1.1.2 - Milgram and situational dissent

The original obedience procedure

Milgram's real aim was to investigate destructive obedience, although participants were told that the research concerned punishment and learning.

The 1963 report described 40 men aged 20–50 from New Haven and nearby communities in the USA. Newspaper advertisements and direct mail attracted volunteers from varied occupations and educational backgrounds. They received **$4.50** for coming to the laboratory at Yale University, and were told the money remained theirs whatever happened. This was a volunteer sample, not a random sample of the population. [Original 1963 report](https://irbmember.web.unc.edu/wp-content/uploads/sites/7878/2017/01/1963_Milgram_BehavioralStudy.pdf)$.

A participant met an experimenter in a grey technician's coat and a man presented as another volunteer. That man was actually a confederate, someone working with the researchers. A rigged draw always made the genuine participant the “teacher” and the confederate the “learner”. The teacher saw the learner strapped into a chair with an electrode attached, then moved to an adjacent room. A real 45 V sample shock helped make the apparatus convincing. The learner received no real shocks during the task.

The teacher read word pairs and tested the learner's memory. After an incorrect answer, the teacher was instructed to operate a shock switch and increase the level by 15 V on the next error. Thirty labelled switches ran from 15 V to 450 V, with warnings suggesting severe danger at the upper levels. The learner's answers and protests were scripted, allowing different participants to face a similar sequence.

In the particular condition described in the 1963 paper, the learner made no audible protest until 300 V, when he banged on the wall and stopped answering. He banged again at 315 V, then became silent. The experimenter instructed the teacher to treat non-response as an incorrect answer and continue. Do not combine this sequence with the earlier spoken protests used in later conditions.

When a teacher hesitated, the experimenter used a sequence of four prods: a request to continue, an appeal to the experiment's requirements, an assertion that continuing was essential, and finally an assertion that there was no choice. If the participant still refused after the fourth prod, the session ended. These escalating statements made withdrawal psychologically difficult even though participants could physically refuse.

The main dependent measure was the maximum shock level administered before refusal. “Fully obedient” meant reaching the maximum 450 V, not that a participant approved of the procedure or felt no distress. The research studied behaviour toward a person believed to be receiving shocks, not actual willingness to harm a known unshocked actor.

Findings and the comparison baseline

The obedience result was striking, but understanding exactly which condition produced it matters.

In the 1963 report, 26 of 40 men reached 450 V, giving 65% full obedience. The other 14 stopped after administering at least 300 V. Participants often sweated, trembled, stuttered or laughed nervously. Distress could accompany obedience; nervous laughter is not evidence that someone enjoyed harming the learner. The finding challenged predictions that almost nobody would continue to the maximum, but it also showed that a substantial minority resisted.

The variations are reported in Milgram's 1974 book, with experiment numbers that identify different conditions rather than publication dates. Experiment 5, the new baseline for subsequent comparisons, retained the teacher, learner, experimenter and increasing-shock task at Yale. It used spoken learner protests, including an earlier mention of a heart condition and a demand for release at 150 V. The laboratory was plainer than the original setting. Again, 26 of 40 reached 450 V (65%). The matching percentage does not make the 1963 condition and Experiment 5 identical. Original book scan.

The three specified variations ask different questions. Experiment 7 changes the authority's physical presence, Experiment 10 changes institutional context, and Experiment 13 changes who issues the increasing-shock instruction. Their participants were different people, so we compare conditions rather than claim the same individual's obedience changed.

A source detail needs care: the original book's Experiment 7 table prints 20.5%, a figure repeated in teaching accounts, but both its count and narrative give 9 of 40, which calculates as 22.5%. Use “about one fifth” for the substantive comparison and identify which reported value you are using if discussing this discrepancy. It does not change the strong reduction from the 65% baseline.

A percentage expresses the count as a share of its total: for the baseline, 26 ÷ 40 × 100 = 65%. A difference between percentages is measured in percentage points. The baseline-to-office comparison is 65 − 47.5 = 17.5 percentage points. This describes the observed difference; it does not establish statistical significance or identify its cause.

Milgram concluded that powerful situational pressures can produce obedience among ordinary volunteers. A percentage is an outcome in a particular procedure; it is not the percentage of humanity that would follow every harmful command.

Telephonic instructions and covert resistance

Experiment 7 investigated whether obedience depended on the experimenter being physically present beside the teacher.

After giving initial instructions, the experimenter left the laboratory and issued further orders by telephone. The increasing-shock task otherwise followed the relevant baseline procedure. Nine of 40 participants reached the maximum, much fewer than the baseline's 26. This supports the conclusion that a nearby authority and direct surveillance helped sustain obedience.

Some participants also resisted covertly. They continued the task but administered lower shocks than ordered, sometimes assuring the experimenter by telephone that they were following the required progression. They did not always openly reject the authority. This matters because resistance is broader than an explicit refusal: a person can undermine an instruction while appearing compliant.

Social impact theory interprets the reduction through lower immediacy: the source of pressure is farther away and separated by a telephone. Agency theory can interpret face-to-face supervision as reinforcing the authority relationship. However, distance and surveillance changed together. The study does not show that spatial closeness alone, independent of monitoring or confrontation, caused the whole difference.

A fictional worker who challenges an unsafe instruction once a supervisor leaves illustrates the mechanism. It is an application of the research, not evidence that workplaces reproduce the experiment's exact percentage.

Institutional context in the rundown office block

Experiment 10 examined whether Yale's reputation was necessary for the high obedience found in the laboratory.

Milgram moved the apparatus to a sparsely furnished office suite in a somewhat rundown commercial building in Bridgeport. Participants were recruited through a mail invitation, with the same age and occupational distributions as at Yale. There was no visible university connection: the study appeared to be run by the invented organisation “Research Associates of Bridgeport”. The personnel and relevant procedure were retained, and participants received the same payment.

Nineteen of 40 reached 450 V (47.5%), compared with 65% in the Yale baseline. Some expressed doubts about the organisation's legitimacy. The direction of the difference is consistent with prestigious institutional backing increasing perceived authority. It therefore suggests one situational route to dissent: questioning whether an organisation is entitled to make its demands.

The conclusion needs qualification. Nearly half still obeyed fully without Yale's visible backing, so that backing was not necessary for every participant. Milgram reported that this reduction was not statistically significant in his comparison. The observed difference therefore should not be described as conclusive proof that institutional prestige caused a population-level reduction. Moving city, setting and apparent organisation also changed a bundle of contextual cues rather than isolating one visual feature of a building.

The ordinary man and the source of orders

Experiment 13 separated the increasing-shock instruction from the formal experimenter's authority.

Three apparent volunteers arrived: the genuine participant and two confederates. The genuine participant became the teacher, one confederate the learner, and the other a person recording timings. The experimenter set up the task and sample shocks but did not specify the progressive shock levels. A staged telephone call took him away. The ordinary participant then proposed increasing shocks by one step after each error and insisted that the teacher follow this system.

Four of 20 teachers reached the maximum (20%); 16 resisted the ordinary man's instructions. The result suggests that an instruction's source matters: insistence from an apparent fellow participant was less effective than direction from the experimenter. Perceived legitimacy and status, rather than the instruction's content alone, help explain obedience.

This was not simply the usual experimenter changing clothes. An additional person, the experimenter's departure and the apparent invention of the shock rule changed the situation. Milgram also acknowledged that the departure could make the staging less credible. The ordinary man was not completely free of authority either, because the experimenter had established the research and the general use of shocks. Thus the study supports a role for authoritative source but cannot assign the entire reduction to clothing or status alone.

Evaluating the research

Milgram's research provides evidence about obedience, while its methods constrain how far that evidence can be generalised.

The scripted learner responses, labelled switches and planned prods helped standardise the task. Measuring maximum shock level made obedience operationally clear. Changing situational features allowed more informative comparisons than simply asking volunteers whether they would obey. Nevertheless, comparisons involved different samples and some manipulations altered several cues at once, so a result should be linked to the actual manipulation rather than a single assumed cause.

There are serious ethical problems. Participants were deceived about the real aim, the other volunteers and the shocks. They could not give fully informed consent to an obedience test, and substantial distress threatened protection from psychological harm. The final prod's assertion that there was no choice undermined a meaningful right to withdraw. Stopping after continued refusal, telling participants their payment was secure and arranging explanation/reassurance and reconciliation with the learner were safeguards, but later reassurance does not erase distress experienced during the procedure. This historical method is not an appropriate learner practical to repeat.

The sample included more occupations than a student-only study, but the original 40 volunteers were all men from one American area in the early 1960s. Volunteering for research at Yale may have selected people who trusted science or institutions. Neither occupational variety nor a clear result makes the sample representative of all genders, countries or periods.

The laboratory situation created a believable moral conflict for many participants, suggested by their distress. Yet administering switches for a learning experiment differs from obeying a manager, military officer or political regime over a long period. Ecological validity concerns this fit to the behaviour being explained. Doubt about whether the shocks were real could also change the meaning of pressing a switch: belief and obedience should be examined, not assumed identical for everyone.

A balanced judgement is that the variations demonstrate substantial situational sensitivity and forms of dissent. They do not show that all people obey blindly, that personality never matters, or that one laboratory result completely explains historical atrocities. Their most useful contribution is a set of testable questions about proximity, legitimacy, responsibility and the conditions that allow people to resist.