In the January/February 2007 edition of the Yale Alumni Magazine, I published an article revisiting Stanley Milgram’s famous obedience experiments. That article itself is adapted from my forthcoming book, The Lucifer Effect: Understanding How Good People Turn Evil (Random House, March 2007). For The Situationist, I am breaking that article into several bite-sized posts as part of a larger series of posts on the Situational Sources of Evil.
* * *
Imagine that you have responded to an advertisement in the New Haven newspaper seeking subjects for a study of memory. A researcher whose serious demeanor and laboratory coat convey scientific importance greets you and another applicant at your arrival at a Yale laboratory in Linsly-Chittenden Hall. You are here to help science find ways to improve people’s learning and memory through the use of punishment. The researcher tells you why this work may have important consequences. The task is straightforward: one of you will be the “teacher” who gives the “learner” a set of word pairings to memorize. During the test, the teacher will give each key word, and the learner must respond with the correct association. When the learner is right, the teacher gives a verbal reward, such as “Good” or “That’s right.” When the learner is wrong, the teacher is to press a lever on an impressive-looking apparatus that delivers an immediate shock to punish the error.
The shock generator has 30 switches, starting from a low level of 15 volts and increasing by 15 volts to each higher level. The experimenter tells you that every time the learner makes a mistake, you have to press the next switch. The control panel shows both the voltage of each switch and a description. The tenth level (150 volts) is “Strong Shock”; the 17th level (255 volts) is “Intense Shock”; the 25th level (375 volts) is “Danger, Severe Shock.” At the 29th and 30th levels (435 and 450 volts) the control panel is marked simply with an ominous XXX: the pornography of ultimate pain and power.
You and another volunteer draw straws to see who will play each role; you are to be the teacher, and the other volunteer will be the learner. He is a mild-mannered, middle-aged man whom you help escort to the next chamber. “Okay, now we are going to set up the learner so he can get some punishment,” the experimenter tells you both. The learner’s arms are strapped down and an electrode is attached to his right wrist. The generator in the next room will deliver the shocks. The two of you communicate over an intercom, with the experimenter standing next to you. You get a sample shock of 45 volts — the third level, a slight tingly pain — so you have a sense of what the shock levels mean. The researcher then signals you to start.
Initially, your pupil does well, but soon he begins making errors, and you start pressing the shock switches. He complains that the shocks are starting to hurt. You look at the experimenter, who nods to continue. As the shock levels increase in intensity, so do the learner’s screams, saying he does not think he wants to continue. You hesitate and question whether you should go on. But the experimenter insists that you have no choice.
In 1949, seated next to me in senior class at James Monroe High School in the Bronx, New York, was my classmate, Stanley Milgram. We were both skinny kids, full of ambition and a desire to make something of ourselves, so that we might escape life in the confines of our ghetto experience. Stanley was the little smart one who we went to for authoritative answers. I was the tall popular one, the smiling guy other kids would go to for social advice.
I had just returned to Monroe High from a horrible year at North Hollywood High School, where I had been shunned and friendless (because, as I later learned, there was a rumor circulating that I was from a New York Sicilian Mafia family). Back at Monroe, I would be chosen “Jimmy Monroe” — most popular boy in Monroe High School’s senior class. Stanley and I once discussed how that transformation could happen. We agreed that I had not changed; the situation was what mattered.
Situational psychology is the study of the human response to features of our social environment, the external behavioral context, above all to the other people around us. Stanley Milgram and I, budding situationists in 1949, both went on to become academic social psychologists. We met again at Yale in 1960 as beginning assistant professors — him starting out at Yale, me at NYU. Some of Milgram’s new research was conducted in a modified laboratory that I had fabricated a few years earlier as a graduate student — in the basement of Linsly-Chittenden, the building where we taught Introductory Psychology courses. That is where Milgram was to conduct his classic and controversial experiments on blind obedience to authority.
Milgram’s interest in the problem of obedience came from deep personal concerns about how readily the Nazis had obediently killed Jews during the Holocaust. His laboratory paradigm, he wrote years later, “gave scientific expression to a more general concern about authority, a concern forced upon members of my generation, in particular upon Jews such as myself, by the atrocities of World War II.”
As Milgram described it, he hit upon the concept for his experiment while musing about a study in which one of his professors, Solomon Asch, had tested how far subjects would conform to the judgment of a group. Asch had put each subject in a group of coached confederates and asked every member, one by one, to compare a set of lines in order of length. When the confederates all started giving the same obviously false answers, 70 percent of the subjects agreed with them at least some of the time.
Milgram wondered whether there was a way to craft a conformity experiment that would be “more humanly significant” than judgments about line length. He wrote later: “I wondered whether groups could pressure a person into performing an act whose human import was more readily apparent; perhaps behaving aggressively toward another person, say by administering increasingly severe shocks to him. But to study the group effect . . . you’d have to know how the subject performed without any group pressure. At that instant, my thought shifted, zeroing in on this experimental control. Just how far would a person go under the experimenter’s orders?”
How far up the scale do you predict that you would go under those orders? Put yourself back in the basement with the fake shock apparatus and the other “volunteer” — actually the experimenter’s confederate, who always plays the learner because the “drawing” is rigged — strapped down in the next room. As the shocks proceed, the learner begins complaining about his heart condition. You dissent, but the experimenter still insists that you continue. The learner makes errors galore. You plead with your pupil to concentrate; you don’t want to hurt him. But your concerns and motivational messages are to no avail. He gets the answers wrong again and again. As the shocks intensify, he shouts out, “I can’t stand the pain, let me out of here!” Then he says to the experimenter, “You have no right to keep me here!” Another level up, he screams, “I absolutely refuse to answer any more! You can’t hold me here! My heart’s bothering me!”
Obviously you want nothing more to do with this experiment. You tell the experimenter that you refuse to continue. You are not the kind of person who harms other people in this way. You want out. But the experimenter continues to insist that you go on. He reminds you of the contract, of your agreement to participate fully. Moreover, he claims responsibility for the consequences of your shocking actions. After you press the 300-volt switch, you read the next keyword, but the learner doesn’t answer. “He’s not responding,” you tell the experimenter. You want him to go into the other room and check on the learner to see if he is all right. The experimenter is impassive; he is not going to check on the learner. Instead he tells you, “If the learner doesn’t answer in a reasonable time, about five seconds, consider it wrong,” since errors of omission must be punished in the same way as errors of commission — that is a rule.
As you continue up to even more dangerous shock levels, there is no sound coming from your pupil’s shock chamber. He may be unconscious or worse. You are truly disturbed and want to quit, but nothing you say works to get your exit from this unexpectedly distressing situation. You are told to follow the rules and keep posing the test items and shocking the errors.
Now try to imagine fully what your participation as the teacher would be. If you actually go all the way to the last of the shock levels, the experimenter will insist that you repeat that XXX switch two more times. I am sure you are saying, “No way would I ever go all the way!” Obviously, you would have dissented, then disobeyed and just walked out. You would never sell out your morality. Right?
Milgram once described his shock experiment to a group of 40 psychiatrists and asked them to estimate the percentage of American citizens who would go to each of the 30 levels in the experiment. On average, they predicted that less than 1 percent would go all the way to the end, that only sadists would engage in such sadistic behavior, and that most people would drop out at the tenth level of 150 volts. They could not have been more wrong.
In Milgram’s experiment, two of every three (65 percent) of the volunteers went all the way up to the maximum shock level of 450 volts. The vast majority of people shocked the victim over and over again despite his increasingly desperate pleas to stop. Most participants dissented from time to time and said they did not want to go on, but the researcher would prod them to continue.
Over the course of a year, Milgram carried out 19 different experiments, each one a different variation of the basic paradigm. In each of these studies he varied one social psychological variable and observed its impact. In one study, he added women; in others he varied the physical proximity or remoteness of either the experimenter-teacher link or the teacher-learner link; had peers rebel or obey before the teacher had the chance to begin; and more.
In one set of experiments, Milgram wanted to show that his results were not due to the authority power of Yale University. So he transplanted his laboratory to a run-down office building in downtown Bridgeport, Connecticut, and repeated the experiment as a project ostensibly of a private research firm with no connection to Yale. It made hardly any difference; the participants fell under the same spell of this situational power.
The data clearly revealed the extreme pliability of human nature: depending on the situation, almost everyone could be totally obedient or almost everyone could resist authority pressures. Milgram was able to demonstrate that compliance rates could soar to over 90 percent of people continuing to the 450-volt maximum or be reduced to less than 10 percent — by introducing just one crucial variable into the compliance recipe.
Want maximum obedience? Make the subject a member of a “teaching team,” in which the job of pulling the shock lever to punish the victim is given to another person (a confederate), while the subject assists with other parts of the procedure. Want resistance to authority pressures? Provide social models — peers who rebel. Participants also refused to deliver the shocks if the learner said he wanted to be shocked; that’s masochistic, and they are not sadists. They were also reluctant to give high levels of shock when the experimenter filled in as the learner. They were more likely to shock when the learner was remote than in proximity.
In each of the other variations on this diverse range of ordinary American citizens, of widely varying ages and occupations and of both genders, it was possible to elicit low, medium, or high levels of compliant obedience with a flick of the situational switch. Milgram’s large sample — a thousand ordinary citizens from varied backgrounds — makes the results of his obedience studies among the most generalizable in all the social sciences. His classic study has been replicated and extended by many other researchers in many countries.
Recently, Thomas Blass of the University of Maryland-Baltimore County [author of The Man Who Shocked The World and creator of the terrific website StanleyMilgram.Com] analyzed the rates of obedience in eight studies conducted in the United States and nine replications in European, African, and Asian countries. He found comparably high levels of compliance in all. The 61 percent mean obedience rate found in the U.S. was matched by the 66 percent rate found across all the other national samples. The degree of obedience was not affected by the timing of the studies, which ranged from 1963 to 1985.
Other studies based on Milgram’s have shown how powerful the obedience effect can be when legitimate authorities exercise their power within their power domains. In one study, most college students administered shocks to whimpering puppies when required to do so by a professor. In another, all but one of 22 nurses flouted their hospital’s procedure by obeying a phone order from an unknown doctor to administer an excessive amount of a drug (actually a placebo); that solitary disobedient nurse should have been given a raise and a hero’s medal. In still another, a group of 20 high school students joined a history teacher’s supposed authoritarian political movement, and within a week had expelled their fellows from class and recruited nearly 200 others from around the school to the cause.
Now we ask the question that must be posed of all such research: what is its external validity, what are real-world parallels to the laboratory demonstration of authority power?
* * *
That question I will turn to in Part II of this series (to be posted next week).