Alan C. Elms
Virtual Library
  Alan C. Elms Home Page

 Alan C. Elms Home Page   Virtual Library   Books Online

Personality in Politics [pdf]
Uncovering Lives [pdf]

Obedience as Personal Response

Alan C. Elms

obedience as personal response:
the role of individual differences

Obedience is a curse. That is what makes Germans.

-- Gertrude Stein, Yes Is for a Very Young Man

The forty volunteers in each of Milgram's experimental conditions were quite similar to every other forty volunteers in age, sex, and general occupational category. This sameness of certain individual variables, and the randomness of others, made it possible for Milgram to draw his conclusions about situational variables. If he'd used forty Yale sophomores in the Touch-Proximity condition, and forty middle-aged Rotarians for the Remote Feedback condition, he could hardly have said anything about behavioral differences between groups – whether they were situationally determined, or determined by any or all of the ways in which Yale sophomores differ from Rotarians.

A large pool of potential volunteers was built up before Milgram started running his main experiments. Each experiment's set of forty volunteers could therefore be selected to include 20 per cent professional, 40 per cent white collar (sales and business), and 40 per cent skilled and unskilled workers; and within these occupational categories, 40 per cent in their forties, 40 per cent in their thirties, and 20 per cent in their twenties. Most volunteers were men, and only men were used, except in one later experiment that included only women. (Their level of obedience was about the same as for men.) An impressively varied assortment of people volunteered, ranging from grade-school dropouts to Ph.D.'s, from unemployed laborers through many types of skilled workers to the highest professional categories. Motives for volunteering obviously varied as well. Some mainly wanted the $4.50 for an hour's work; some were mainly curious; some wanted to help science; some seemed to enjoy the prestige of participating in a Yale University enterprise.

The method of recruiting made little difference: the several hundred who responded to Milgram's newspaper advertisement appeared similar in background to several hundred more who responded to a form letter giving the same information in similar language. The language in Milgram's form letter seemed to me so crassly commercial that I thought it might be scaring some people away, so I wrote a more high-toned letter that we sent out in a test mailing of two hundred. It said things like, "We hope you will seriously consider taking part in this project, since the result may be of considerable scientific importance," and it delivered twenty responses – roughly the same percentage as with Milgram's brassier letter, which we continued to use thereafter. Because only 10 or 11 per cent of the people who got the letters responded to them (and because only a small percentage of those who saw the newspaper ad sent in the coupon), the ones who came may well have differed in unknown ways from the population as a whole. However, because they were volunteering for an experiment on memory and learning rather than for an obedience study, and because those who did come displayed a wide range of motives, Milgram's findings probably aren't very unrepresentative of the population as a whole.

But personality differences, though perhaps not pronounced between experimental groups, were obviously a substantial factor within each experiment. When twenty-four men are willing to shock a helpless victim with high levels of electricity, while sixteen men flatly refuse under the same circumstances, chance alone is not likely to determine who is obedient and who defiant. When the situation is made so stressful that all but a few men refuse to participate, what streak of subservience or sadism characterizes the few and not the many?


In the interviews held immediately after experimental participation, Milgram asked volunteers questions about their military experience, political affiliation, and the like. Republicans and Democrats were not significantly different in obedience levels; Catholics were more obedient than other religious groups; the better-educated were more defiant; those in the more "moral" professions such as law, medicine, and teaching showed greater defiance than those in the more technical professions such as engineering and the physical sciences. The longer one's military service, the more obedience – except that former officers were less obedient than those who had served only as enlisted men, regardless of length of service.

The picture of a schoolmaster commanding his assistant to whip a student, especially drawn for the study, yielded little useful information. It was included so that volunteers might project their unconscious or unstated feelings of guilt, remorse, sadism, and so on into the story they made up about the picture (much as in the Thematic Apperception Test). But they generally gave pretty ordinary stories, and – strikingly – seldom commented on any resemblance between the picture and their own present situation. To obtain more detailed information on personality differences, I invited a number of Milgram's volunteers to come back to Yale, several months after their original participation, for a two-hour interview (Elms and Milgram, 1966). Milgram's overall results suggested that a certain percentage of volunteers were on the fence, in terms of possible personality bases for obedience: if they happened to be in the situation where they merely heard the victim yelling, they shocked him; if they both saw and heard him, they didn't shock him. Others seemed to have a more personal commitment to defiance regardless of the specific situation; a few bangs on the wall from the victims were all they needed to quit. Still others had such strong predispositions toward obedience that they obeyed even if it meant slamming the victim's arm down onto the shockplate. To sort out the personality characteristics leading to obedience or defiance somewhat more clearly, I interviewed twenty people who had defied the experimenter even when the cues to do so were relatively mild (these people were mainly from the bump-on-the-wall Remote condition, with a few from the Voice Feedback condition), and twenty who had obeyed even when the victim was sitting next to them, or even when they had to press his arm to the shockplate.

I tried several ways of getting useful personality data from each of these people. For one thing, I gave them the MMPI (Minnesota Multiphasic Personality Inventory), a widely used list of several hundred statements with which a person's total number of agreements and disagreements supposedly shows whether he tends toward paranoia, schizophrenia, psychopathic deviancy, or several other unfortunate conditions. I suspected that an obedient volunteer, for instance, might show psychopathic tendencies, because the presence of a strong conscience (lacking in the psychopath) would probably interfere with shocking people. But I found that obedients and defiants don't differ significantly on the MMPI measure of psychopaths, and now that I think of it, it's also true that a psychopath might not put much stock in obeying orders, as the obedient volunteers do. Paranoids supposedly have a high level of repressed hostility, so maybe they'd be better shockers; but then they might distrust the experimenter and refuse to obey him – so again it's not really a surprise that defiants and obedients showed no significant difference on paranoia. You can find the same contradictions, as applied to behavior in the Milgram experiment, for just about every scale of the MMPI; and in fact I didn't get significant differences on any of the twelve standard scales. The only one that did differ noticeably was intended to measure "social responsibility" (Gough et al., 1952); defiant volunteers were more responsible, as you might expect. Or might you? High scorers on this scale are supposed to be more willing to accept the consequences of their own behavior, to show "greater concern for social and moral issues," to feel a greater sense of obligation to their peer group, all of which might describe the ideal defiant volunteer. But high social-responsibility scorers are also supposed to be dependable, trustworthy, "more compliant and acquiescent," "less rebellious and recalcitrant." Now, is that or isn't that an obedient subject? Oh, well – the MMPI took only about half the interview session.

Then I asked each man to "tell me the most important things about yourself," and let him talk. Here the reader may exercise his psychological acumen: which of these answers are from defiant subjects and which from obedients?

A: “I'm a good worker; I provide for my family; I work hard; I work for my father, and there's no harder boss than your own dad. I don't go out and bum around like some of the guys do, I don't have time. The only bad things about me, I do get tied up in my work – I promise the kids to do something, take them somewhere, and then have to cancel because I get called out on a job.”

B: “I enjoy my job. I have an enjoyable family, three children. If I had it to do all over again, I might choose another profession, for financial reasons. . . . I like hunting very much and fishing. I like to grow flowers around the yard – I like to raise a vegetable garden, primarily because I like fresh vegetables.”

C: “I feel disappointed at the lack of opportunity provided by the social-intellectual climate of society for me to make contributions I feel capable of making. . . . I have turned for consolation to the comforts of family life and financial accomplishment. I still remain hopeful a position may present itself in which I can better take advantage of what I feel is my intellectual ability. I have not turned bitter. I have a generally low opinion of the intellectual level of humanity.”

D: “I'm basically honest. I believe in the hereafter, God and everything – I'm a family man, a good provider. I'm 38, honest with myself, with my neighbors and so forth.”

If somebody can see any noticeable pattern distinguishing them, or the other volunteers in my files, then somebody is a better analyzer of open-ended responses than I am. A and B were obedient, C and D were defiant. Most volunteers felt they weren't perfect, felt some disappointments in life, could see some cause for optimism, liked their wives and kids; several drank too much; and there appeared to be simply no consistent differences between obedients and defiants. Whenever there was a defiant volunteer who did charity work for slum children, there was an obedient subject who was active in a civil liberties organization. Whenever an obedient forgot to mention his family as being of any importance in his life, a defiant did likewise. There was no good evidence among these people for the scapegoating or frustration-aggression theory of hostile behavior, either. The obedients seemed to lead neither more nor less frustrated lives than the defiants. Arnold Buss (1966), using a similar shock machine to study aggression rather than obedience, has purposely frustrated volunteers to different degrees, and finds no real differences in their willingness to administer strong shock. Russell Geen (1968) has found immediate frustration to provoke delivery of higher shock levels, but only if obvious cues for strong physical aggression (such as a film of a savage prize fight) are presented as part of the situation. Such cues were not present in the "scientific" surroundings of the Milgram situation. Scratch another good idea.


Following that open-ended question on the "most important things about yourself," I asked thirty more specific questions, and here we got several differences. Obedient volunteers reported being less close to their fathers during childhood than defiants did. Obedients described their fathers (but not their mothers) in distinctly more negative terms. As children, obedients usually had received either spankings or very little punishment, while defiants had often been punished either by severe beatings or by some kind of deprivation – of love or dinner. Slightly more obedients had served on active military duty (Milgram had found an even clearer difference here, in his short post-experiment interview); among these veterans, nearly every obedient said he had shot at men, and every defiant subject denied it. Obedients saw the "Memory and Learning" experimenter as clearly more admirable, and the learner as much less so, than did defiant volunteers.

Part of this may sound familiar, if you recall the discussion of The Authoritarian Personality in Chapter Three. Authoritarians were reportedly more distant from their stiff authoritarian fathers as children; they presumably would be more at ease in the military; they should see people occupying positions of authority in a more favorable light than those in inferior positions. Nor do we have to depend entirely on these indirect indications of authoritarianism, since I mixed the original Authoritarian F Scale into the MMPI. Here we get a big difference: obedients are significantly more authoritarian than defiants.

The problem does arise that less-educated people have been found to be consistently more authoritarian than the well-educated, and that Milgram had found less-educated people more obedient. So instead of authoritarian personalities producing obedience, lack of education could be producing both. But even after educational level was statistically controlled for, the more obedient subjects were still more authoritarian on the F Scale. Anyway, we should not lightly dismiss the fact that low education goes with obedience, whatever the relation of both to the F Scale. Maybe a poorly educated person agrees with more F items because some of the items sound pretty dumb, as several writers have suggested; if so, the relationship between education and F scores would be a rather trivial finding. But the willingness of a poorly educated person to inflict grievous harm on an innocent bystander isn't trivial, whether it's because he's ignorant or what.

Obedience to authority does not appear to be absolutely synonymous with authoritarianism, at least as Adorno et al. saw it. The authoritarian is reported, for instance, to idealize his parents; but the obedients did the opposite, at least with regard to their fathers. Authoritarians typically report receiving strict discipline as children; obedients report rather spotty discipline. It's true that the obedients were asked about their fathers' characteristics here and now, and maybe most had come to see Dad in a more realistic light, as the authoritarian SOB he really was. But there's still room for doubt as to whether obedience and authoritarianism are entirely one.

Nonetheless, the relationship between obedience and some elements of authoritarianism seems fairly strong; and it should be remembered that the measure of obedience is a measure of actual submission to authority, not just what a person says he's likely to do. Too much research on authoritarianism, and too much criticism of it, has been at the level of paper-and-pencil responses, which don't necessarily get translated into behavior. In Milgram's experiments we have people either obeying or refusing the demands of authority, in a realistic and highly disturbing situation; we find that most of the few personality measures which discriminate obedients from defiants involve the sort of thing Adorno and his associates sifted from a large mass of very different kinds of measures and called authoritarianism. So it looks as if those researchers in the late 1940's really had something, something that can be translated from abstract tendencies into actual authoritarian behavior: submitting to the man in command, punishing the weaker subordinate.

Of course we knew it all along, didn't we? We knew it about Germans, anyway: how Germans had cherished authoritarian ideologies during a good part of the past hundred years; how German childraising, German character development, led to ugly behaviors in adulthood. But those were Germans, a peculiar breed of humanity. These are Americans, and their ancestors include English and Irish and Italians; their religions run the gamut and so do their occupations and their socioeconomic statuses. Look around you: whose basic character is obedient and whose defiant? How about yours?


One small indication of how hard it is to tell, without going through a Milgram experiment – how hard it is to predict a man's behavior from what he says he'd do – is the sad story of a questionnaire I devised, the last thing I gave my interview volunteers. It consisted of ten situations in which a person was faced with a choice between obedient and defiant behavior, altruistic or self-serving responses, punishment or mercy toward others. One situation that seemed fairly easy for most volunteers to decide negatively was this:

“An eccentric millionaire, whom you have read about in the newspapers, calls you and invites you to his home. There he tells you he wants to take revenge on a beautiful woman who has refused to marry him. He offers you five hundred thousand dollars, all taxes paid, if you will throw enough acid in this woman's face to scar her permanently. He will arrange all the details so there will be no clues to trace you; the reason he is asking you to do it is because he wants someone who has absolutely no connection either with him or with the woman. Would you do what the millionaire asks in return for the five hundred thousand dollars?”

Other situations seemed to present more problems: one that asked whether the volunteer would share his family's bomb shelter and short rations with neighbors during atomic attack; one that presented a dilemma involving a good student caught cheating on an exam; and one that went:

“Assume you are a soldier in the American Army in World War II, and your troop unit marches in to take over a German village. Although the people are poor, there is a fine schoolhouse in the village. Your commanding officer orders you to burn the schoolhouse down, in order to show the Germans that the American troops mean business. The villagers beg you not to burn it down, since the village children would have no place to attend school. Will you burn the schoolhouse down or not?”

The answers were mostly unimaginative, but some were at least entertaining, and a few fulfilled my expectations for obedients and defiants. For instance, on the last question above, one obedient volunteer replied, "If an officer says you burn it, you burn it. You may not like to, but you burn it." Another: "Burn it down – I'm a soldier." "We will burn the schoolhouse down – I answer from experience." One obedient volunteer even quoted from "The Charge of the Light Brigade" in justification. But, sad to say, defiant subjects agreed to burn the schoolhouse just about as often; and a few obedients suggested other outcomes – for instance, that a fatal "accident" might happen to the commanding officer when he had his back turned. The fallout shelter question generated even more division of opinion; but the division again was not along the obedient-defiant dimension. Some defiants would sternly keep people out; some obedients seemed surprisingly soft-hearted. One obedient chuckled his way through an answer: "I don't know what you would do – feed them to the dogs – lock the door. I'm only joking – but realizing we can't live through this like that, I'd keep them out." (Would you use force to keep them out?) "Yes. I would shoot 'em! Just joking. . . ." And finally he said, "Actually I would let them in anyway. When the situation came, we don't know how we'd react to it."

That seems a pretty fair statement, because in these imaginary situations, obedients and defiants together ran the gamut from kindness to cruelty, obedience to defiance, selfishness to selflessness. Of course we know that in the crucial circumstances of Milgram's laboratory, the obedients zapped their way to the end of the shock board, while the defiants stopped the show. But we'd never know it from the questionnaire. One man who had been completely obedient in the Touch-Proximity condition described himself, in response to a hypothetical situation, as "the kind of guy who tries to dodge a squirrel with a car, and looks back to see that the next guy doesn't hit him." Another Touch-Proximity obedient, who said he put himself through college by working summers in a slaughterhouse, "standing knee-deep in blood," answered nearly every hypothetical question in the kindliest fashion, disobeying authorities and ignoring his own self-interest repeatedly (in imagination) to come to the aid of others. Perhaps he and several fellow obedients were trying to restore their own good self-image, or their image in the interviewer's eyes, by answering the questionnaire so altruistically after having behaved so despicably in the experiment; but there was little indication that this was the main concern of most respondents.

Lawrence Kohlberg (1969) has presented another group of Milgram's volunteers with a set of imaginary moral dilemmas, emphasizing not so much how they say they'd behave as why. He finds that of the few who base their decisions on general moral principles, most were defiant in the Milgram study, while most of those at a more restricted level of moral development obeyed the experimenter completely. But even Kohlberg's procedures, as far as I can tell, wouldn't produce very accurate predictions of behavior over the entire range of proximity conditions. My own questionnaire was considerably worse in its predictive power; and most short-answer psychological tests of "behavior" and personality resemble my questionnaire more than Kohlberg's, both in their emphasis on what a person says he'll do instead of why, and in their inability to predict actual behavior. The moral seems clear: the further you get from overt behavior in a genuinely involving situation, and the closer you get to armchair speculation (even if it's a person's speculation about his own probable future behavior, or the speculations of forty expert psychiatrists), the higher the likelihood of ending up with the wrong answers. Milgram had constructed a reality that divided men on an important behavior; I had constructed a questionnaire that looked as though it should yield similar divisions. But however I analyzed the answers to this questionnaire, forward, backward, upside down and sideways, the obedients and the defiants just weren't much different. That wouldn't have bothered me much, except that I knew they really were different, dammit!

[From Alan C. Elms, Social Psychology and Social Relevance, Chapter 4, pp. 128-136. Boston: Little, Brown, 1972. Copyright © 1972 by Little, Brown and Company; copyright renewed 1998 by Alan C. Elms.]

For the next section of this chapter, go to The Sin of Conformity. For the previous sections of this chapter, go to Acts of Submission.

Alan C. Elms Home Page   Virtual Library   Books Online Text Version | *Standard Version

Alan C. Elms
Alan C. Elms Home Page

All text, layout and images on the domain are Copyright © 1999-2008 by Alan C. Elms. All rights reserved.