a ruse. The problem is that the same can be said with the dog.

[Jeff Corey]

Let me grab an adult beverage and I'll get back to you.
State dependent learning and all that.


<--- Schatzi didn't like that experiment either.

[technoextreme]

Let me grab an adlt beverage and I'll get back to you.
State dependent learning and all that.

I'm an electrical engineer not a psychologist. I just so happen to know that people have been hit with something that produces about two million volts. The immediate assumption is that they were insane to even put themselves in such a position but it is perfectly harmless. *Note in that case it was perfectly harmless. Please don't assume that means it's a good idea to go sticking a fork in the outlet*

[Jeff Corey]

I'm an experimental psychologist and would like to know, "How many Amperes?"

When I was an aide to a shrink who gave electroconvulsive therapy, he would always says, "It ain't the volts that fry you, it's the amps."

[technoextreme]

I'm an experimental psychologist and would like to know, "How many Amperes?"

When I was an aide to a shrink who gave electroconvulsive therapy, he would always says, "It ain't the volts that fry you, it's the amps."
That's an excellent question. I honestly have no idea namely because Im talking about the original two story van de graaf generator whose discharges sound like a gun going off. It's got to be low enough because otherwise they wouldn't make it a science experiment where people put themselves at risk of getting hit. It would make one heck of a science demonstration if it was dangerous..... My god.... The person is dead!!!!
http://mos.org/sln/toe/cage.html

[Jeff Corey]

And how about that Nick Tesla and his coil. He appeared on the stage in NYC about 100 years ago covered with lightning.

[Smackety]

I use the term hypothetical in the sense that it is an inferred, assumed or theoretical mechanism whereby two unobservable cognitions are purported to be in conflict and generate an unobservable aversive emotional state, which theoretically prompts a person to somehow modify one or both of the unobservable cognitions to be more congruent.
How this applies to Zimbardo's controversial interpretation of his questionable results escapes me.

I am saying that what Zimbardo saw was, in part, cognitive dissonance at work.

[autumn1971]

I agree that the methodology and conclusions may be, respectively, open to much criticism and not scientifically validated. I do submit that anyone who has ever worked behind a counter in the US can confirm that most of the people encountered by said workers, regardless of any prior opinion granted by title or ostentatious display of wealth, are utter and complete anuses.
Perhaps I am simply showing where I would show up on a scale of "guards", but I have gained an almost total contempt for the public: merely by working in positions that exposed me to them.
I'm in a particularly cynical mood right now.

[Lonewulf]

I agree that the methodology and conclusions may be, respectively, open to much criticism and not scientifically validated. I do submit that anyone who has ever worked behind a counter in the US can confirm that most of the people encountered by said workers, regardless of any prior opinion granted by title or ostentatious display of wealth, are utter and complete anuses.
Perhaps I am simply showing where I would show up on a scale of "guards", but I have gained an almost total contempt for the public: merely by working in positions that exposed me to them.
I'm in a particularly cynical mood right now.Just remember that people have a tendency to remember the hits more than the misses. 5 people come up, do the business, and are polite, can go unremembered, whereas one guy coming in as an a-hole will give memories for a lifetime.

[ImaginalDisc]

Did you listen to either podcast? The first is pretty short and makes excellent argument. I'm not one to argue via link so if you want me to I will summarize. It's just that he does the job better than I could and it is not a big investment of time.

Let me know.

It wasn't working. I had to tell my firewall that Brian Dunning isn't evil. :-p