Tuesday, January 16, 2007

How 'bout them apples?

When it comes to social psychology, everyone seems to know two names: Milgram and Zimbardo. It's funny, because they both studied essentially the same thing - under what circumstances otherwise normal, compassionate people will act to hurt others.

I happen to prefer Milgram's experiments. They're a lot sounder, experimentally and ethically. Zimbardo sort of threw a bunch of undergraduates into a prison and waited to see what would happen. Milgram tested subjects one by one, systematically changing variables such as the immediacy of the victim, the presence of the experimenter, the presence of peers, even the gender of the subjects. Also, while many people have problems with the psychological harm caused to subjects (would you want to know that you were capable of shocking a person to near death, just because you were ordered to?), there was nowhere near the risk of physical harm to the subjects in Zimbardo's experiment (which had to be stopped after only 6 days).

Anyway, those of you who aren't familiar with the Stanford Prison Experiment can now see it whenever you'd like... on YouTube.

SciAm Blog has links to the videos as well as an excellent summary of the experiment itself.

What I want to talk is this quote from Zimbardo:

"It's not that we put bad apples in a good barrel. We put good apples in a bad barrel. The barrel corrupts anything that it touches."

This quote very clearly delinates between two modes of thought. There's the idea that people are inherently either bad or good, and that determines their behavior. This is called dispositional attribution of behavior. Then there's the idea, which Zimbardo ascribes to, that it's a matter of circumstance. That would be situational attribution of behavior.

Obviously, it doesn't have to be either/or. A bad situation can bias the behavior patterns of an individual. To go back to the Milgram experiment, about 60% of people were willing to shock the victim when he was in another room, but only 30% of people shocked the victim when they had to hold his arm to a "shock plate". On the one hand, you can see how changing the situation drastically decreased compliance. On the other hand, you can see that 30% of people were still willing to shock the victim even when they had to see the consequences and physically apply the shocks themselves. (Or, on the bright side, you can see that 40% of people still refused to shock the victim even when doing so would be relatively easy.)

You can even see this happening with the Stanford Prison Experiment. Some of the guards were quite cruel, attacking prisoners with fire extinguishers and imposing long periods of solitary confinement and forced exercise, but others were kinder to the prisoners and even did favors for them. Of course, none tried to end the experiment. But the point is that while the situation can change the likelihood of particular behaviors in groups of people, the chances of any two people exhibiting the behavior can be drastically different. There's something in the disposition.

This is the kind of discussion that has no clean ending. I'll leave you with part of a diary entry I wrote this summer, when I was at the Yerkes Primate Center. The capuchins I was working with were house in groups of fifteen, and in one of the two groups, an intense hierarchy had developed where one family, the G-family, was consistently attacked. One day a particularly intense fight broke out, and it happened to be the same day that Israel began bombing Lebanon.

I went to my desk and checked my e-mail and clicked on BBCnews. When a conflict has existed since before you were born, sometimes it seems eternal. And I can't bring myself to blame anybody, because who do you blame? The individuals on both sides who escalate the conflict - the suicide bombers, the war criminals - they're just individuals. As long as the situation exists, there will always be someone who will react that way to it. Colleen, a grad student in my lab, told me that capuchins in the wild don't act like ours do. The low-ranking capuchins don't get beat up, they just run away. But Goya has nowhere to run to. We do all we can to protect her, shouting at the other monkeys when they go after her, threatening them with the hose, distracting them with peanuts and froot loops. We're her greatest defenders. But we put her in that situation. And we're putting the new G baby into that situation.

But at the same time, right next door you have another colony, existing (relatively) peacefully. We put them in that situation, too.

I guess what I'm trying to say is that you can't change bad human behavior without changing the bad situation. But at the same time, changing the situation won't fix everything. There will always be bad apples.


Nemo85 said...

Thing about the Stanford people is that they knew it was unethical and that it would utterly unpublishable from the get-go. However, they just wanted to see, and the experiment wasn't going to last very long, etc. However, it went so incredibly bad so incredibly fast (which is the remarkable part; it happened in a matter of hours, not days).

Milgram I can see as ethical as the teacher who "discriminated" against anyone not blue-eyed in her class. Yes, it causes discomfort at the time, but all the subjects afterward said they learned something.

Going with the philosophy that if an experiment is of no bennefit to a subject, then it shouldn't be approved of, would eliminate almost all psychological testing. My ERP experiment is of no bennefit to the subject, but it is also of no harm. When the informed consent was given, they were told that the subjects would not be harmed, and were assured that the other guy was being harmed.

In sum, Milgram was less concerned about whether people would harm eachother anonymously; but how much they trust someone with a lab coat who says "I know what I'm doing." It's not weakness, it's compliance. We are trained to believe authority figures, after all, they know more than we do.

I wrote a treatise about this at some point to my dad about how that's how cops stay in business. That's how they can bust parties by standing in the doorway and saying "you know what you did," because we, societally or otherwise, are not trained to say "I may know what I did, but humor me, what did I do?" The abuse of this power, of course, is when the cop then smacks you with a charge of harassing an officer. Etc.

There was a case at McDonald's a few years ago where a guy prank called the place and convinced the manager to to strip search a female employee (aged barely 17) into stripping and performing sexual acts on him. How? By pretending he was a cop on the phone. The prankster (for lack of a better word) wasn't caught, but the two people involved were charged in the crime.

*the above comments are based on information from a special report on ABC in which subjects in Milgram and the guy who did the Stanford Experiment were interviewed.

Shauna said...

Of course, the reason for almost all basic research is not whether it will help the subject, because treatments are way too far in the future, but simply the acquisition of knowledge. Sometimes I wonder if all experimental subjects in basic research should be other scientists who believe in its benefit. The other day I was a subject in Tara's ERP experiment and as part of the procedure she told me there'd be no benefit to me. I was like, "Hell, of course there will be. I want to know the outcome!" Then again, doing all your experiments on scientists would bias things just the littlest bit. ;)

If behavior is caused by the situation, those McDonald's employees should not have been prosecuted. If behavior is caused by the individual, than the subjects in the Zimbardo and Milgram experiments "should" have been prosecuted. Where do you draw the line?

Rachel McKinney said...

This actually falls somewhat in line with a little bit of the metaphysics of persons stuff I've been up to lately: what are the sorts of properties that a person has necessarily, versus the sorts of properties that a person has only contingently? How well does empirical inquiry set us up for investigating this distinction?

One move that I want to make is to say that when it comes to dispositional properties, it is very difficult to apportion causal responsibility between intrinsic features (say, genetically-regulated "bad-appleness") and extrinsic features (say, environmental "bad barrel-embeddness").

Here is an argument from Louise Antony that I want to quickly sketch:

Think of the salt in my saltshaker. This salt has certain dispositional properties: it's solid now, but would dissolve if placed in a pot of water. So the fact that my salt is now solid isn't a necessary property of my salt: it's a contingent property -- the salt is solid because it's placed under certain activating conditions. Dispositional properties are functional properties: salt doesn't have a "dissolving nature" -- it has a nature such that it dissolves when placed in water.

The same is going to be true, I think, for pretty much any human property we can think of. When we say that, for example, "language use is part of the human essence," what we mean is an empirical generalization that every human being is disposed to acquire language under an extremely large range of "normal" human environments. So when we characterize properties as determined by, say, the human genome, what we really mean is "by almost every human genotype under the known range of environmental variation." Because we don't have access to the behavioral aspects of genotypes (rather, only their phenotypic expressions), this creates an epistemic difficulty: by virtue of what are we justified in saying property X is a "natural" property of humans if all we have access to is X's contingent expression in a widely-varied set of activating conditions?

The trick, then, for psychologists, sociologists and other empirical knowledge-workers is, I think, to develop their work to take this kind of radical contingency into account. I mean, I don't know much of anything about nativism or social constructivism as it takes place in your field, but I think the epistemic dimension to these metaphysical "nature vs. nurture" questions gets overlooked all too easily: by virtue of what are we warrented in saying that bad apple-ness is a part of a person's "nature" if necessarily all we have access to is a result of putatively intrinsic features PLUS activating conditions?

Shauna said...

The trick, then, for psychologists, sociologists and other empirical knowledge-workers is, I think, to develop their work to take this kind of radical contingency into account.

But is that really necessary? You feel the need to qualify statements about a behavior as being expressed "by almost every human genotype under the known range of environmental variation" but... I don't. I don't know if I'm speaking as a scientist to a philosopher or as me to you, but if the environmental variation is out of the known range, well, why do we care how the human genotype responds in a situation that will never occur?

Obviously if there's a situation we consider impossible that actually isn't, it's important for us to reach out and consider how the human genotype responds to it. And in order to do that, we need to keep an open mind towards new situations and extreme behaviors. But what I mean to say is, I think focusing on that obscures the importance of a broad range of human behavior.

To bring this back to the topic, when we speak of a person's "nature" we're not talking about inherent badness, even if I used those phrases as a shortcut in my post. A person is not bad, a behavior is bad. (Well, that too can be argued, but for the sake of sanity...) A person's nature is the combination of behaviors they will exhibit in all the possible situations they can be in. Hopefully there are patterns to these behaviors (otherwise us psychologists are in for a headache).

Anonymous said...

Not exactly on topic, but this pioneering 1967 essay by Naomi Weisstein


relied on Milgram's findings, among a lot of other material. Definitely worth reading.

Kuda Boalha said...

I would not say that Stanford Experiment was a complete failure, everyone learned something from it, even if it was that no such experiment should ever take place, and we are still talking about it, are we not?