milgram experiment historischer hintergrund

Philosophia Scientiæ

Travaux d'histoire et de philosophie des sciences

Home Numéros 28-2 Stanley Milgram’s Obedience Studi...

Stanley Milgram’s Obedience Studies: A Critical Review of the Most Influential Explanatory Accounts

In 1960 Stanley Milgram wondered if ordinary people would, as many Germans did during the Holocaust, obey higher orders to harm innocent people. Soon after, he ran the New Baseline experiment: at the behest of a scientific authority, 65 percent of subjects inflicted what appeared to be potentially lethal shocks on another person with a mild heart condition. Although, to date, there is no widely accepted explanation that can account for this finding, three theories continue to attract a disproportionate amount of attention: the Incredulity Hypothesis, Agentic State , and Engaged Followership . The aim of this paper is to present an overview of how Milgram invented his basic procedure and then use the insights gained from his journey of discovery as a foundation from which to critically review these three explanatory accounts. Succeeding this critical review, a lesser well-known theory will be presented; one that is supported by Milgram’s original results.

En 1960, Stanley Milgram s’est demandé si des personnes ordinaires obéiraient à des ordres leur prescrivant de nuire à des innocents, ainsi que de nombreux Allemands l’avaient fait pendant la Shoah. Pour répondre à cette question, il conçut une expérimentation à l’issue de laquelle, à la demande d’une autorité scientifique, 65 % des sujets ont infligé ce qui semblait être des chocs potentiellement létaux à une autre personne atteinte d’une pathologie cardiaque légère. À ce jour, aucune explication de cette découverte ne fait consensus. Trois théories continuent toutefois d’accaparer l’essentiel de l’attention : l’hypothèse de l’incrédulité, l’hypothèse de l’état agentique et l’hypothèse du suivisme engagé. L’objectif de cet article est de présenter un aperçu de la manière dont Milgram a conçu sa procédure de base, puis d’utiliser cette connaissance de la préhistoire de sa découverte comme base pour examiner de manière critique ces trois explications. À la suite de cette revue critique, une théorie moins connue sera présentée, laquelle est corroborée par les résultats originaux de Milgram.

The authors would like to thank Emeritus Professor Robert Gregory of Victoria University of Wellington, New Zealand, for his comments. All responsibility rests with the authors.

A piece of literature makes its way into canon based largely on the meaning it imparts in our lives. Milgram’s experiments are indisputably in the canon. And yet, no one can agree on the theme [...] What message has Milgram sent us, in what sort of bottle, on which sea? Lauren Slater [2004, 61]

1 Introduction

1 In 1960, the American social psychologist Stanley Milgram, both inspired and terrified by the Holocaust, wondered if ordinary (American) people would follow a higher authority’s orders to harm an innocent person. In his New Baseline experiment, undertaken at Yale University, the subject was informed by an actor playing the role of a social science “experimenter” that they were to participate in an experiment purportedly exploring the effects of punishment on learning. The subject was allocated the role of “teacher” and was taken to a small room where they observed the experimenter firmly strap another actor fulfilling the role of “the learner” into an electric chair. In conversation, the learner warned all present that they had a mild heart condition. After reassuring the learner that although the shocks “may be painful they’re not dangerous,” the experimenter, with subject in tow, left the small room and entered a larger adjoining room where the subject was seated in front of a shock generator: an electrical device with 30 switches that increased in 15-volt increments ranging from 15 to 450 volts.

2 The experimenter then instructed the subject to ask the learner a list of word pair questions, which the learner purposefully answered incorrectly with regularity. The experimenter instructed the subject to increase the intensity of the “punishment” by 15 volts for every incorrect answer received. No shocks were really administered, but if the subject hesitated to do as they were told, the experimenter urged them to follow his shock-inflicting instructions by deploying a set of increasingly insistent prods. If the subject complied, at the 150-volt switch, the learner exclaimed his heart was bothering him. During the second half of the experiment, the learner complained further about his heart and then started engaging in agonized screams until, across the last third of the remaining shocks, he went silent, implying something had gone terribly wrong. To this, the experimenter urged the subject to treat all subsequent unanswered questions as incorrect and continue inflicting further shocks. Milgram’s research question was: would the subject comply or defy the experimenter’s insistence that they continue to inflict what appeared to be excruciating shocks on a vulnerable person?

3 Essentially, this procedure forced subjects into making a choice: side with the learner and stop inflicting painful shocks, or side with the experimenter and collect further data by inflicting more shocks. The subject could not simultaneously appease both parties. They were faced with resolving a dilemma: was it more important to help the experimenter or the learner? Milgram argues this dilemma has a moral dimension: when observers of the experiment are asked “what constitutes appropriate behavior in this situation, they unfailingly see disobedience as proper” [Milgram 1974, 6]. Disobedience was proper because of the experimenter’s “[...] fundamental breach of moral conduct to hurt another person against his will” [1974, 41]. For outsiders looking in, the choice is obvious: it is more important to help an innocent person with a mild heart condition avoid intense shocks than it is to collect data. But herein lies the attraction of laypersons and scholars alike to this research: refusing to inflict shocks appears to be the morally correct course of action, yet the New Baseline experiment showed that 65 percent of subjects instead chose to inflict what they were led to believe were three consecutive 450-volt shocks on an innocent person. Thus, an intriguing question arises: why did most subjects repeatedly pursue what outsiders unanimously agree was the wrong choice?

4 Over the last half century or so, a range of theories purporting to explain Milgram’s counter-intuitive results have been proposed. Although some explanatory accounts have generated a greater following than others, scholars specialising in the Obedience Studies have failed to converge in support of any one of them. That said, the central aim of this paper is to present and critically review the three most prominent explanatory accounts for Milgram’s baseline results: the Incredulity Hypothesis , the Agentic State , and Engaged Followership . The final section of this article will briefly present a less known theory, the validity of which, it will be argued, is bolstered by Milgram’s own results. Before presenting these theoretical accounts, however, it is important to note that one potentially fruitful strategy to better understand the inner workings of the Obedience Studies is to delineate their start-to-finish invention. Doing so, we believe, will provide a strong base from which to critically assess the quality of all these theoretical accounts. Thus, what follows will first present an abridged overview of Milgram’s inventive journey of discovery.

2 Invention and key findings

5 During the Nazi war crimes trials, many ordinary Germans justified their harmful actions by arguing that they were just following higher orders. Milgram therefore wondered what would happen if he ran a social science experiment where an authority figure instructed ordinary people to follow orders to hurt another person. For this experiment to garner scholarly attention, Milgram knew it needed to obtain eye-catching results—nobody would be surprised by an experiment that generated a low rate of harmful obedience. He therefore set out with a preconceived goal: the official baseline experiment had to “maximize obedience [...]” [Milgram cited in Russell 2011, 158]. Milgram, however, did not have an experimental procedure capable of producing this result, so he had to invent one.

6 His first attempt at inventing a procedure capable of maximising obedience was rudimentary. Initially he planned to hire an actor to play an authority figure who, for some reason, required the subject to obey his commands to physically assault another person. As Milgram noted in one document (circa 1960):

We can then start out by giving the subject commands from the lower end. (Tap him.) And gradually proceed to more intensive commands. (Slug him.) Some kind of scale of this sort is needed. A second requirement [...] is to give the command some institutional justification. [Milgram cited in Russell 2018, 48]

7 Inherent in this idea is the manipulative “foot-in-the-door” technique: persons are more likely to agree to a significant request if it is preceded by a comparatively insignificant request [Freedman & Fraser 1966]. Milgram later termed his inclusion of manipulative techniques like this as Binding Factors, which are powerful bonds that can entrap a person into engaging in actions they would prefer not to do [1974, 148]. But were “intensive commands” to engage in an aggressive and intensifying “scale[d]” physical assault likely to maximise obedience? Two problems with this idea probably became apparent. First, putting aside the ethical difficulties of running an experiment where a person was physically beaten(!), surely subjects would feel too closely connected to the order’s harmful end result. Second, subjects would surely need a reason—an “institutional justification”—for instigating the assault. Subsequent procedural changes indicate Milgram likely sensed then resolved both problems. That is, Figure 1 (below), a document written by Milgram, indicates subjects were to harm another person by inflicting intensifying electrical shocks. This document also identifies the “institutional justification” Milgram intended to use to encourage harm-infliction: before starting, subjects were to accept a “pledge to obey”:

Fig. 1: A document from Milgram’s personal archive titled “Studies in Obedience...” [cf. Russell 2018, 49]

8 The contents of this document are as follows:

Studies in Obedience

• Wa[i]ver of responsibility—from experimenter—For Germa[n]y
• Panel [Sketch of shock generator with the final shock labeled “LETHAL”]
• The War Situation—2 naive S[ubject]’s. One must shock the other – 1 way switch. Can be controlled by E[xperimenter].
• Working in teams:
• The Pledge. Subjects pledge to obey. Because of certain possible hazards, the S. must adhere carefully to the instructions of the Exp[erimen]t[e]r

9 In this document Milgram seems to have incorporated into his emerging procedure what he understood to be the Nazi regime’s most effective techniques of manipulation: displacement of responsibility, hierarchical obedience to authority, a pledge to obey harmful orders, taking small steps toward a radical outcome, all in the context of war.

10 But the document’s most eye-catching feature is Milgram’s idea to utilize a shock generator as the means of harm infliction. This device was an example of what he later termed a Strain Resolving Mechanism. Strain Resolving Mechanisms help reduce the stress people normally experience when inflicting harm [Milgram 1974, 153–164]. That is, a person who presses switches that harm another person will experience less stress than had they physically beaten them with their fists.

11 However, one weakness with his above research idea was that surely the “pledge to obey” was too closely associated with the Nazi regime whereby Milgram’s inclusion of it risked alerting subjects to the study’s actual research question. To increase the research idea’s viability, Milgram required a more veiled justification for harm infliction. Indeed, by October 1960 Milgram had discarded his pledge to obey in favour of an more opaque—better disguised— justification for harm infliction: the subject, in the role of a “teacher”, was to inflict shocks on a “learner” to help determine if punishment improved learning [Russell 2018, 56–59]. In doing so, however, Milgram added another Binding Factor to his procedure: to help advance scientific knowledge on the topic of punishment’s effect on learning, subjects would be pressured into inflicting every shock. With his shock machine and a less transparent justification for shock infliction, Milgram envisioned two essential components that moved him closer to preconceived goal achievement.

12 To see if his inchoate procedure worked, Milgram’s students ran the first pilot trial and to his astonishment, most subjects inflicted every shock. Despite his surprise, Milgram remained critical of his student’s “not very well controlled” pilot [Milgram cited in Blass 2004, 68]. More specifically, because of the students’ disjointed procedure, fake-looking shock generator, and weak acting skills [Russell 2018, 64], Milgram suspected some subjects completed the experiment because they did not believe the learner was being harmed. Milgram sought to overcome these weaknesses [Blass 2004, 75], [Russell 2018, 68, 76].

13 It also transpires that Milgram used the pilots as a means of increasing the emerging procedure’s efficacious power. He closely observed the defiant pilot subjects’ strategies to prematurely ending the experiment and, after doing so, he then tried to envision counter-escape strategies—binding or strain-resolving techniques—that the experimenter could deploy against future subjects who might try to use similar tactics. Doing so better ensured that during the official experiments reluctant subjects were more likely to continue inflicting further shocks. One example of Milgram utilizing this technique comes from a “Report on Pilot Res[e]arch” dated 4 December 1960, where he noted some subjects informed the experimenter it was the voluntary learner’s prerogative to “leave whenever he wants to” [Russell 2018, 64]. Thus, these pilot subjects’ awareness of a person’s right to withdraw from an experiment emboldened them to resist the experimenter’s demands that they continue. To block off this potential escape route, Milgram—clearly drawing on an idea mentioned in Figure 1—proposed “the following change”: the experimenter should respond to such future resistance by stating “I [h]ave responsibility for this situation [...]. Proceed with the next question” [Russell 2018, 65]. By the official experiment this statement was changed to “I’m responsible for anything that happens to him. Continue, please” [Milgram 1974, 74]. Milgram’s change here resulted in the addition of another Strain Resolving Mechanism to the emerging procedure: if future subjects accepted the experimenter’s offer, they could, if they so chose, displace all responsibility for shock infliction on to him (thereby potentially reducing their own strain over shock infliction).

14 Milgram’s varied techniques of discovering, then adding an array of Strain Resolving Mechanisms and Binding Factors to his emerging procedure—many of which are not mentioned here [see Russell 2011, 2018]—in part explains how he managed to construct a powerful official baseline procedure. Basically, Milgram—in the role of a goal-directed social engineer—kept adding Binding Factors and Strain Resolving Mechanisms to the emerging procedure, which cumulatively helped him achieve his preconceived goal, that is, maximization of the official baseline’s completion rate. The high completion rate of his final pilot trial signaled to Milgram he was finally ready to run the first official baseline. He then ran his first official experiment: the Remote Feedback condition, where the learner, who could not be seen or heard, suddenly kicked the adjoining wall at the 300-volt switch and thereafter went silent. This experiment obtained a 65 percent completion rate. Soon after running this first baseline, Milgram lost access to his laboratory space and was moved to a new location (which put pressure on him to start over). Around the same time, he came to suspect that in the newly allocated laboratory he could, on starting over, probably run a new yet far more disturbing (impressive?) baseline condition, which he suspected would still obtain a high completion rate. The New Baseline (or “cardiac” condition), which, as earlier described, involved a screaming learner distressed by his heart condition. The New Baseline also obtained a 65 percent completion rate.

15 After Milgram had achieved his preconceived goal of obtaining a high baseline completion rate, he then pursued his strategy for theoretical development. That is, to discover why most subjects completed the New Baseline , he planned to run a score or so of slight baseline variations, against which he could test both earlier subjects’ post-experimental justifications for completing and the accuracy of various hypotheses as they came to him. Across the many slight baseline variations, he planned to test and eliminate these justifications and hypotheses until the most accurate account emerged. For example, some subjects claimed they only completed the experiment because it was conducted at Yale University—a respected research institution. To test the validity of this explanation, Milgram ran the Institutional Context condition: a baseline replication run in a shabby Bridgeport office under the fictitious title of Research Associates of Bridgeport . This slight variation, however, obtained a 47.5 percent completion rate; a result “not significantly lower than that obtained at Yale”, thus eliminating this potential explanation for the baseline result [Milgram 1974, 69]. Milgram also wondered if inherent human aggressiveness was responsible for the baseline result. Consequently, he ran a variation termed the Subject Chooses Shock Level condition where subjects were instructed to shock at any intensity of their choosing. Most subjects—97.5 percent—chose to inflict low-intensity shocks. Because nearly all subjects, on their own volition, chose not to seriously hurt the learner, inherent human aggressiveness did not generate the high baseline completion rate.

16 The theory Milgram came to favour was that ordinary people seem to have a strong proclivity to obey what they believe to be legitimate commands. He favoured this account because when subjects were given orders to inflict shocks, as in the New Baseline , 65 percent completed the experiment. But when no orders were given, as in the Subject Chooses Shock Level condition, only 2.5 percent completed the experiment. The 62.5 percent difference between both experiments bolstered Milgram’s presupposition that higher orders seemed to be centrally involved in generating the baseline result.

17 However, some of the results from his baseline variations conflicted with this conclusion. One of several possible examples was the Teacher in Charge condition. Before the start of this variation the experimenter stated that he had to leave the laboratory, but as he was departing, he told an actor-teacher to invent a punishment regime, which the other subject-teacher was to implement. The actor-teacher decided to increase the intensity of punishment by 15 volts for each incorrect answer, and then instructed the subject to fulfill the role of shock inflictor. In this condition, 55 percent of subjects did as the man of equal status told them and inflicted every shock. Milgram was confused: “we were surprised aond [ sic ] somewhat troubled that 11 [out of 20 subjects] should be full[y] obedient under common man authority” [Milgram cited in Russell 2018, 99]. Milgram was troubled because this result conflicted with his favourite theory that obedience to an authority caused most subjects to complete the New Baseline condition. Thereafter, Milgram’s confidence in developing a comprehensive theory of obedience to authority declined [Milgram cited in 2018, 99]. In short, Milgram was baffled by his occasionally contradictory baseline variation results, which saw him, across most of the 1960s, abandon his goal of developing a comprehensive theory [2018, 101].

18 In December 1961 Milgram submitted for consideration his first Obedience Study journal article. This paper described and presented the results of the Remote-Feedback condition (the first baseline he, as mentioned, later rejected in favour of the New Baseline ; whereby the latter became the procedure that all subsequent slight variations were modeled upon). However, because in this article Milgram failed to provide a concise theoretical account for the result, reviewers twice recommended its rejection [Blass 2004, 114], with one, Edward E. Jones, adding that the 65 percent completion rate was, in support of the above overview of Milgram’s inventive journey, at best a “triumph of social engineering” [Parker 2000, 112, cited in Russell 2018, 72]. With no theory, Milgram [1963, 377] then seems to have appeased his reviewers by adding 13 “features” inherent to the procedure that he thought helped explain why most subjects completed the experiment. In 1963, this article was accepted for publication. Milgram’s research journey shows he preferred—probably with concision and simplicity in mind—a mono-causal rather than a multi-causal explanatory account for his high baseline completion rate. But, as mentioned, across most of the 1960s he failed to develop one.

19 Interestingly, Burger [2009, 3], who undertook a partial replication of the Obedience studies, came to a similar multi-causal conclusion to Milgram [1963] when he (Burger) presented “four features of the situation that likely contributed to the high rates of obedience [...]”: “Obedience to Authority”, “Gradual Increase in Demands”, “Limited Sources of Information in a Novel Situation” and “Responsibility Not Assigned or Diffused...” Other scholars have come to similar multi-causal conclusions, see [Bauman 1989, 151–164], [Hollander & Turowetz 2017, 671–672], [Miller 2014, 560–564]. The ensuing array of terms (“factors” and “mechanisms”)—many possible explanatory accounts—reflects a statement Milgram made in his research notes: “[It is not that] an explanation is lacking; but rather [...] there are so [ma]ny to cho[o]se from” [Stanley Milgram Papers, Box 46, Folder 175, Titled: “Notes: Proximity”]. Here Milgram is alluding to a weakness with the multi-causal explanatory accounts: they inject greater ambiguity (or complexity) over the question of why exactly subjects completed the baseline. That is, their many possible explanations collectively blend back into a rather unsatisfying hazy mystery that only returns one back to the original question: why did they do it?

20 Milgram’s above inventive journey may, in fact, have generated this failure, across the 1960s, to produce a singular mono-causal, over his above vague multi-causal, theoretical account. That is, as shown, when Milgram constructed his basic procedure, he added so many different manipulative techniques of social engineering—all those Strain Resolving Mechanisms and Binding Factors—to the baseline procedure that his limited score of experimental variations proved incapable of isolating all their independent influences. That is, there are many different potential explanations because there were many individual Strain Resolving Mechanisms and Binding Factors inherent within the procedure.

21 Although Milgram, Burger, and many others have presented multi-causal accounts to help explain the Obedience Studies, more ambitious mono-causal theories have been proposed. These mono-causal theories, if accurate, have a major advantage over the multi-causal accounts: they provide greater precision and parsimony over the question of why subjects completed the New Baseline . This potential for greater precision and parsimony essentially renders mono-causal theories the holy grail of explanatory accounts, and thus, compared to the unsatisfying multi-causal theories, they tend to attract enormous scholarly attention. Indeed, this is, at least in part, why the three most prominent theoretical accounts of the Obedience Studies are all mono-causal. The potential of these three most prominent theoretical accounts to shed light on why exactly subjects completed the New Baseline renders it important that their validity be assessed. Doris, Niemi et al. [2024] term the first of these three mono-causal accounts the Incredulity Hypothesis .

3 The Incredulity Hypothesis

22 Milgram [1972, 139] argued that successfully deceiving most subjects into believing that the shock generator was seriously harming the learner was absolutely “critical” for the methodological strength (internal validity) of the experiments. Deception was important because if most subjects did not believe the learner was being harmed, then there would not have been any “conflict” of conscience [Milgram 1963, 378] over whether or not to continue helping the experimenter collect his data. It is therefore widely accepted that “the entire [methodological] foundation of the obedience research rests on the believability of the victim’s increasingly mounting suffering” [Miller 1986, 143]. Successfully deceiving most subjects was also of great importance to Milgram because it made the next step in his research possible: generalizing his findings to the outside world (ecological validity); most controversially, to the Holocaust [Milgram 1963, 371].

23 But if critics could show the basic procedure failed to successfully deceive most subjects, then Milgram could no longer, as he frequently did, generalize from his findings to the outside world. Put differently, one cannot generalize to, say, the Holocaust or elsewhere, from a methodologically weak foundation. And over the past half century or so numerous scholars have challenged Milgram’s assertion that he successfully deceived his subjects into believing they were inflicting dangerous shocks.

24 The first scholars to challenge Milgram on these grounds was Orne & Holland [1968] They argued that the baseline result can be explained by the obedient subjects’ accurate suspicion that the experiment was a ruse. Subjects believed it to be a ruse because they could not imagine that any researcher or prestigious institution would allow an innocent person to be exposed to danger. Therefore, subjects concluded that, despite evidence to the contrary, no serious harm would come from their completing the experiment: doing so would be “all right” [Orne & Holland 1968, 287].

25 Yet if subjects did not believe the learner was being shocked, why did most complete the experiment? Orne & Holland argued that most subjects completed the baseline experiment because they were influenced by demand characteristics . Demand characteristics are when subjects attempt to detect the meaning and purpose of an experiment and, on doing this, they then engage in behaviors they think are likely to please the researcher by confirming their probable hypothesis. Thus, for Orne & Holland, most of Milgram’s subjects accurately detected the experiment’s covert aim: would they follow malevolent orders to harm an innocent person. Subjects, having established that the learner was not really being harmed, then proceeded to please Milgram by inflicting every shock asked of them whilst feigning explicit indicators of nervousness. Therefore, those subjects who completed the experiment did so to help Milgram achieve his goal of obtaining a high completion rate whilst also knowing that no harm would come from their doing so.

26 So how valid is Orne & Holland’s argument? In conflict with their accusation that demand characteristics generated the baseline result, Milgram’s inventive journey illustrates that when the Nazi-sounding “pledge to obey” justification for harm infliction was substituted in favour of an experiment that was (apparently) exploring the effects of punishment on learning, he clearly tried to hide from subjects his study’s actual objective of determining if ordinary people would follow a malevolent authority’s instructions. This inventive journey also shows that Milgram actively went to great lengths to protect his study from demand characteristics (recalling his skepticism with the first student-run pilot and, thereafter, the numerous ways he attempted to strengthen the basic procedure). What convinced Milgram he had successfully deceived most subjects was, of course, their signs of “extreme tension”: sweating, trembling, stuttering, groaning, biting their lips and digging their nails into their flesh [Milgram 1963, 375]. But one particularly convincing line of evidence detracting from Orne & Holland’s thesis was that if the obedient subjects knew the shocks were fake, why after completing the New Baseline experiment did 73 percent of them decline an opportunity to experience a 450-volt shock [Milgram 1974, 57]? Milgram showed no patience for Orne & Holland’s thesis, arguing their,

[...] suggestion that the subjects only feigned sweating, trembling, and stuttering to please the experimenter is pathetically detached from reality, equivalent to the statement that hemophiliacs bleed to keep their physicians busy. [Milgram 1972, 140]

27 Eckman [1977, 94] agreed: “when one reads the actual transcripts of Milgram’s subjects’ verbal behavior, it is hard to conclude it was all a put-on. There was just too much conflict and stress.” Thus, “[t]o invoke the charge ‘demand characteristics’ against Milgram’s work is foolish” [1977, 95].

28 Although there are strong grounds for dismissing the validity of Orne & Holland’s central argument that demand characteristics generated Milgram’s results, their issue of trust—that obedient subjects did not believe the shocks were dangerous and therefore that Milgram failed to deceive them—has had a powerful influence on the subsequent literature, see [Gibson 2019, 38–39], [Perry 2013, 173, 258], [Perry, Brannigan et al. 2020]. Much of this influence, however, traces not to Orne & Holland, but to Don Mixon. Across a series of publications in the 1970s and 1980s, Mixon provided a new and more persuasive explanation for the subjects’ stressed reactions, which, as a result, elevated Orne & Holland’s issue of trust to new and more influential heights.

29 Mixon set out by arguing that subject displays of stress and anxiety were, contra Orne & Holland, genuine. However, this stress was not due to, as Milgram argued, a conflict of conscience, but because of their exposure to an ambiguous situation where the information coming from the experimenter and learner was contradictory: the shocks were apparently harmful, but not dangerous; the experimenter was calm, but the learner was screaming in agony. Milgram’s ambiguous procedure therefore generated confusion among his subjects. As Mixon observed, “No wonder many subjects showed such stress. What to believe? The right thing to do depends on which actor is believed” [Mixon 1989, 33]. Then, according to Mixon, because “increasingly large chunks of the social and physical world that we live in can be understood only by experts” [1989, 35], most of the obedient subjects chose to resolve the stressfully ambiguous situation confronting them by trusting the authority figure’s word: he said that although the shocks “may be painful” they would “cause no permanent tissue damage [...]” [Milgram 1974, 56]. Thus, for the obedient subjects, the study must have been benign.

30 In support of Mixon’s claims, the above invention of the Obedience Studies illustrates that during the pilot studies Milgram indeed purposefully injected ambiguity into his emerging basic procedure. For example, Milgram changed the designated title of the last button on his proposed shock machine from “LETHAL” (Figure 1) to “XXX” (official shock generator [Milgram 1974, 28]), presumably because he sensed the definitiveness surrounding a word like “LETHAL” might detract from his preconceived goal to “maximize obedience [...]”. And there are other examples were Milgram removed clarity and replaced it with ambiguity [see Russell 2018, 123]. Thus, there is support for the claim that Milgram purposefully created an ambiguous and thus confusing situation where, in pursuit of stress reduction, subjects might have decided to side with the only expert present. Again, many contemporary scholars, like Burger, see merit in Mixon’s argument:

When you’re in that situation, wondering, should I continue or should I not, there are reasons to do both. What you do have is an expert in the room who knows all about this study, and presumably has been through this many times before with many participants, and he’s telling you, [t]here’s nothing wrong. The reasonable, rational thing to do is to listen to the guy who’s the expert when you’re not sure what to do. [Burger cited in Perry 2013, 359]

31 Although Mixon’s argument is both different to and more convincing than Orne & Holland’s, the conclusions of both groups of scholars on the issue over whether or not Milgram successfully deceived his obedient subjects are the same. To clarify, because, for Mixon, obedient subjects believed the “expert” experimenter when he said the shocks were not dangerous, what is implied is that Milgram’s all-important attempts to deceive them into believing they were seriously harming the learner must have failed. Therefore, both Orne & Holland and Mixon agree: obedient subjects trusted that the experimenter would not allow the learner to experience dangerous shocks and this belief—that Milgram’s attempts at deception failed—mono-causally explains why they decided to complete the baseline experiment.

32 More recently, Perry, Brannigan et al. [2020] discovered compelling evidence in Milgram’s personal archive that lends weight to Orne & Holland and Mixon’s point of agreement. They first note:

In his first journal article about his obedience research, Milgram [...] stressed the dramaturgical credibility of the experiment. He emphasized that “[w]ith few exceptions subjects were convinced of the reality of the experimental situation” [...]. The implication was that the subjects fully believed that what was happening was real, and despite indications that the learner was in increasing pain, 26 out of 40 proceeded to administer the maximum shock. [Perry, Brannigan et al. 2020, 92]

33 However, in conflict with this, Perry, Brannigan et al. add that what Milgram failed to mention was that in 1962 he instructed his research assistant Taketo Murata to analyse the subjects’ post-experimental interview survey data and that he do so with a specific focus on the issue of subject deception. The report Murata wrote compared:

[...] the degrees of obedience between those subjects who said they were doubtful that the shocks were painful and those who were certain they were. [...] Murata [1962:1] wrote, “The following is a condition-by-condition analysis to determine whether shock level reached was affected by the extent to which the subject believed that the learner was actually receiving shock.” [...] The report’s main results were a comparison of the mean shock levels administered by subjects who had “fully believed” [...] in the reality of shocks versus those classified as having “not fully believed” [...]. Murata found that in 18 of 23 experiments, those subjects who fully believed the learner was getting painful shocks gave lower levels of shock than subjects who doubted the shocks were real. [Perry, Brannigan et al. 2020, 94]

34 The reverse also applies: “Those who were less successfully convinced by the cover story were more obedient” [2020, 99]. Put differently: “obedience increases with skepticism of pain” [2020, 98], a conclusion that, across most of Milgram’s experiments, bolsters Orne & Holland’s and Mixon’s point of agreement that Milgram failed to sufficiently deceive his subjects into believing the shocks were dangerous.

35 Perry, Brannigan et al. further reinforced their point by referencing Hollander & Turowetz’s analysis of obedient subjects’ post-experimental justifications for completing. That is, across the five conditions for which, at Milgram’s personal archive, audio-recordings were available, Hollander & Turowetz [2017, 661–662] found the most common reason 46 obedient subjects provided for inflicting every shock was because they did not think the learner was being harmed (33 subjects, or 72  percent).

36 So what emerges across Orne & Holland, Mixon, and Perry, Brannigan  et al .’s research is a common denominator: Milgram’s baseline procedure failed to sufficiently deceive the “obedient” subjects into believing the shocks were harmful and, as a result, these subjects developed a belief that no harm would come from their compliant actions. And this mono-causal account explains why most New Baseline subjects completed the experiment. Although all three accounts are technically methodological critiques of the Obedience Studies, all provide dismissive explanatory accounts for Milgram’s apparently counter-intuitive results.

• 1 As mentioned, Mixon argued that Milgram’s purposeful insertion of ambiguity left subjects with lit (...)

37 However, the validity of Orne & Holland, Mixon, and Perry et al. ’s arguments have been challenged [see Russell 2018, 120–126], [Russell & Gregory 2021, Russell 2024]. The crux of this counter-critique is that largely because of the partition separating the subjects from the learner, subjects in the New Baseline could not know for certain if the shocks really were harmful. The wall introduced a strong dose of ambiguity into the New Baseline procedure. Unable to see the learner, “many” subjects were left in, as Perry notes, “a state of uncertainty and stress” [2013, 162]. This uncertainty meant any decision by a skeptical subject to inflict every shock necessitated they take a major risk: their suspicion that the experiment was possibly, probably, or even most certainly harmless could be wrong and, if so, the learner would be seriously injured. For subjects who took this risk, an important question emerges: was it morally correct for them to prioritise their mere suspicion the experiment was fake and inflict every shock because they thought everything would be “all right”? 1

38 There is, it transpires, a morally correct and inherently safe resolution to the dilemma over whether to stop or continue inflicting shocks: if a subject was, as all must have been, unsure if the shocks were real, the safest (moral) choice was to err on the side of caution and prematurely end the experiment [Coutts 1977, 520 cited in Russell 2018, 125]. Doing so eliminated all risk of a subject’s hunch being wrong and instead guaranteed the safety of a fellow human being [Russell & Gregory 2021]. Although this ethically cautious type of problem-solving was uncommon, a minority of subjects exhibited it across numerous experimental conditions. For example, one suspicious and uncooperative subject later said, “When I decided that I wouldn’t go along with any more shocks, my feeling was ‘plant or not [...] I was not going to take a chance that our learner would get hurt’ ” [Russell 2018, 124]. Another subject in a different condition noted he “wasn’t sure” if the learner “was getting the shocks,” but when “he started to complain vigorously [...] I refused to go on” [Russell 2018, 142]. In fact, one subject explicitly stated he was certain the subject was not being shocked, but he refused to trust the experimenter:

Teacher “I don’t believe this! [...]”

E xperimenter “You don’t believe what? [...]”

Teacher “I don’t believe you were giving him the shock.”

E xperimenter “Then why, why won’t you continue?”

Teacher “Well I, I just don’t want to take a chance [...]” [Russell 2018, 124]

39 This subject did not believe the shocks were real but, largely because of the wall separating him from the learner, he could not be certain. The subject’s uncertainty in the ambiguous situation confronting him dictated that he could not afford to take that chance and—contra Mixon—trust the expert in charge because there was still a possibility his hunch might be wrong. And this subject was obviously aware of the consequences that such a mistake would have on the learner.

40 But in the New Baseline condition most subjects resolved the experiment’s inherent moral dilemma differently: despite the risk that the shocks might have been real, 65 percent of them assumed that no harm would come from their completing the experiment and that doing so would probably be “all right.” Of course, because all New Baseline subjects remained uncertain over the reality of the shocks, deciding to risk another person’s well-being over a mere hunch was unethical (especially when compared to the option of prematurely stopping what might have been a dangerous experiment).

41 Thus, the mystery at hand returns: why did 65 percent of New Baseline subjects choose to act so unethically? Perhaps the most discussed explanatory account of the Obedience Studies was that eventually offered by Milgram [1974]; probably because of the widely shared assumption that surely the inventor of this experiment knew why most completed the baseline.

4 The Agentic State

42 In 1974 Milgram published his book Obedience to Authority: An Experimental View. To give this book an air of completeness, Milgram greatly desired it include a mono- over a multi-causal theoretical account. Thus, from the late 1960s he started developing his best attempt at producing such a theory: his so-called Agentic State . So, what did this mono-causal theory entail?

43 In his book Milgram, drawing on evolutionary theory, argued that, like many animals, humans function in hierarchical structures. Due to the survival value of organization over disorganisation, he argued hierarchical structures have, through breeding, led humans to gain advantages over their competitors in the natural environment. Milgram is careful to point out that obedience to those higher in the chain of command is not innate. “Rather, we are born with a potential for obedience, which then interacts with the influence of society to produce the obedient man” [1974, 125]. He then argued that mankind’s potentially destructive impulses are frequently kept in check by the conscience of an autonomously acting individual. However, when an individual is introduced into a hierarchical chain of command, a homeostatic-driven internal change can take hold in which conscience-driven control over their actions can be suppressed and supplanted by a higher authority’s demands. Milgram termed this process the Agentic Shift , which is where “[...] a man [ sic ] feels responsible to the authority directing him but feels no responsibility for the content of the actions that the authority prescribes” [1974, 145–146].

44 After this shift, the previously autonomous individual enters the Agentic State , which is where “the individual no longer views himself as responsible for his [ sic ] own actions but defines himself [ sic ] as an instrument for carrying out the wishes of others” [1974, 134]. In this state subjects became so-called exemplary functionaries: solely concerned with—to the full satisfaction of the experimenter—how effectively they carried out their instructions.

45 Milgram believed two main factors put subjects into the Agentic State : his previously mentioned Strain Resolving Mechanisms and Binding Factors. In explaining both compliance and defiance, Milgram [1974, 154] succinctly captured his theory as follows:

O ;  B  > ( s − r )

D ;  B  < ( s − r )

in which O  represents obedience; D , disobedience; B , binding factors;  s , strain; and r , the strain-resolving mechanisms. Obedience is the outcome when the binding factors are greater than the net strain (strain as reduced by the resolving mechanisms), whereas disobedience results when net strain exceeds the strength of the binding forces.

46 Considering Milgram’s earlier journey of discovery, one can see why he found aspects of this post hoc theory appealing. That is, the Agentic State’s above concise mathematical-like formula is not only mono- over multi-causal [Gibson 2019, 61], but it is also a theory of social engineering that closely aligns with his invention of the Obedience Studies: the more manipulative forces he added, the more efficacious his basic procedure seemed to become.

47 That said, Milgram’s theory has attracted much criticism [see Russell 2018, 132]. One telling critique is that the Agentic State cannot explain the variance in completion rates across the different experimental conditions [Blass 2004, 216]. Neither can it account for individual differences within conditions [Gibson 2019, 56–57]. So although the Agentic State attempted to offer a comprehensive mono-causal or “totalizing” explanation that illuminated “behaviour across individuals, across experimental conditions, and across time” [2019, 61], it failed to deliver.

48 Even scholars noted for being positively disposed toward Milgram’s research admit there is “very little empirical support” for the Agentic State [Miller 2004, 210] and that the book’s “weakest” chapter is that which presents his theory [Blass 2004, 216]. Again, after Milgram had been trapped for approaching a decade in a mono-causal theoretical drought, the Agentic State was likely his last-ditch attempt at offering such an explanation. It transpires that the creator of the experiments did not know why subjects completed his basic procedure.

49 Although in 2004, Obedience-study specialist Arthur Miller argued, after the demise of the Agentic State , that there is “no conclusive theory to account for destructive obedience—or defiance, either” [2004, 233], a decade later he alluded to a “[...] most promising recent development [...]” [2004, 568]. Miller was referring to Engaged Followership theory.

5 Engaged Followership

50 The leading authors of Engaged Followership theory—Haslam & Reicher—argue that subjects in the Obedience Studies:

[...] are able to inflict harm on others not because they are unaware that they are doing wrong, but rather because—as engaged followers—they know full well what they are doing and believe it to be right [...]. [Haslam, Reicher et al. 2015 a , 79]

51 This engagement, they add, involved subjects identifying with a “noble” cause [2015 a , 79], specifically the subjects’ personal affiliation with the experimenter’s “scientific mission”: the research’s purported quest to explore the effects of punishment on learning [2015 a , 77]. Thus, the experimenter’s explicit and implied appeals to his scientific mission—and the greater good so frequently associated with it—persuaded then motivated most baseline subjects to endure the victim’s pained reactions until every shock had been inflicted. As the authors themselves conclude:

[...] it appears that the behavior of Milgram’s participants was underpinned by active commitment to his [scientific] endeavor and an associated desire to prove themselves as worthy subjects. Indeed, it was this commitment and desire that appear to have steeled them to persist in the face of the significant stressors that the task exposed them to. [Haslam & Reicher 2017, 71]

52 As the pseudonymous subject Morris Braverman worded it: “[...] in the interest of science, one goes through with it” [Milgram 1974, 54].

53 Haslam & Reicher have published a variety of empirical studies that they argue provides evidence in favour of their theory [Reicher, Haslam et al. 2012, Haslam, Reicher et al. 2014, 2015 b,a ].

54 According to Gibson, the “crucial variable” to emerge from this research is that of “social identification” and, as a result, Haslam & Reicher found:

[...] conditions that promote identification with the experimenter lead to higher levels of completion; conditions that promote higher levels of identification with the learner lead to higher levels of defiance. [Gibson 2019, 59]

55 If true, Engaged Followership theory purports to explain not only why subjects completed the New Baseline , but also why they resisted:

[...] being torn between two insistent voices—the one representing science, the other representing the community of ordinary people—the question of which one the participant heeds ultimately depends on which category he identifies with most. [Reicher, Haslam et al. 2012, 319]

56 This intriguing theory, however, has attracted some criticism. For example, Gibson [2019, 60] notes, despite the above evidence in support of their account, Haslam & Reicher’s theory has never directly been tested on an Obedience Study-type procedure, adding:

[...] it is notable that the only study that involves direct measures of naïve participants’ levels of identification in a Milgram-esque paradigm [Haslam et al. , 2014] found that there were no effects of identification on participants’ likelihood of completing the experiment, likelihood of completing post-experiment measures, or on the length of time that participants continued with the experiment. This raises the question of how well the somewhat-limited evidence available to date actually supports this account of Milgram’s findings.

• 2 An issue with Hollander & Turowetz’s thesis is if most obedient subjects, as they assert, complete (...)

57 Other critics like Hollander & Turowetz note that during Milgram’s post-experimental interviews, subjects frequently stated they completed the experiment for a range of reasons that had nothing to do with advancing the experiment’s scientific aims. More specifically, and as touched on earlier, Hollander & Turowetz note that out of their sample of 46 obedient subjects, most provided the following (sometimes multiple) accounts for completing: because they did not think the learner was being harmed: 33 (72 percent), because they were following instructions: 27 (59 percent), or because they felt contractually obligated into completing: 5 (11 percent). Subjects only mentioned they inflicted every shock because of the importance of the experiment 11 times (24 percent) [Hollander & Turowetz 2017, 661–662]. Hollander & Turowetz thus suspect there are likely “multiple empirically grounded”, thus a multi -causal array of “explanations” of the Obedience Studies [Hollander & Turowetz 2018, 302]. 2

58 Because Hollander & Turowetz found subjects offered a variety of different justifications for completing, they accused Haslam & Reicher of advancing a mono-causal explanatory account that is simply not true [2017, 671–672], [see also Gibson 2019, 61]. Haslam & Reicher [2018, 293] responded:

3 Although they may not have intended for their theory to be mono-causal, it does closely resemble s (...) In proposing the “engaged followership” account our intention was never to try to replace one monolithic orthodoxy [Milgram’s Agentic State ] with another and to try to have the final word. We entirely endorse H[ollander] & T[urowetz]’s argument that it would be both futile and wrong to posit a single process to account for what Milgram found [...]. 3

59 Thus, here, on top of now agreeing that there are probably a variety of possible reasons subjects inflict every shock, Haslam & Reicher willingly concede that Engaged Followership theory can only account for why some , and not all, subjects completed the Obedience Studies.

60 That said, Haslam & Reicher have countered aspects of Hollander & Turowetz’s critique. For example, Haslam & Reicher used Engaged Followership theory to account for Hollander & Turowetz’s finding that most obedient subjects in their sample later justified completing because they did not think the learner was being harmed:

Participants are induced to cooperate because they accept the Experimenter’s assurances about the study which in turn is because they accept the credibility of his statements as a scientist. [Haslam & Reicher 2018, 294]

They then, in different words, repeat this point:

More concretely, to say “I imposed painful shocks against the learner’s will because I think science is a good thing” seems contradictory unless one goes on to explain that because one believes in science one also trusts that a legitimate scientific experiment will not harm people. [Haslam & Reicher 2018, 295]

61 The problem, however, with Haslam & Reicher’s counter argument is, in an attempt to defend their embattled theory, they deploy one of Mixon’s foundational arguments from his methodological critique of the Obedience Studies: because “large chunks” of the modern world “can be understood only by experts”, ordinary people in confusing situations place their trust in experts who know best: the experimenter said the seemingly painful shocks were not dangerous, so obedient subjects knew shock-infliction was harmless. Therefore, according to Mixon, most subjects completed the New Baseline not because, as Milgram argued, they buckled under the coercive power of malevolent authority, but because they knew it was safe to do so. Of course, if Mixon’s methodological critique is correct (and Haslam & Reicher’s above quotes inadvertently implies it is), then what is the use in them trying to provide an explanation for the results of a study that has such a flimsy methodological foundation? Put differently, Haslam & Reicher’s pro-Mixon counter argument reinvigorates the ability of their theory to explain the results of a social science experiment that—according to Mixon—is methodologically flawed.

62 Engaged Followership theory is not immune to another perhaps even more damning criticism. Haslam & Reicher cannot evade the following fact: among all those subjects who later justified that they completed the experiment to help advance “science,” some of this group also simultaneously tried to undermine (sabotage?) the experiment’s purported scientific goals [Russell 2018, 230–231]. For example, consider the pseudonymous and fully “obedient” Eleonor Rosenblum who, during her post-experimental debrief, justified completing because: “It is an experiment [...]. So I had to do it” [Milgram 1974, 83]; later also declaring “I’m one for science [...]” [1974, 83]. So, although Haslam & Reicher now concede that not every subject offered this justification for completing, here is a clear case where Engaged Followership theory seems to rather precisely account for why a subject inflicted every shock. As a supporter of science, Rosenblum believed that inflicting the shocks was, as Haslam & Reicher would put it, “right [...]” [Haslam, Reicher et al. 2015 a , 79].

63 However, when Rosenblum was later reintroduced to the unharmed learner she confided to him, in a statement whose accuracy was corroborated by the measured recording of her performance, “ ‘Did you hear me stressing the [correct] word[?] I was hoping that you would hear me’ ” [Milgram 1974, 82]. That is, Rosenblum covertly attempted to provide the learner with the correct response to her questions by vocally emphasizing the correct answer. Here Rosenblum engaged in what Milgram termed “subterfuge” [1974, 159], because she hoped her provision of the right word would enable the learner to correctly answer her questions thereby enabling her to avoid shocking him. Of course, an obedient subject who genuinely identified with the experiment’s scientific mission and believed shock infliction to be “right”, should not have covertly tried to undermine the very purpose that they (apparently) so identified. So, although Rosenblum’s behaviour appears to be a textbook case in support of Haslam & Reicher’s theory—she said she inflicted the shocks it for science—it appears their wilting theoretical account cannot even explain her behaviour. What further detracts from the potency of Haslam & Reicher’s theory is that Rosenblum was no outlier. As Milgram observed:

Some subjects could be observed signaling the correct answer to the victim by stressing it vocally as they read the multiple-choice words aloud. That is, they attempted to prompt the learner and thus prevent his receiving shocks. [Milgram 1974, 159]

And Milgram’s detection of acts of subterfuge have independently been confirmed by others:

Hannah wasn’t the only subject who defied Williams [the experimenter] by accentuating the right answer—I’d heard it on a number of recordings. [...] another subject said: “I did everything in my power to emphasise the correct answers [...] and I hoped the supervisor watching it didn’t catch on [...].” [Perry 2013, 124]

• 4 Haslam & Reicher could respond to this point by noting that, as they say, “[...] even when partici (...)

64 There was even some diversity in these subjects’ techniques of subterfuge: some gave the appearance they were inflicting intensifying shocks, but instead covertly attempted to avoid fully depressing the shock switches [Perry 2013, 197], [Gibson 2019, 183]. Subterfuge occurred so frequently that Milgram had to instruct the experimenter to ensure subjects performed their set tasks correctly, with him telling one errant subject: “Please read the words in an even tone” [Perry 2013, 196]. This subject then attempted to lightly brush the shock switches, to which the experimenter responded: “That didn’t make contact. Press the switch down completely” [2013, 197]. Nonetheless, the key point being is that Engaged Followership theory struggles to account for subterfuge-type behaviours. 4

65 Although Haslam & Reicher’s theory exhibits various weaknesses, it still offers an important piece of the theoretical puzzle; remembering that when Milgram was inventing the experiments he would not have gotten far without the legitimating justification of harm-infliction for science . As Burger [2009, 3] argued:

Milgram’s experimenter was granted the legitimacy of authority by virtue of his association with the experiment, the university, and perhaps even science.

66 However, we suspect the aura of science is less explanatory and more of a singular and no doubt important Milgram-like Strain Resolving Mechanism: any means that, as mentioned, helped subjects offset the tension typically associated with harm infliction. And the term for this specific Strain Resolving Mechanism is Moral Inversion . Moral Inversion is where harm infliction is persuasively transformed from “something evil” (shocking an innocent person) into something “good” (advancing human learning) [Adams & Balfour 1998, xx, cited in Russell 2018, 59].

67 For us, the above criticisms against Engaged Followship theory sees its relegation from a mono-causal, to a partial, to—at best—a minor and incomplete explanation for Milgram’s results (again, better subsumed under the overriding banner of Moral Inversion ). We agree with Hollander & Turowetz that “[...] the puzzle of Milgramesque behaviour is more complex than engaged followership at present allows” [2017, 672].

68 In conclusion of this section, a critical assessment of what are arguably the three most discussed explanatory accounts as to why ordinary people completed Milgram’s New Baseline experiment has been provided. We dismiss all three accounts, largely because, in a variety of ways, their assertions are inconsistent with Milgram’s findings. To avoid leaving the reader at an explanatory dead-end, we also believe less prominent theories should be granted greater critical attention. The next section presents one such theory.

6 Fear of a confrontation

69 As already touched on, one potentially promising theory is inherent within Milgram’s earlier invention of his procedure: as Milgram, in the role of a social engineer, added more strain resolving mechanisms and binding factors to his basic procedure, the cumulatively more coercive the experiment became. This explanation may help explain how Milgram ensured most subjects completed the baseline, but it does not account for why exactly they chose to do so. Thus, the motivational impetus remains a mystery. In his book Milgram proposed, beyond his Agentic State , another reason why some subjects may have decided to complete. He observed:

If sympathetic concern for the victim were the exclusive force, all subjects would have calmly defied the experimenter. Instead, there were both obedient and defiant outcomes, frequently accompanied by extreme tension. [...] Therefore there must be a competing [...] inhibition that precludes activation of the disobedient response. [Milgram 1974, 42–43]

70 If the competing inhibition could be identified, the base motive/s would be revealed. One possible competing inhibition Milgram mentioned that may have precluded disobedience was:

The learner cannot break off and at the same time protect the authority’s definitions of his own competence. Thus, the subject fears that if he breaks off, he will appear arrogant, untoward, and rude. Such emotions [...] suffuse the mind and feelings of the subject, who is miserable at the prospect of having to repudiate the authority to his face. [Milgram 1974, 150]

• 5 This explanation is multi-causal because if most subjects completed for this reason, a minority of (...)
• 6 We suspect the previously discussed Eleonor Rosenblum was a subject whose behaviour is explicable (...)

71 Others have taken this potential explanation for completing further with what, it will be noted, is a multi-causal account: “a socially awkward confrontation with the experimenter was actually the most common—but not the only—[...] motivational force behind the New Baseline experiment’s high completion rate” [Russell 2018, 201], [see also Russell & Gregory 2011]. 5 For the reader, this probably sounds like too trivial a reason for completing, but Milgram’s results lend much weight in favour of it: in nearly all experiments where subjects, in order to prematurely end the experiment, did not have to initiate and then engage in a direct confrontation with the experimenter, none, or few of them inflicted every shock (average completion rate: 11.25 percent ). This account suggests that in the absence of having to engage in a direct confrontation to stop the experiment, nothing or little competed with most subjects’ sympathy for the learner’s predicament, leaving them free to act on their personal preference to stop inflicting further shocks. However, in nearly all the experiments that necessitated subjects, in order to end them prematurely, initiate then engage in a direct confrontation with the experimenter, relatively large percentages of subjects inflicted every shock (average completion rate: 54.61 percent (see Table 1 below)). 6 A statistical test based on Table 1 justifies the conclusion that the hypothesis that a fear of confrontation is causally relevant for “obedient” behaviour should be accepted (Chi Square obedient vs. disobedient in the two kinds of conditions, 1  d.f. = 123 ; p < 0.01.

Table 1: Experimental conditions where siding with the learner did or did not require the subject to initiate then engage in a confrontation with the experimenter. [Russell & Gregory 2011], [Russell 2018, 79–80; 92–96; 202]

72 The most obvious exception to this general rule was the Relationship condition, where subjects were instructed to bring another person along to the laboratory who was at least an acquaintance they had known for two or more years (3 of the 20 subjects, however, brought a relative). On arrival, one of the pair became the teacher and the other the learner (then after the learner was strapped into the electric chair and the teacher and experimenter left the “shock” room, Milgram appeared before the learner and explained the experiment’s actual purpose: to test if their friend would shock them. Milgram then quickly trained these learners how to react to the intensifying “shocks.” But in conflict with Mixon’s assertion that people in ambiguous situations tend to side with experts, 85 percent of subjects in this condition refused to do so [Russell 2014]. What the Relationship condition illustrated is that a teacher’s fear of having to confront the experimenter in order to stop the experiment diminished greatly when the victim was someone they knew. In fact, this fear appeared to completely evaporate when the acquaintance happened to be a family member: 0 out of 3 (0 percent) completed. This result—despite being Milgram’s greatest ethical breech—hints at an important insight that, we would argue, alludes to the inner-workings of the basic procedure: excluding the Relationship condition, all of Milgram’s baselines and experimental variations involved a stranger as the victim. And with stranger victims, it was obviously, relatively speaking, easier for Milgram to coerce (socially engineer) many people with his barrage of binding factors and strain resolving mechanisms into completing (especially when compared to him pushing subjects into hurting a person who was an acquaintance and, especially, a family member). In sum, factors like victim relationship in conjunction with Milgram’s mention of the fear of a confrontation—and, not discussed here, the incomparably powerful strain-resolving power of the shock generator [Russell 2018, 239–245]—are all potential insights that we believe require greater critical attention. See Russell [2018] for a fuller account of this theory.

7 Conclusion

73 This article has presented and critically assessed what are arguably the three most discussed and thus prominent explanatory accounts as to why ordinary people completed Milgram’s New Baseline experiment. The theories reviewed included the Incredulity Hypothesis , the Agentic State , and Engaged Followership . We dismissed all three theories because, in a variety of ways, their assertions are inconsistent with what Milgram found. Consequently, we believe less prominent theories exploring subject motivation—like the subjects’ fear of having to engage in a confrontation with the experimenter—requires closer critical attention.

74 But despite our assertions on what theories deserve less or more attention, what should become apparent to the reader is that in relation to what is perhaps both the most revered and reviled experiment in social psychology—a study that for well over half a century has, across both scholarly circles and in popular culture, sustained enormous attention—the gains on the theoretical frontlines have been miniscule. But the positive side of this theoretical deficit is that for budding social scientists, this likely means that the mystery of why subjects completed the baseline experiment remains.

Bibliography

Bauman, Zygmunt [1989], Modernity and the Holocaust , Ithaca: Cornell University Press.

Blass, Thomas [2004], The Man Who Shocked the World: The life and legacy of Stanley Milgram , New York: Basic Books.

Burger, Jerry M. [2009], Replicating Milgram: Would people still obey today, American Psychologist , 64(1), 1–11, doi: 10.1037/a0010932 .

Doris, M., John, Niemi, Laura, et al. [2024], The incredulity hypothesis and the enduring legacy of the obedience experiments, Philosophia Scientiæ , 28(2), 53–89.

Eckman, Bruce K. [1977], Stanley Milgram’s obedience studies, Et Cetera , 34, 88–99.

Freedman, Jonathan L. & Fraser, Scott C. [1966], Compliance without pressure: The foot-in-the-door technique, Journal of Personality and Social Psychology , 4(2), 195–202, doi: 10.1037/h0023552 .

Gibson, Stephen [2019], Arguing, Obeying and Defying: A Rhetorical Perspective on Stanley Milgram’s Obedience Experiments , Cambridge University Press, 1st edn., doi: 10.1017/9781108367943 .

Haslam, S. Alexander & Reicher, Stephen D. [2017], 50 years of Obedience to authority : From blind conformity to engaged followership, Annual Review of Law and Social Science , 13(1), 59–78, doi: 10.1146/annurev-lawsocsci-110316-113710 .

Haslam, S. Alexander & Reicher, Stephen D. [2018], A truth that does not always speak its name: How Hollander and Turowetz’s findings confirm and extend the engaged followership analysis of harm‐doing in the Milgram paradigm, British Journal of Social Psychology , 57(2), 292–300, doi: 10.1111/bjso.12247 .

Haslam, S. Alexander, Reicher, Stephen D., et al. [2014], Nothing by mere authority: Evidence that in an experimental analogue of the Milgram paradigm participants are motivated not by orders but by appeals to science, Journal of Social Issues , 70(3), 473–488, doi: 10.1111/josi.12072 .

Haslam, S. Alexander, Reicher, Stephen D., et al. [2015 a ], “Happy to have been of service”: The Yale archive as a window into the engaged followership of participants in Milgram’s ‘obedience’ experiments, British Journal of Social Psychology , 54(1), 55–83, doi: 10.1111/bjso.12074 .

Haslam, S. Alexander, Reicher, Stephen D., et al. [2015 b ], Shock treatment: Using immersive digital realism to restage and re-examine Milgram’s “obedience to authority” research, PLoS ONE , 10(3), e109 015, doi: 10.1371/journal.pone.0109015 .

Hollander, Matthew M. & Turowetz, Jason [2017], Normalizing trust: Participants’ immediately post-hoc explanations of behaviour in Milgram’s “obedience” experiments, British Journal of Social Psychology , 56(4), 655–674, doi: 10.1111/bjso.12206 .

Hollander, Matthew M. & Turowetz, Jason [2018], Multiple compliant processes: A reply to Haslam and Reicher on the engaged followership explanation of ‘obedience’ in Milgram’s experiments, British Journal of Social Psychology , 57(2), 301–309, doi: 10.1111/bjso.12252 .

Milgram, Stanley [1963], Behavioral study of obedience, The Journal of Abnormal and Social Psychology , 67(4), 371–378, doi: 10.1037/h0040525 .

Milgram, Stanley [1965], Some conditions of obedience and disobedience to authority, Human Relations , 18(1), 57–76, doi: 10.1177/001872676501800105 .

Milgram, Stanley [1972], Interpreting obedience: Error and evidence. A reply to Orne and Holland, in: The Social Psychology of Psychological Research , edited by A. G. Miller, New York: Free Press, 138–154.

Milgram, Stanley [1974], Obedience to Authority: An experimental view , New York: Harper and Row.

Miller, Arthur G. [1986], The Obedience Experiments: A case study of controversy in social science , New York: Praeger.

Miller, Arthur G. [2004], What can the Milgram obedience experiments tell us about the Holocaust? Generalizing from the social psychology laboratory, in: The Social Psychology of Good and Evil , edited by A. G. Miller, New York: Guilford Press, 193–237.

Miller, Arthur G. [2014], The explanatory value of Milgram’s obedience experiments: A contemporary appraisal, Journal of Social Issues , 70(3), 558–573, doi: 10.1111/josi.12078 .

Mixon, Don [1989], Obedience and Civilization: Authorized Crime and the Normality of Evil , London: Pluto Press.

Orne, Martin T. & Holland, Charles Howard [1968], Some conditions of obedience and disobedience to authority. On the ecological validity of laboratory deceptions, International Journal of Psychiatry , 6(4), 282–293.

Perry, Gina [2013], Beyond the Shock Machine: The Untold Story of the Milgram Obedience Experiments , Melbourne: Scribe.

Perry, Gina, Brannigan, Augustine, et al. [2020], Credibility and incredulity in Milgram’s obedience experiments: A reanalysis of an unpublished test, Social Psychology Quarterly , 83(1), 88–106, doi: 10.1177/0190272519861952 .

Reicher, Stephen D., Haslam, S. Alexander, et al. [2012], Working toward the experimenter: Reconceptualizing obedience within the Milgram paradigm as identification-based followership, Perspectives on Psychological Science , 7(4), 315–324, doi: 10.1177/1745691612448482 .

Russell, Daniel C. [2009], Practical Intelligence and the Virtues , Oxford: Oxford University Press.

Russell, Nestar [2011], Milgram’s obedience to authority experiments: Origins and early evolution, British Journal of Social Psychology , 50(1), 140–162, doi: 10.1348/014466610x492205 .

Russell, Nestar [2014], Stanley Milgram’s obedience to authority “Relationship” Condition: Some methodological and theoretical implications, Social Sciences , 3(2), 194–214, doi: 10.3390/socsci3020194 .

Russell, Nestar [2018], Understanding Willing Participants. Milgram’s Obedience Experiments and the Holocaust , vol. 1: Milgram’s Obedience Experiments and the Holocaust, Cham: Palgrave MacMilllan, doi: 10.1007/978-3-319-95816-3 .

Russell, Nestar [2024], Stanley Milgram’s obedience studies: A contemporary ethical and methodological assessment, Philosophia Scientiæ , 28(2), 107–129.

Russell, Nestar & Gregory, Robert [2011], Spinning an organizational “Web of obligation”? Moral choice in Stanley Milgram’s “Obedience” experiments, The American Review of Public Administration , 41(5), 495–518, doi: 10.1177/0275074010384129 .

Russell, Nestar & Gregory, Robert [2021], Are Milgram’s obedience studies internally valid? Critique and counter-critique, Open Journal of Social Sciences , 09(02), 65–93, doi: 10.4236/jss.2021.92005 .

Slater, Lauren [2004], Opening Skinner’s Box: Great Psychological Experiments of the Twentieth Century , New York: W. W. Norton.

1 As mentioned, Mixon argued that Milgram’s purposeful insertion of ambiguity left subjects with little choice but to side with the only expert present: the experimenter. Mixon therefore suspected that if Milgram eliminated—instead of purposefully inserted—ambiguity into the basic procedure, subjects would have known if the shocks really were harmful and acted accordingly (if the shocks were real, they would not have completed the experiment). We disagree. Rather, the ambiguity Milgram injected into the basic procedure introduced the possibility that the obedient subjects’ hunch that the learner was not seriously being harmed could have been wrong, which, as mentioned, was a possibility that came with potentially devastating consequences for the learner. If all subjects could have been certain that the learner was not being seriously harmed, there would have been no chance of them being wrong. And if there was no possibility of being wrong, then subjects would not have been faced with resolving a moral dilemma, which is what the experiment ultimately tested. In sum, no ambiguity, no moral dilemma.

2 An issue with Hollander & Turowetz’s thesis is if most obedient subjects, as they assert, completed Milgram’s basic procedure because they did not think the learner was being harmed , they are effectively siding with Orne & Holland, Mixon, and Perry et al .’s point of agreement: Milgram’s attempts to deceive his subjects failed. Therefore, it can be argued that Hollander & Turowetz are, by implication, supporters of the Incredulity Hypothesis , and thus the survival of their thesis—recently converted into a book—will be tied to its fate.

3 Although they may not have intended for their theory to be mono-causal, it does closely resemble such an account: “Our reconceptualization also has the advantage of invoking a single set of processes to explain not only when people follow authority but also when they resist it” [Reicher, Haslam et al. 2012, 322].

4 Haslam & Reicher could respond to this point by noting that, as they say, “[...] even when participants follow the experimenter, there is still a pull toward the learner—a concern with his fate. [...] To some extent at least, participants attend to both voices” [Reicher, Haslam et al. 2012, 323]. Our issue with this response is with what they say next: “Identification determines which is accorded most weight” [2012, 323]. It is unclear who Rosenblum accorded the most weight—had the experiment really been exploring punishment on learning, the experimenter would have obtained tainted ‘data’ and she also ignored the fettered learner’s desperate requests to be freed. For reasons discussed in the Footnote 6, the person we suspect Rosenblum accords the most weight (“identified” most with) was herself: her needs and desires.

5 This explanation is multi-causal because if most subjects completed for this reason, a minority of subjects competed for another or even other quite different reasons [see Russell 2009, 127]. Although this explanation is multi-causal, instead of presenting a wide and potentially confusing array of potential factors or mechanisms (like [Milgram 1963], [Burger 2009]), in line with the advantage with the mono-causal accounts, the above explanation retains elements of exactitude and simplicity: it asserts why most subject complete.

6 We suspect the previously discussed Eleonor Rosenblum was a subject whose behaviour is explicable by the fear that, in order to help the learner by stopping the experiment, she would have to initiate then engage in a direct confrontation with the experimenter. In fear of this solution, she then sought out, regarding the moral dilemma before her, an alternative and more attractive course of action. That is, she came to suspect that if she vocally emphasized the correct answer to the learner, although she would technically taint the ‘accuracy’ of the scientist’s data, doing so would enable her to avoid having to shock the learner (which would thereby obviate any learner-generated pressure that she help them by initiating then engaging in a direct confrontation with the experimenter).

List of illustrations

Bibliographical reference.

Nestar Russell and Raphaël Künstler , “Stanley Milgram’s Obedience Studies: A Critical Review of the Most Influential Explanatory Accounts” ,  Philosophia Scientiæ , 28-2 | 2024, 3-31.

Electronic reference

Nestar Russell and Raphaël Künstler , “Stanley Milgram’s Obedience Studies: A Critical Review of the Most Influential Explanatory Accounts” ,  Philosophia Scientiæ [Online], 28-2 | 2024, Online since 24 May 2024 , connection on 29 October 2024 . URL : http://journals.openedition.org/philosophiascientiae/4328; DOI : https://doi.org/10.4000/11pu0

About the authors

Nestar russell.

University of Calgary (Canada)

By this author

• Stanley Milgram’s Obedience Studies: An Ethical and Methodological Assessment [Full text] Published in Philosophia Scientiæ , 28-2 | 2024

Raphaël Künstler

CLLE–ERRAPHIS, Université de Toulouse (France)

The text only may be used under licence CC BY-NC-ND 4.0 . All other elements (illustrations, imported files) are “All rights reserved”, unless otherwise stated.

Full text issues

• 28-3 | 2024 Kurt Goldstein: Biology, Anthropology, and the Clinic
• 28-2 | 2024 Revisiting Stanley Milgram’s Experiment
• 28-1 | 2024 The Richness and Variety of Neo-Kantianism: Helmholtz, Cassirer, Vaihinger
• 27-3 | 2023 Poincarean studies (II)
• 27-2 | 2023 Poincarean studies (I)
• 27-1 | 2023 The "Vichy parenthesis"? Trajectories of academics and disciplinary reconfigurations during the Occupation and the immediate post-war period
• 26-3 | 2022 Praxeological Gestalts
• 26-2 | 2022 Heritage Building in Mathematics (18th-20th Centuries)
• 26-1 | 2022 La désuétude conceptuelle : abandon ou transformation ?
• 25-3 | 2021 L’analyse dans les mathématiques grecques
• 25-2 | 2021 Mathematics and philosophy in Leibniz through the lens of his unpublished manuscripts
• 25-1 | 2021 The Peano School: Logic, Epistemology and Didactics
• 24-3 | 2020 Readings and Legacy of the Jules Vuillemin's La Philosophie de l’algèbre
• 24-2 | 2020 Resemblance in Contemporary Philosophy
• 24-1 | 2020 Mathematics in Military Academies (18th and 19th Centuries)
• 23-3 | 2019 International Scientific Circulations since the Beginning of the 20th Century: New Perspectives for Study
• 23-2 | 2019 Expérimentation dans les sciences de la nature Expérimentation dans les sciences humaines et sociales
• 23-1 | 2019 Y a-t-il encore de la place en bas ?
• 22-3 | 2018 Sur la philosophie scientifique et l’unité de la science
• 22-2 | 2018 Case studies in social epistemology. Argumentation, public deliberation and collaborative practices
• 22-1 | 2018 Science(s) et édition(s) des années 1780 à l'entre-deux-guerres
• 21-3 | 2017 Don't Believe a Thing! A Tribute to Pascal Engel
• 21-2 | 2017 Raymond Ruyer
• 21-1 | 2017 Homage to Galileo Galilei 1564-2014
• 20-3 | 2016 Le scepticisme selon Jules Vuillemin
• 20-2 | 2016 Circulations et échanges dans l'espace euro-méditerranéen (XVIIIe-XXIe siècles)
• 20-1 | 2016 Le kantisme hors des écoles kantiennes
• 19-3 | 2015 The Bounds of Naturalism
• 19-2 | 2015 Circulations et échanges mathématiques
• 19-1 | 2015 Logic and Philosophy of Science in Nancy (II)
• 18-3 | 2014 Logic and Philosophy of Science in Nancy (I)
• 18-2 | 2014 Hugo Dingler et l’épistémologie pragmatiste en Allemagne
• 18-1 | 2014 Standards of Rigor in Mathematical Practice
• 17-3 | 2013 Tacit and Explicit Knowledge: Harry Collins’s Framework
• 17-2 | 2013 The Mind–Brain Problem in Cognitive Neuroscience
• 17-1 | 2013 The Epistemological Thought of Otto Hölder
• 16-3 | 2012 Alan Turing
• 16-2 | 2012 Modal Matters
• 16-1 | 2012 From Practice to Results in Logic and Mathematics
• 15-3 | 2011 L'espace et le temps
• 15-2 | 2011 La syllogistique de Łukasiewicz
• 15-1 | 2011 Hugh MacColl after One Hundred Years
• 14-2 | 2010 Louis Rougier, De Torricelli à Pascal
• 14-1 | 2010
• 13-2 | 2009
• 13-1 | 2009
• 12-2 | 2008 Normes et santé
• 12-1 | 2008 (Anti-)Realisms: The Metaphysical Issue
• 11-2 | 2007
• 11-1 | 2007 Karl Popper : un philosophe dans le siècle
• 10-2 | 2006 Louis Rougier : vie et œuvre d'un philosophe engagé
• 10-1 | 2006 Jerzy Kalinowski : logique et normativité
• 9-2 | 2005 Aperçus philosophiques en logique et en mathématiques
• 9-1 | 2005
• 8-2 | 2004 Logique & théorie des jeux
• 8-1 | 2004 Le problème de l’incommensurabilité, un demi-siècle après

Cahiers spéciaux

• CS 7 | 2007 Louis Rougier : vie et œuvre d'un philosophe engagé
• CS 6 | 2006 Constructivism: Mathematics, Logic, Philosophy and Linguistics
• CS 5 | 2005 Fonder autrement les mathématiques

Collection numérisée

Présentation.

• The journal
• Guide for authors
• Technical Guide

Informations

• Legal information and credits
• Publishing policies

Appels à contributions

• Appels en cours
• Appels clos

Newsletters

• OpenEdition Newsletter

In collaboration with

• Journal supported by the Institut des sciences humaines et sociales (InSHS) of the CNRS, 2023-2024

Electronic ISSN 1775-4283

Read detailed presentation  

Site map  – Contact  – Legal information and credits  – Syndication

Privacy Policy  – About Cookies  – Report a problem

OpenEdition member  – Published with Lodel  – Administration only

You will be redirected to OpenEdition Search

Stanley Milgram’s Obedience Experiments: A Report Card 50 Years Later

• Published: 09 October 2013
• Volume 50 , pages 623–628, ( 2013 )

Cite this article

• Augustine Brannigan 1  

23k Accesses

4 Citations

18 Altmetric

Explore all metrics

Avoid common mistakes on your manuscript.

Fifty years ago Stanley Milgram published the first report of his studies of obedience to authority. His work (1963) forged the mindset of how social scientists over the next two generations came to explain the participation of hundreds of thousands of Germans in the mass murder of European Jews during the Holocaust. Milgram’s model was Adolph Eichmann who was convicted and executed for his role in the deportation of European Jews to death camps created in Poland for their eradication. Eichmann’s legal defense, that he was ‘just following orders,’ suggested that the final solution to the ‘Jewish problem’ in Europe was engineered by desk murderers remotely positioned in hierarchies of authority across the Nazi bureaucracy. Submission was unquestioned because the decision to eradicate the Jews originated from the sovereign authority. Milgram’s murderers were loyal automatons.

Milgram attracted his subjects from the wider community in New Haven and Bridgeport. He recruited an astounding 780 subjects. His work was identified by Roger Brown as ‘the most important psychological research’ done in his generation. Where Hannah Arendt speculated philosophically that the ranks of Holocaust perpetrators such as Eichmann were unremarkable non-entities, Milgram described in an experimental idiom the ease with which New Haven citizens could be transformed into brutal Nazis without much difficulty. Milgram’s work also provoked questions about the ethical treatment of human subjects in a way that helped to shape future policies for the treatment of volunteers in experimental studies. It alerted funding agencies to the necessity of risk assessment of those deliberately misled in studies premised on subject deception. Milgram also championed the proposition that grave questions of human morality could be examined following the experimental methods that proved to be so effective in the natural sciences. He also contributed to the dogma in social psychology that ‘the situation’ is one of the most, if not the most, important determinant of social behavior. His 1974 book was promoted widely in the popular press, and created a media storm. His scientific portrayal of ‘the banality of evil’ inspired an artistic outpouring of films, and plays, and remains a point of relevance in studies of holocaust history today. Stanley Milgram died in 1984 at the age of 51.

Stanley Milgram 1933-1984

Enter Gina Perry. Perry is an Australian journalist and writer who took an interest in the Milgram study after learning through personal acquaintances that several persons who participated in a replication of the obedience study at La Trobe University in Melbourne in 1973 and 1974 continued to suffer trauma decades later. That unpublished replication involved some 200 subjects. She turned her attention to the original study, and spent 4 years researching the Milgram archives at Yale University. She listened to 140 audio recordings of the original experiments, and dozens of hours of conversations involving former subjects in the postmortem debriefing with a psychiatrist. She interviewed former subjects and experts familiar with the research, family members of the actors, and read the mountains of documentation and correspondence accumulated during the study. The conclusions she draws from her investigation were disturbing, and will fundamentally challenge the way scholars interpret Milgram and his experiment.

What is known about Milgram and his background? He was interested in the conditions that led to the expression of deep antisemitism in Germany during the Second World War. Some thought that pathological conformity might have had national roots. In his doctoral work, he investigated national differences in defiance of group pressure to conform to judgments that subjects thought were incorrect. In this work, he adopted the protocols of Solomon Asch. Indeed, he found national differences in conformity between Norwegian and French subjects, but nothing that illuminated the German case. As a young professor he sought to raise the ante by creating conditions in which subjects were compelled, not just to say something they thought to be untrue, but to act in a gravely inappropriate way. Most students with any postsecondary training in recent years will recall the experiment. The ‘cover story’ was a learning experiment in which potential teachers and learners were recruited from the public by newspaper ads and direct mail solicitation. Persons who presented individually for the study drew names out of a hat for their respective roles. A lab-coated scientist explained the need to determine the effectiveness of punishment on the learning process. The subjects were deceived from the start. The learner (Jim McDonough) and the scientist (John Williams) were amateur actors. Participants were paid $5 for participation and carfare, a very significant compensation at the time in 1961. They witnessed the physical restraint of the learner in an attached room where he was to be ‘tested.’ The shocking appliance consisted of 30 switches numbered from 15 to 450 V. It buzzed and snapped and exuded technological credibility as the teacher administered the punishments up to a level labeled as ‘severe shock.’ To start things, the teacher read a long list of word pairs, and then, under the supervision of Mr. Williams, the scientist, proceeded to test whether the learner retained knowledge of the information just presented to him. Now the drama began in earnest. The learner apparently had a terrible memory. The shocks escalated in accord with the errors. And he began to protest with increasingly painful moans and screams. This was all piped back to the learner through speakers. The response pattern was all predetermined, and designed to create an increasingly disagreeable dilemma for the subject: either shock or walk, obey or defy. Each experiment lasted about 50 min, and resulted in levels of agitation among some subjects that were unprecedented in previous psychological research. Milgram’s question was simply how far the teacher would continue to issue painful shocks before defying Mr. Williams’s directives to continue. Milgram’s critical finding was that 65 % of ordinary persons would administer levels of punishment that would appear to be lethal, even where, in one condition, the learner was depicted as a person with an existing heart condition. Despite this apparently profound level of aggression, the conventional wisdom suggested that the agitation associated with the exercise dissipated immediately as the subjects were ‘de-hoaxed.’ All the civility that had been suspended was restored by the debriefing, and having been made whole again, the subject and the scientist departed company on good terms. That was the myth.

Traumatized subjects

Perry discovered a different picture. Herb Winer was ‘boiling with anger’ for days after the experiment (p. 79). At the time, like Milgram, he was an untenured professor at Yale. He confronted Milgram in his office with his concerns about the experiment, particularly about pressure to shock someone with a heart condition. His trauma was so intense that he confided in Perry, nearly 50 years later, that his memory of the event would be ‘among the last things I will ever forget’ (p. 84). After the cover story was explained, Winer became an admirer of Milgram, ‘although he will never forgive him for what he put him through.’ Bill Lee was another subject tracked down by Perry. Bill Menold was unsure of whether the study was a sham or not, but he found it ‘unbelievably stressful…I was a basket case on the way home’ (p. 52). He confided that night in a neighbor who was an electrician to learn more about electrical shocks. Hannah Bergman (a pseudonym) still recalled the experiment vividly after half a century. Her recollections suggested that she ‘was ashamed—and frightened.’ Her son told Perry that ‘it was a traumatic event in her life which opened some unsettling personal issues with no subsequent follow-up’ (p. 112). A New Haven Alderman complained to Yale authorities about the study: ‘I can’t remember ever being quite so upset’ (p. 132). One subject (#716) checked mortality notices in the New Haven Register , for fear of having killed the learner. Another subject (#501) was shaking so much he was not sure he would be able to drive home; according to his wife, on the way home he was shivering in the car and talked incessantly about his intense discomfort until midnight (p. 95). Subject 711 reported that ‘the experiment left such an effect on me that I spent the night in a cold sweat and nightmares because of fears that I might have killed that man in the chair’ (p. 93). None of the previous histories of these experiments even hinted at such reactions, nor was any of this ever reported in the university curriculum. What caused all the trauma?

To say that the de-hoaxing left a lot to be desired would be a gross understatement. In his first publication, Milgram had written that steps were taken ‘to assure that the subject would leave the laboratory in a state of well-being. A friendly reconciliation was arranged between the subject and the victim, and an effort was made to reduce any tensions that arose as a result of the experiment’ (Milgram 1963 : 374). Also, ‘at the very least all subjects were told that the victim had not received dangerous electric shocks.’ Perry’s review of the archives indicates that this was simply not the case. In fact, Perry reports that 75 % of the subjects were not immediately debriefed in any serious way until the last 4 out of 23 conditions. Perry reports that subjects in conditions one to eighteen, around 600 people, left the lab believing that they had shocked a man, with all that dramatized agony etched on their conscience (p. 92). This was corroborated by Alan Elms, Milgram’s research assistant in the first 4 conditions. ‘For most people who took part, the immediate debrief did not tell them there were no shocks’ (p. 90). In addition, many of the subjects who met after the completion of the study with the psychiatrist, Dr. Paul Errera, similarly reported they received no debriefing at all (p. 89–107).

At minimum a debriefing would have involved an explanation that the scientist and the learner were actors, the shocking appliance was a fake, all the screams were simulated, and that the teachers were the focus of the study. Perry reports that even where some account was given by Milgram to the subjects, they were told that their behaviours, whether obedient or defiant, were natural and understandable, and that the shocking device had been developed to test small animals, and was harmless to people. So even when it occurred, the debriefing, in Perry’s words, ‘turned out to be another fiction’ (p. 90). In addition, the debriefing was remarkably brief—two minutes—and did not involve any question-answer interaction with the experimenter. Milgram did not want future subjects to be contaminated by accounts from prior subjects about the true nature of the experiment, and so he withheld such information until the experiment was virtually over. A fuller explanation was mailed to subjects a year later, but it does not seem to have consoled any of those interviewed by Perry.

The Skeptical Subjects

If many subjects were traumatized, there were significant others who had their doubts about the cover story (p. 156). One subject wrote to Professor Milgram the day after his participation. He had inferred that the ‘draw’ for roles was fixed, and that both pieces of paper probably had the word ‘teacher’ written on them. He found the learner unaccountably ‘disinterested’, and was suspicious of all the one-way glass mirrors. He also noticed that the learner was not given his check at the same time as himself. Another noticed that the learner’s check was dog-eared from what appeared to be frequent use. Others engaged in reality testing by asking the learner to tap on the wall if he could hear him. No response. One lowered the shock level intentionally, and the learner seemed to express increased pain despite this. Others were simply sceptical that Yale would permit anyone to absorb such punishment. Some commented on the fact that no one with a cardiac condition which was under medical surveillance would submit to such intense agitation. Another noted that there was a speaker in the learner’s room, and the sound from the voice did not appear to be coming through the door, as he would have expected. And many suggested that the sounds appeared to be audio recordings. All this was noted in the archives. Under these conditions, the subjects simply played along as required by the experiment, since they assumed that no one would purposely be hurt, and it was all for the good of science.

Milgram was aware of this scepticism, but he dismissed it as a reaction formation. He reasoned that the subjects had acted shamefully, then, in self-defense, they denied anyone was injured, and that they had not done any harm. Perry also turned up a report by a research assistant, Taketo Murata. He examined the maximum shock levels in all 23 conditions, but divided the subjects into those who appeared to be doubters, and those who appeared to be true believers. ‘Taketo found that people most likely to disobey and give lower-voltage shocks were those who said they believed someone really was being hurt’ (p. 164). Perry comments further that ‘only half of the people who undertook the experiment fully believed it was real, and of those, two-thirds disobeyed the experimenter’ (p. 163). There was another area of information leakage that must have piqued the curiosity of some teachers. There were numerous cases where the subjects practically shouted out the correct answer to the learner, but this communication never made a difference in his response. Also, numerous teachers, frustrated by the learner’s poor performance, offered to switch places during the experiment, but again, this offer did not attract any interest or response. This did not always result in outright disbelief, but created some suspicions that things were not exactly as they seemed.

The Secret Experiment

In his 1974 book, Obedience to Authority , Milgram gave the fullest account of the various conditions he investigated. He reported 18 conditions. In the archives, Perry came across 24, although one, the ‘Educated Opinion’ condition, was not actually an experiment, but an estimation by psychiatrists and university students of the probability that average subjects would be fully obedient given the conditions described to them. Among the unpublished investigations, Perry discovered a remarkable condition that Milgram had kept secret. This was the study of ‘intimate relationships.’ Twenty pairs of people were recruited on the basis of a pre-existing intimacy. They were family members, fathers and sons, brothers-in-law, and good friends. One was randomly assigned to the teacher role, the other to the learner role. After the learners were strapped into the restraining device, Milgram privately explained the ruse to them, and encouraged them to vocalize along the lines employed by the actor in response to the shocks in previous conditions. The ‘intimate relationships’ study produced one of the highest levels of defiance of any condition: 85 %. It also produced a great deal of agitation to teachers as the learners begged their friends or family members by name to be released. One subject (#2435) went ballistic with the scientist’s pressure, and started shouting at him for encouraging him to injure his own son.

Perry speculated that Milgram was ambivalent about this condition for two reasons. On the one hand, ‘Milgram might have kept it secret because he realized that what he asked subjects to do in Condition 24 might be difficult to defend’ (p. 202). After all, he abused their mutual trust and intimacy to turn the one against the other. On the other hand, the results countered the whole direction of Milgram’s argument about the power of bureaucracy. Perry found a note in the archives in which Milgram confessed that ‘within the context of this experiment, this is as powerful a demonstration of disobedience that can be found.’ When people believed that someone was being hurt, and that it was someone close to them, ‘they refused to continue’ (p. 202). Given its implication, the finding was never reported.

This suggests that, to an extent, Milgram cherry-picked his results for impact. Perry notes that Milgram worked to produce the astonishingly high compliance rate of 65 %. He assumed that he needed a plurality of his subjects, but not a figure so high that it begged credibility. In pilot studies he tweaked the design repeatedly. At first, there was no verbal feedback from the learner, and every subject, when commanded, went indifferently to the maximum shock. Such a response would suggest that subjects did not actually assume they were doing anything harmful. The verbal feedback from the learner was introduced to create resistance. Milgram also explored a number of Stress Reducing Mechanisms and Binding Factors to optimize compliance. Stress was reduced, for example, by framing the actions as part of a legitimate learning experiment, and by advising the subjects that there was no permanent damage from the shocks. The binding factors included the gradual 30 step increments from the lowest to the highest shock level on the supposition that once they started, the movement up the shock scale would signal their acceptance of the protocol one step at a time.

Perry also found that there was often a Mexican standoff between the subjects and Mr. Williams as to their point of defiance. This was particularly evident in the all-female design. In their histories of the experiments Blass ( 2004 ) and Miller ( 1986 ) created the impression that the scientist would use 4 specific prods to encourage the subjects to continue, since that was what Milgram published. ‘If the Subject still refused after this last [fourth] prod, the experiment was discontinued’ (Blass 2004 : 85). The subjects were always free to break off. After listening to the Female Condition (condition 20), Perry concluded: ‘this isn’t what the tapes showed’ (p. 136). Mr. Williams did not adhere strictly to the protocol. This was reflected in postmortem interviews with Dr. Errera, where three women from the Female Condition suggested that they had been ‘railroaded’ by Williams (p. 135). He would not relent in his pressure. In one case (#2026), he brought the subject a cup of coffee while she sat idle for 30 min before succumbing to repeated pressure to continue. Another subject was prompted a total of 26 times. This suggests, not only that the results could be cherry-picked between conditions, but also that in any one condition the scientist could elevate the compliance rate by departure from the protocol and the relentless application of pressure. The resulting 65 % compliance in condition 20 was equivalent to the previous highs achieved in two earlier conditions. In the remote feedback design (condition 1), the victim apparently pounded on the wall to signal distress. This reduced compliance from 100 to 65 %. In the cardiac condition, all the elaborate moaning, screaming and demands to be released (condition 5), resulted surprisingly in the same 65 % compliance level. How could such radically dissimilar feedback result in identical levels of compliance? This might be explained in part by the degree to which the scientist adhered strictly to, or departed from, the 4 prod protocol. As Perry’s analysis of the Female Condition suggests, the various treatments were simply not standardized. Milgram’s conclusion that there were no gender differences in aggression based on a comparison of outcomes in condition 5 and condition 20 does not bear scrutiny.

In his face-to-face dealing with subjects, Milgram assured them that their reactions were normal and understandable. Yet in his book he describes the compliant subjects as acting in ‘a shockingly immoral way’ (1974: 194). In his notes, he describes them as ‘moral imbeciles’ capable of staffing ‘death camps’ (Perry 2012 : 260). In the 1974 coverage of his book on the CBS network ’60 minutes’ program, he portrays the compliant subjects as New Haven Nazis (p. 369), and asserts that one would be able to staff a system of death camps in America with enough people recruited from medium-sized American towns.

The Obedience Legacy

What are the implications of Perry’s critique for the place Milgram is accorded in the canons of historical psychology? After all, we are told that Milgram’s results were essentially replicated in 2009. I offer five observations. First, there are many shortcomings in Perry’s work. Her evidence is anecdotal in the sense that she was not able to canvas systematically large numbers of former subjects, and after 50 years, this is not surprising. She did have access to all the postmortem interviews of former subjects with psychiatrist Paul Errera, but only 120 of the original 780 subjects were asked to attend these meetings, and only 32 did. Of the 780 individual cases, only 140 audio-recordings were available for Perry’s use. Nonetheless, there was convergence among the surviving subjects about the enduring levels of trauma from which they continued to suffer, both in the US and Australia. She also discovered that the majority of subjects were not appropriately debriefed in a timely manner. From the archival materials, there was a mountain of evidence of scepticism among subjects who were not deceived by the cover story. This was a fact Milgram stubbornly refused to acknowledge. Additionally, she discovered a totally new, and previously unreported condition, the ‘intimate relationships’ study, which dramatically altered the significance of the entire experimental initiative. None of these findings were reported in the previous histories of the experiment by Blass or Miller. The value of her book is to question the scientific and moral significance of the obedience study as it was originally reported.

Second, her analysis of Milgram’s formal advocacy of the primacy of ‘the situation’ with the harsh moral judgments Milgram offered in print and in privacy regarding the obedient subjects illuminates the enormous disconnect between Milgram’s scientific posturing and his conclusions about the defects of those who obeyed. Perry: ‘He associated obedient behaviour with lower intelligence, less education, and the working classes’ (p. 298). Defiant subjects were smart, educated and middle class. The obedient were impervious to the suffering they caused, were remorseless and ‘unthinkingly obedient’ (p. 243). Yet the behavior exhibited in the lab was frequently marked by deep anguish and empathy, even by persons compelled to obey, not by ‘the situation,’ but by the relentless badgering by Williams, their own self-doubts, and a sense that all was done for defensible, scientific reasons. Just as Kant’s scientist must presuppose space and time before the laws of physics and chemistry can be deduced, the psychologist must presuppose that all human experience occurs in situations. But situations do not explain differences in behaviour.

Third, the recent historiography of the Holocaust in the work of Browning and Goldhagen emphasized the agency of the ordinary perpetrators. These were persons who acted, not out of fear of reprisal from superiors, not out of duress or fear of disobedience, nor were they blindly obedient. They acted out of a sense of duty and responsibility. Cesarani’s and Lipstadt’s accounts of Eichmann depict a man who was the Final Solution’s greatest advocate. These accounts stand in dramatic contrast to the banality of evil that Milgram’s work perpetuated, the image of automatons who had no moral agency once they put on a uniform. The experimental account he attempted to construct actually misrepresented the original phenomenon, just as the accounts of his subjects suppressed their agency, and their struggles with him in his own lab.

Fourth, anytime someone raises questions about the validity or reliability of the Milgram experiment, he or she is told that it has already been replicated all over the world, so that criticizing Milgram is essentially a waste of time. Milgram reported that his work had been replicated in Australia, Germany, Italy, and South Africa, suggesting the findings were universal. However, according to Perry, ‘the Australian study found significantly lower levels of obedience than Milgram’s’, as did the Italian and Germany studies; the South African study was a student report based on 16 subjects (p. 307). In 2009 Jerry Burger reported in American Psychologist that he had partially replicated Milgram’s condition 5 (the cardiac condition). His original study was financed by the ABC network as a reality TV show, and was designed to produce comparable results. It was not designed to examine any of the methodological questions raised by Perry. In a second co-authored publication, Burger suggested that the central phenomenon he studied was something other than obedience, since the prods that appeared to look most like direct commands were ones that were singularly ineffective in producing compliance. In addition, his subjects were told that the learner, like themselves, could disengage for any reason at any point in the process. What could one infer from all that hollering that failed to result in the learner ever quitting?

Last point. Social psychology is an awkward discipline since it occupies a cognitive space already filled with common sense, the judgment of ages, and insight derived from lived experience. It has undertaken to replace this sort of human knowledge with something based on rationality derived from adherence to the experimental method. In Rodrigues and Levine’s 100 Years of Experimental Social Psychology (1999), the contributors came to two conclusions. First, they repeatedly noted that the field had not resulted in a significant body of cumulative, non-trivial knowledge. Second, whatever it was that formed the core of the discipline did not share substantial consensus among its contributors (Brannigan 2002 ). Perry’s disturbing investigation of the Milgram archives will not change any of that.

Further Reading

Blass, T. 2004. The Man who Shocked the World: the life and legacy of Stanley Milgram . New York: Basic Books.

Google Scholar  

Burger, J., Girgis, Z. M., & Manning, C. C. 2011. In their own words: Explaining obedience to authority through an examination of participants’ comments. Social Psychological and Personality Science , 2 , 460–466.

Article   Google Scholar  

Brannigan, A. 2002. The Experimental Turn in Social Psychology. Society: Social Science and Modern Society , 39 (5), 74–79.

Perry, Gina. 2012. Behind the Shock Machine: The untold story of the notorious Milgram psychology experiments , Melbourne and London: Scribe Books. New York: The New Press, 2013.

Milgram, S. 1963. Behavioral Study of Obedience. Journal of Abnormal Psychology , 67 , 371–78.

Milgram, S. 1974. Obedience to Authority . New York: Harper & Row.

Miller, A. G. 1986. The Obedience Experiments: A case study of controversy in social science . New York: Praeger.

Download references

Author information

Authors and affiliations.

Department of Sociology, University of Calgary, 2500 University Dr NW, Calgary, Alberta, Canada, T2N 1N4

Augustine Brannigan

You can also search for this author in PubMed   Google Scholar

Corresponding author

Correspondence to Augustine Brannigan .

Rights and permissions

Reprints and permissions

About this article

Brannigan, A. Stanley Milgram’s Obedience Experiments: A Report Card 50 Years Later. Soc 50 , 623–628 (2013). https://doi.org/10.1007/s12115-013-9724-3

Download citation

Published : 09 October 2013

Issue Date : December 2013

DOI : https://doi.org/10.1007/s12115-013-9724-3

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

• Find a journal
• Publish with us
• Track your research

Rethinking One of Psychology's Most Infamous Experiments

In the 1960s, Stanley Milgram's electric-shock studies showed that people will obey even the most abhorrent of orders. But recently, researchers have begun to question his conclusions — and offer some of their own.

In 1961, Yale University psychology professor Stanley Milgram placed an advertisement in the New Haven Register . “We will pay you $4 for one hour of your time,” it read, asking for “500 New Haven men to help us complete a scientific study of memory and learning.”

Only part of that was true. Over the next two years, hundreds of people showed up at Milgram’s lab for a learning and memory study that quickly turned into something else entirely. Under the watch of the experimenter, the volunteer—dubbed “the teacher”—would read out strings of words to his partner, “the learner,” who was hooked up to an electric-shock machine in the other room. Each time the learner made a mistake in repeating the words, the teacher was to deliver a shock of increasing intensity, starting at 15 volts (labeled “slight shock” on the machine) and going all the way up to 450 volts (“Danger: severe shock”). Some people, horrified at what they were being asked to do, stopped the experiment early, defying their supervisor’s urging to go on; others continued up to 450 volts, even as the learner pled for mercy, yelled a warning about his heart condition—and then fell alarmingly silent. In the most well-known variation of the experiment, a full 65 percent of people went all the way.

Until they emerged from the lab, the participants didn’t know that the shocks weren’t real, that the cries of pain were pre-recorded, and that the learner—railroad auditor Jim McDonough —was in on the whole thing, sitting alive and unharmed in the next room. They were also unaware that they had just been used to prove the claim that would soon make Milgram famous: that ordinary people, under the direction of an authority figure, would obey just about any order they were given, even to torture. It’s a phenomenon that’s been used to explain atrocities from the Holocaust to the Vietnam War’s My Lai massacre to the abuse of prisoners at Abu Ghraib. “To a remarkable degree,” Peter Baker wrote in Pacific Standard in 2013, “Milgram’s early research has come to serve as a kind of all-purpose lightning rod for discussions about the human heart of darkness.”

In some ways, though, Milgram’s study is also—as promised—a study of memory, if not the one he pretended it was.

More than five decades after it was first published in the Journal of Abnormal and Social Psychology in 1963, it’s earned a place as one of the most famous experiments of the 20th century. Milgram’s research has spawned countless spinoff studies among psychologists, sociologists, and historians, even as it’s leapt from academia into the realm of pop culture. It’s inspired songs by Peter Gabriel (lyrics: “We do what we’re told/We do what we’re told/Told to do”) and Dar Williams (“When I knew it was wrong, I played it just like a game/I pressed the buzzer”); a number of books whose titles make puns out of the word “shocking”; a controversial French documentary disguised as a game show ; episodes of Law and Order and Bones ; a made-for-TV movie with William Shatner; a jewelry collection (bizarrely) from the company Enfants Perdus; and most recently, the biopic The Experimenter , starring Peter Sarsgaard as the title character—and this list is by no means exhaustive.

But as with human memory, the study—even published, archived, enshrined in psychology textbooks—is malleable. And in the past few years, a new wave of researchers have dedicated themselves to reshaping it, arguing that Milgram’s lessons on human obedience are, in fact, misremembered—that his work doesn’t prove what he claimed it does.

The problem is, no one can really agree on what it proves instead.

To mark the 50th anniversary of the experiments’ publication (or, technically, the 51st), the Journal of Social Issues released a themed edition in September 2014 dedicated to all things Milgram. “There is a compelling and timely case for reexamining Milgram’s legacy,” the editors wrote in the introduction, noting that they were in good company: In 1964, the year after the experiments were published, fewer than 10 published studies referenced Milgram’s work; in 2012, that number was more than 60.

It’s a trend that surely would have pleased Milgram, who crafted his work with an audience in mind from the beginning. “Milgram was a fantastic dramaturg. His studies are fantastic little pieces of theater. They’re beautifully scripted,” said Stephen Reicher, a professor of psychology at the University of St. Andrews and a co-editor of the Journal of Social Issues ’ special edition. Capitalizing on the fame his 1963 publication earned him, Milgram went on to publish a book on his experiments in 1974 and a documentary, Obedience , with footage from the original experiments.

But for a man determined to leave a lasting legacy, Milgram also made it remarkably easy for people to pick it apart. The Yale University archives contain boxes upon boxes of papers, videos, and audio recordings, an entire career carefully documented for posterity. Though Milgram’s widow Alexandra donated the materials after his death in 1984, they remained largely untouched for years, until Yale’s library staff began to digitize all the materials in the early 2000s. Able to easily access troves of material for the first time, the researchers came flocking.

“There’s a lot of dirty laundry in those archives,” said Arthur Miller, a professor emeritus of psychology at Miami University and another co-editor of the Journal of Social Issues . “Critics of Milgram seem to want to—and do—find material in these archives that makes Milgram look bad or unethical or, in some cases, a liar.”

One of the most vocal of those critics is Australian author and psychologist Gina Perry, who documented her experience tracking down Milgram’s research participants in her 2013 book Behind the Shock Machine: The Untold Story of the Notorious Milgram Psychology Experiments . Her project began as an effort to write about the experiments from the perspective of the participants—but when she went back through the archives to confirm some of their stories, she said, she found some glaring issues with Milgram’s data. Among her accusations: that the supervisors went off script in their prods to the teachers, that some of the volunteers were aware that the setup was a hoax, and that others weren’t debriefed on the whole thing until months later. “My main issue is that methodologically, there have been so many problems with Milgram’s research that we have to start re-examining the textbook descriptions of the research,” she said.

But many psychologists argue that even with methodological holes and moral lapses, the basic finding of Milgram’s work, the rate of obedience, still holds up. Because of the ethical challenge of reproducing the study, the idea survived for decades on a mix of good faith and partial replications—one study had participants administer their shocks in a virtual-reality system, for example—until 2007, when ABC collaborated with Santa Clara University psychologist Jerry Burger to replicate Milgram’s experiment for an episode of the TV show Basic Instincts titled “ The Science of Evil ,” pegged to Abu Ghraib.

Burger’s way around an ethical breach: In the most well-known experiment, he found, 80 percent of the participants who reached a 150-volt shock continued all the way to the end. “So what I said we could do is take people up to the 150-volt point, see how they reacted, and end the study right there,” he said. The rest of the setup was nearly identical to Milgram’s lab of the early 1960s (with one notable exception: “Milgram had a gray lab coat and I couldn’t find a gray, so I got a light blue.”)

At the end of the experiment, Burger was left with an obedience rate around the same as the one Milgram had recorded—proving, he said, not only that Milgram’s numbers had been accurate, but that his work was as relevant as ever. “[The results] didn’t surprise me,” he said, “but for years I had heard from my students and from other people, ‘Well, that was back in the 60s, and somehow how we’re more aware of the problems of blind obedience, and people have changed.’”

In recent years, though, much of the attention has focused less on supporting or discrediting Milgram’s statistics, and more on rethinking his conclusions. With a paper published earlier this month in the British Journal of Social Psychology , Matthew Hollander, a sociology Ph.D. candidate at the University of Wisconsin, is among the most recent to question Milgram’s notion of obedience. After analyzing the conversation patterns from audio recordings of 117 study participants, Hollander found that Milgram’s original classification of his subjects—either obedient or disobedient—failed to capture the true dynamics of the situation. Rather, he argued, people in both categories tried several different forms of protest—those who successfully ended the experiment early were simply better at resisting than the ones that continued shocking.

“Research subjects may say things like ‘I can’t do this anymore’ or ‘I’m not going to do this anymore,’” he said, even those who went all the way to 450 volts. “I understand those practices to be a way of trying to stop the experiment in a relatively aggressive, direct, and explicit way.”

It’s a far cry from Milgram’s idea that the capacity for evil lies dormant in everyone, ready to be awakened with the right set of circumstances. The ability to disobey toxic orders, Hollander said, is a skill that can be taught like any other—all a person needs to learn is what to say and how to say it.

In some ways, the conclusions Milgram drew were as much a product of their time as they were a product of his research. At the time he began his studies, the trial of Adolf Eichmann, one of the major architects of the Holocaust, was already in full swing. In 1963, the same year that Milgram published his studies, writer Hannah Arendt coined the phrase “the banality of evil” to describe Eichmann in her book on the trial, Eichmann in Jerusalem .

Milgram, who was born in New York City in 1933 to Jewish immigrant parents, came to view his studies as a validation of Arendt’s idea—but the Holocaust had been at the forefront of his mind for years before either of them published their work. “I should have been born into the German-speaking Jewish community of Prague in 1922 and died in a gas chamber some 20 years later,” he wrote in a letter to a friend in 1958. “How I came to be born in the Bronx Hospital, I’ll never quite understand.”

And in the introduction of his 1963 paper, he invoked the Nazis within the first few paragraphs: “Obedience, as a determinant of behavior, is of particular relevance to our time,” he wrote. “Gas chambers were built, death camps were guarded; daily quotas of corpses were produced … These inhumane policies may have originated in the mind of a single person, but they could only be carried out on a massive scale if a very large number of persons obeyed orders.”

Though the term didn’t exist at the time, Milgram was a proponent of what today’s social psychologists call situationism: the idea that people’s behavior is determined largely by what’s happening around them. “They’re not psychopaths, and they’re not hostile, and they’re not aggressive or deranged. They’re just people, like you and me,” Miller said. “If you put us in certain situations, we’re more likely to be racist or sexist, or we may lie, or we may cheat. There are studies that show this, thousands and thousands of studies that document the many unsavory aspects of most people.”

But continued to its logical extreme, situationism “has an exonerating effect,” he said. “In the minds of a lot of people, it tends to excuse the bad behavior … it’s not the person’s fault for doing the bad thing, it’s the situation they were put in.” Milgram’s studies were famous because their implications were also devastating: If the Nazis were just following orders, then he had proved that anyone at all could be a Nazi. If the guards at Abu Ghraib were just following orders, then anyone was capable of torture.

The latter, Reicher said, is part of why interest in Milgram’s work has seen a resurgence in recent years. “If you look at acts of human atrocity, they’ve hardly diminished over time,” he said, and news of the abuse at Abu Ghraib was surfacing around the same time that Yale’s archival material was digitized, a perfect storm of encouragement for scholars to turn their attention once again to the question of what causes evil.

He and his colleague Alex Haslam, the third co-editor of The Journal of Social Issues ’ Milgram edition and a professor of psychology at the University of Queensland, have come up with a different answer. “The notion that we somehow automatically obey authority, that we are somehow programmed, doesn’t account for the variability [in rates of obedience] across conditions,” he said; in some iterations of Milgram’s study, the rate of compliance was close to 100 percent, while in others it was closer to zero. “We need an account that can explain the variability—when we obey, when we don’t.”

“We argue that the answer to that question is a matter of identification,” he continued. “Do they identify more with the cause of science, and listen to the experimenter as a legitimate representative of science, or do they identify more with the learner as an ordinary person? … You’re torn between these different voices. Who do you listen to?”

The question, he conceded, applies as much to the study of Milgram today as it does to what went on in his lab. “Trying to get a consensus among academics is like herding cats,” Reicher said, but “if there is a consensus, it’s that we need a new explanation. I think nearly everybody accepts the fact that Milgram discovered a remarkable phenomenon, but he didn’t provide a very compelling explanation of that phenomenon.”

What he provided instead was a difficult and deeply uncomfortable set of questions—and his research, flawed as it is, endures not because it clarifies the causes of human atrocities, but because it confuses more than it answers.

Or, as Miller put it: “The whole thing exists in terms of its controversy, how it’s excited some and infuriated others. People have tried to knock it down, and it always comes up standing.”

About the Author

More Stories

The Definitive April Fool's Prank Bracket

The State Department Is Spring Breaking, Badly

Milgram's Obedience to Authority experiments: origins and early evolution

Affiliation.

• 1 Victoria University of Wellington, New Zealand. [email protected]
• PMID: 21366616
• DOI: 10.1348/014466610X492205

Stanley Milgram's Obedience to Authority experiments remain one of the most inspired contributions in the field of social psychology. Although Milgram undertook more than 20 experimental variations, his most (in)famous result was the first official trial run - the remote condition and its 65% completion rate. Drawing on many unpublished documents from Milgram's personal archive at Yale University, this article traces the historical origins and early evolution of the obedience experiments. Part 1 presents the previous experiences that led to Milgram's conception of his rudimentary research idea and then details the role of his intuition in its refinement. Part 2 traces the conversion of Milgram's evolving idea into a reality, paying particular attention to his application of the exploratory method of discovery during several pilot studies. Both parts illuminate Milgram's ad hoc introduction of various manipulative techniques and subtle tension-resolving refinements. The procedural adjustments continued until Milgram was confident that the first official experiment would produce a high completion rate, a result contrary to expectations of people's behaviour. Showing how Milgram conceived of, then arrived at, this first official result is important because the insights gained may help others to determine theoretically why so many participants completed this experiment.

©2010 The British Psychological Society.

Publication types

• Historical Article
• Authoritarianism*
• Cooperative Behavior*
• Dominance-Subordination*
• History, 20th Century
• Power, Psychological*
• Psychological Theory*
• Psychology, Social / history*
• Punishment / history*
• Research / history*
• United States

Personal name as subject

• Stanley Milgram

• Architecture and Design
• Asian and Pacific Studies
• Business and Economics
• Classical and Ancient Near Eastern Studies
• Computer Sciences
• Cultural Studies
• Engineering
• General Interest
• Geosciences
• Industrial Chemistry
• Islamic and Middle Eastern Studies
• Jewish Studies
• Library and Information Science, Book Studies
• Life Sciences
• Linguistics and Semiotics
• Literary Studies
• Materials Sciences
• Mathematics
• Social Sciences
• Sports and Recreation
• Theology and Religion
• Publish your article
• The role of authors
• Promoting your article
• Abstracting & indexing
• Publishing Ethics
• Why publish with De Gruyter
• How to publish with De Gruyter
• Our book series
• Our subject areas
• Your digital product at De Gruyter
• Contribute to our reference works
• Product information
• Tools & resources
• Product Information
• Promotional Materials
• Orders and Inquiries
• FAQ for Library Suppliers and Book Sellers
• Repository Policy
• Free access policy
• Open Access agreements
• Database portals
• For Authors
• Customer service
• People + Culture
• Journal Management
• How to join us
• Working at De Gruyter
• Mission & Vision
• De Gruyter Foundation
• De Gruyter Ebound
• Our Responsibility
• Partner publishers

Your purchase has been completed. Your documents are now available to view.

Milgram für Historiker

Reichweite und grenzen einer übertragung des milgram-experiments auf den nationalsozialismus.

• Thomas Sandkühler and Hans-Walter Schmuhl

Stanley Milgram was the first who tried to apply the results of his experiment on National Socialism. Historical science has hardly picked up on this subject with the exception of the American historian Christopher Browning. Despite of some serious problems which have occured by transferring the Milgram-experiment onto National Socialism we are convinced that the possibilities Milgram has opened up for contemporary history have not been exhausted yet. In this connection we would like to plead for a stronger distinction of types of perpetrators, taking into account the latest results in criminology. The Milgram-experiment refers methodically to local studies of massacres and genocides. Its application on the bureaucracy of destruction seems particularly promising to us. Also, there should be included the rescuers of jews into the research on perpetrators as a controlling body.

© 1998 by Lucius & Lucius, Stuttgart

• X / Twitter

Supplementary Materials

Please login or register with De Gruyter to order this product.

Journal and Issue

Articles in the same issue.