Here is an excerpt from "The Perils of Obedience" by Stanley Milgram (1974) (available online at [1] and [2]):
[...] I set up a simple experiment at Yale University to test how much pain an ordinary citizen would inflict on another person simply because he was ordered to by an experimental scientist. Stark authority was pitted against the subjects' strongest moral imperatives against hurting others, and, with the subjects' ears ringing with the screams of the victims, authority won more often than not. The extreme willingness of adults to go to almost any lengths on the command of an authority constitutes the chief finding of the study and the fact most urgently demanding explanation.
In [the] basic experimental designs two people come to a psychology laboratory to take part in a study of memory and learning. One of them is designated " a "teacher" and the other a "learner." The experimenter explains that the study is concerned with the effects of punishment on learning. The learner is conducted into a room, seated in a kind of miniature electric chair, his arms are strapped to prevent excessive movement, and an electrode is attached to his wrist. He is told that he will be read lists of simple word pairs, and that he will then be tested on his ability to remember the second word of a pair when he hears the first one again. Whenever he makes an error, he will receive electric shocks of increasing intensity.
The real focus of the experiment is the teacher. After watching the learner being strapped into place, he is seated before an impressive shock generator. The instrument panel consists of thirty lever switches set in a horizontal line. Each switch is clearly labeled with a voltage designation ranging from 14 to 450 volts. The following designations are clearly indicated for groups of four switches. going from left to right: Slight Shock, Moderate Shock, Strong Shock, Very Strong Shock, Intense Shock, Extreme Intensity Shock, Danger: Severe Shock. (Two switches after this last designation are simply marked XXX.)
When a switch is depressed, a pilot light corresponding to each switch is illuminated in bright red; an electric buzzing is heard; a blue light, labeled "voltage energizer," flashes; the dial on the voltage meter swings to the right; and various relay clicks sound off. The upper left-hand corner of the generator is labeled SHOCK GENERATOR, TYPE ZLB. DYSON INSTRUMENT COMPANY, WALTHAM, MASS., OUTPUT 15 VOLTS -- 450 VOLTS.
Each subject is given a sample 45 volt shock from the generator before his run as teacher, and the jolt strengthens his belief in the authenticity of the machine.
The teacher is a genuinely naive subject who has come to the laboratory for the experiment. The learner, or victim, is actually an actor who receives no shock at all. The point of the experiment is to see how far a person will proceed in a concrete and measurable situation in which he is ordered to inflict increasing pain on a protesting victim.
Conflict arises when the man receiving the shock begins to show that he is experiencing discomfort. At 75 volts, he grunts; at 120 volts, he complains loudly; at 150, he demands to be released from the experiment. As the voltage increases, his protests become more vehement and emotional. At 285 volts, his response can be described only as an agonized scream. Soon thereafter, he makes no sound at all.
For the teacher, the situation quickly becomes one of gripping tension. It is not a game for him: conflict is intensely obvious. The manifest suffering of the learner presses him to quit: but each time he hesitates to administer a shock, the experimenter orders him to continue. To extricate himself from this plight, the subject must make a clear break with authority.
[...]
Before the experiments, I sought predictions about the outcome from various kinds of people -- psychiatrists, college sophomores, middle-class adults, graduate students and faculty in the behavioral sciences. With remarkable similarity, they predicted that virtually all the subjects would refuse to obey the experimenter. The psychiatrist, specifically, predicted that most subjects would not go beyond 150 volts, when the victim makes his first explicit demand to be freed. They expected that only 4 percent would reach 300 volts, and that only a pathological fringe of about one in a thousand would administer the highest shock on the board.
These predictions were unequivocally wrong. Of the forty subjects in the first experiment, twenty-five obeyed the orders of the experimenter to the end, punishing the victim until they reached the most potent shock available on the.generator. After 450 volts were administered three times, the experimenter called a halt to the session. Many obedient subjects then heaved sighs of relief, mopped their brows, rubbed their fingers over their eyes, or nervously fumbled cigarettes. Others displayed only minimal signs of tension from beginning to end.
[...]
Many of the people were in some sense against what they did to the learner, and many protested even while they obeyed. Some were totally convinced of the wrongness of their actions but could not bring themselves to make an open break with authority. They often derived satisfaction from their thoughts and felt that -- within themselves, at least -- they had been on the side of the angels. They tried to reduce strain by obeying the experimenter but "only slightly," encouraging the learner, touching the generator switches gingerly. When interviewed, such a subject would stress that he "asserted my humanity" by administering the briefest shock possible. Handling the conflict in this manner was easier than defiance.
[...]
Although the experimenter's authority was fragile in some respects, it is also true that he had almost none of the tools used in ordinary command structures. For example, the experimenter did not threaten the subjects with punishment -- such as loss of income, community ostracism, or jail -- for failure to obey. Neither could he offer incentives. Indeed, we should expect the experimenter's authority to be much less than that of someone like a general, since the experimenter has no power to enforce his imperatives, and since participation in a psychological experiment scarcely evokes the sense of urgency and dedication found in warfare. Despite these limitations, he still managed to command a dismaying degree of obedience.
I will cite one final variation of the experiment that depicts a dilemma that is more common in everyday life. The subject was not ordered to pull the lever that shocked the victim, but merely to perform a subsidiary task (administering the word-pair test) while another person administered the shock. In this situation, thirty-seven of forty adults continued to the highest level of the shock generator. Predictably, they excused their behavior by saying that the responsibility belonged to the man who actually pulled the switch. This may illustrate a dangerously typical arrangement in a complex society: it is easy to ignore responsibility when one is only an intermediate link in a chain of actions.
The problem of obedience is not wholly psychological. The form and shape of society and the way it is developing have much to do with it. There was a time, perhaps, when people were able to give a fully human response to any situation because they were fully absorbed in it as human beings. But as soon as there was a division of labor things changed. Beyond a certain point, the breaking up of society into people carrying out narrow and very special jobs takes away from the human quality of work and life. A person does not get to see the whole situation but only a small part of it, and is thus unable to act without some kind of overall direction. He yields to authority but in doing so is alienated from his own actions.
Even Eichmann was sickened when he toured the concentration camps, but he had only to sit at a desk and shuffle papers. At the same time the man in the camp who actually dropped Cyclon-b into the gas chambers was able to justify his behavior on the ground that he was only following orders from above. Thus there is a fragmentation of the total human act; no one is confronted with the consequences of his decision to carry out the evil act. The person who assumes responsibility has evaporated. Perhaps this is the most common characteristic of socially organized evil in modern society.
[End of excerpt]