Lesson 6 of 8 · 19 min
Milgram (1963): obedience
Describe the aim, procedure, findings and conclusions of Milgram's (1963) obedience study, explain why people obeyed, and evaluate its ethics and its effect on ethical guidelines for researchers.
What you will be able to do
- Describe the aim, participants and procedure of Milgram (1963).
- Describe the findings, including the break-off points and signs of tension, and the conclusion.
- Explain obedience with the agentic state, legitimate authority and gradual escalation, and how Milgram's later variations changed it.
- Evaluate the study's ethics, and describe its effect on ethical guidelines for psychological researchers.
Stanley Milgram began his experiments at Yale University in 1961, the year Adolf Eichmann, one of the organisers of the Holocaust, was tried in Jerusalem and defended himself by saying he had followed orders. Milgram opened his 1963 paper with the Holocaust: millions were killed on command, and that could only happen "if a very large number of persons obeyed orders". He wanted to measure, in a laboratory, how far ordinary people would go when an authority told them to hurt someone.
Key idea
Milgram (1963) at a glance
Aim: to investigate how far people would obey an authority's instructions to give increasingly painful electric shocks to another person.
Participants: 40 men aged 20 to 50 from New Haven and nearby towns, recruited by a newspaper advertisement and direct mail for "a study of memory and learning" (a self-selected sample). Their jobs ranged from postal clerks and labourers to teachers, salesmen and engineers. Each was paid $4.50 and told the money was theirs whatever happened.
Measure: the highest of 30 shock levels each participant gave before refusing to go on. Milgram called those who went to the last switch obedient and those who stopped earlier defiant.
Procedure. At Yale, each participant met the experimenter (a 31-year-old biology teacher in a grey technician's coat, stern and impassive) and another "participant", a mild-mannered 47-year-old accountant who was really a confederate. A rigged draw (both slips said "Teacher") always made the real participant the teacher and the confederate the learner.
The learner was strapped into a chair in the next room with an electrode on his wrist. The teacher was given a real 45-volt sample shock, then sat at a shock generator with 30 switches from 15 to 450 volts in 15-volt steps, labelled in groups from Slight Shock to Danger: Severe Shock, with the last two marked XXX. None of the shocks to the learner were real.
The teacher read word pairs to the learner, then tested him. For every wrong answer the teacher had to give a shock, one level higher each time, announcing the voltage aloud. The learner gave about three wrong answers to every right one. At 300 volts he pounded on the wall and his answers stopped appearing; the experimenter said to treat no answer as a wrong answer. He pounded again at 315 volts, and after that there was silence.
The experimenter's prods, used in order whenever the teacher hesitated
"Please continue," or "Please go on."
"The experiment requires that you continue."
"It is absolutely essential that you continue."
"You have no other choice, you must go on."
If the teacher still refused after the fourth prod, the session ended.
Move the slider along the 30 switches and watch how many of the 40 participants had stopped by each one.
Predictions. Milgram described the procedure to 14 Yale psychology seniors and asked what 100 Americans would do. They predicted that between 0 and 3% would go to 450 volts, a mean of 1.2%. Colleagues guessed few would go past Very Strong Shock.
Findings.
- Every participant went to at least 300 volts. No one stopped earlier.
- 5 stopped at 300 V, 4 at 315 V, 2 at 330 V, and 1 each at 345, 360 and 375 V: 14 defiant participants.
- 26 of the 40 (65%) went all the way to 450 volts, two steps past Danger: Severe Shock.
- Participants believed the situation was real: asked how painful the last shocks were, on a 14-point scale, their mean answer was 13.42.
- Many showed extreme tension: sweating, trembling, stuttering, biting their lips, digging their nails into their flesh. 14 showed nervous laughter, and 3 had full-blown, uncontrollable seizures; one was so violent the session was stopped.
| Highest shock given (V) | Number of participants |
|---|---|
| 300 | 5 |
| 315 | 4 |
| 330 | 2 |
| 345 | 1 |
| 360 | 1 |
| 375 | 1 |
| 450 (the maximum) | 26 |
| Total | 40 |
Using the table, calculate the percentage of participants who stopped at 300 V or 315 V.
Worked example
Analyse Milgram's break-off data
The table shows the highest shock given by each of Milgram's (1963) 40 participants.
| Highest shock (V) | 300 | 315 | 330 | 345 | 360 | 375 | 450 |
|---|---|---|---|---|---|---|---|
| Participants | 5 | 4 | 2 | 1 | 1 | 1 | 26 |
(a) Calculate the percentage of participants who went to 450 V. (b) Calculate the mean highest shock. (c) Identify the median, and explain why it is a better summary than the mean here. (d) Compare the result with the Yale seniors' prediction of 1.2%.
Step 1: (a) Percentage obedient
.
Step 2: (b) Mean
Total of the highest shocks: . Mean V.
Step 3: (c) Median
With 40 scores the median is the mean of the 20th and 21st in order. The first 14 are below 450 V, so the 20th and 21st are both 450 V: median V. The data are skewed (most at the maximum, a tail of lower values), and the mean of 405 V is a voltage few people actually stopped at, so the median describes the typical participant better.
Step 4: (d) Compare with the prediction
The seniors predicted 1.2% would go to 450 V; 65% did, more than fifty times as many. People greatly underestimate the power of the situation.
.
Try it yourself first, then check the working one step at a time.
Answer
(a) 65%. (b) 405 V. (c) Median 450 V; the scores are skewed towards the maximum, so the median is more typical. (d) 65% against a predicted 1.2%: obedience was more than fifty times the prediction.
Method
Percentage: count ÷ total × 100.
Mean from a frequency table: multiply each value by its frequency, add, divide by the total.
Median: the middle position in order; for n = 40, the 20th and 21st.
Watch out for
Averaging the seven voltages in the table without weighting by the number of participants: that gives 353.6 V.
Saying 65% "obeyed" without saying to what: 65% went to the maximum shock.
Conclusion. Milgram concluded that the tendency to obey a legitimate authority is far stronger than people predict: ordinary people will act against their own values and harm an innocent person when instructed, even by an authority "who has no special powers to enforce his commands". The behaviour came from the situation, not from cruel personalities: the participants were distressed by what they were doing, yet most obeyed.
Why did they obey? Milgram and later researchers point to:
- Legitimate authority. The experimenter, the laboratory coat and Yale University signalled a right to give instructions.
- The agentic state. Participants saw themselves as carrying out the experimenter's wishes, so the responsibility felt like his. Some asked who would be responsible if the learner was hurt, and went on once told it would be the experimenter.
- Gradual escalation. Each shock was only 15 V more than the last, so there was no obvious point at which to stop.
- Distance from the victim. The learner was out of sight and, after 315 V, silent.
Note
Milgram's variations
Milgram went on to run many variations, reported in Obedience to Authority (1974), each with a new group of participants. Obedience fell when the learner was in the same room or the teacher had to touch him, when the experimenter gave orders by telephone, when the orders came from an "ordinary person" instead of the experimenter, and most of all when two other "teachers" (confederates) refused to continue. It fell somewhat when the study moved from Yale to a run-down office building. The variations are not part of the 1963 paper the syllabus names, but they help explain its result.
Each bar is a separate group of participants. Which features of the situation made obedience fall?
One participant said afterwards, "I wouldn't have done it on my own. He was in charge, so it was on him." Which explanation of obedience does this best show?
Ethics. The study caused a storm. Diana Baumrind (1964) argued that it harmed participants and broke the trust between researcher and participant. Measured against the ethical principles you use in your own practicals:
- Protection from harm: participants suffered extreme stress, including seizures, and some left knowing they had been willing to hurt someone.
- Deception and informed consent: they were told the study was about memory and learning, that the shocks were real and that the draw was fair, so they could not give informed consent.
- Withdrawal rights: they were told they were paid whatever happened, but the prods ("You have no other choice, you must go on") told them they could not leave.
- Debriefing: Milgram did dehoax every participant: each met the unharmed learner in "a friendly reconciliation", and Milgram later reported that most were glad to have taken part. Critics replied that a debriefing cannot undo what a person has learnt about themselves.
For each step of the procedure as Milgram ran it, decide whether it is acceptable or which principle it breaches.
History
How Milgram changed the rules
Milgram's study, together with medical scandals of the same era, pushed psychology to write down and enforce ethical rules. The American Psychological Association published research principles in 1973, and committees that must approve research before it starts became standard. Today in Australia, every study with human participants must be approved by a human research ethics committee under the National Statement on Ethical Conduct in Human Research. It requires informed consent, a real right to withdraw at any time without penalty, deception only when there is no alternative and the risk is low, and debriefing.
Milgram's exact procedure could not be approved today. When Burger (2009) repeated it in the United States, he screened out people at risk, told participants three times they could leave, and stopped every session at 150 V, because in Milgram's later variation with the learner's protests most people who went past 150 V continued to the end. Most of Burger's participants were still willing to go past 150 V, at a rate not significantly different from Milgram's, suggesting obedience had changed little in nearly 50 years.
Explain how Milgram's (1963) study breached the ethical principle of withdrawal rights, and how a researcher would protect it today. (2 marks)
Exam tip
Evaluating the study
Balance the critique. Strengths: a standardised procedure (fixed prods, a scripted learner) made it replicable; a varied community sample, not only students; and the findings are important, which Milgram used to justify the cost. Limitations: the ethics above; a laboratory with an unusual task (ecological validity is debated); an all-male American sample in 1961–62; and Australian psychologist Gina Perry's study of the archives (2012) found the experimenter often went beyond the four standard prods and that many participants later said they doubted the shocks were real.
Summary
Aim: how far people would obey an authority's instructions to shock another person.
Procedure: 40 men aged 20–50; rigged draw; teacher and learner; 30 switches, 15–450 V; learner pounds at 300 V and 315 V, then silence; four prods.
Findings: all 40 reached 300 V; 26 of 40 (65%) went to 450 V, against a predicted 1.2%; extreme tension, with 14 laughing nervously and 3 having seizures.
Explanations: legitimate authority, the agentic state, gradual escalation, distance from the victim; obedience fell in variations that changed these.
Ethics: harm, deception, no informed consent and undermined withdrawal rights; participants were dehoaxed.
Effect on guidelines: helped bring in enforced codes and ethics committees; in Australia every study needs approval under the National Statement.
Finished the lesson?
Try the 3 quick checks above and the lesson completes itself, or mark it complete now.