Stanley Milgram wanted to understand how ordinary people could carry out atrocities under orders, a question made urgent by the Holocaust. His 1963 study became one of the most famous — and most controversial — experiments in psychology, and it is core AQA Social Influence content.
Forty American men volunteered for what they were told was a study on memory and learning, run at Yale. Each became a “teacher” who delivered apparent electric shocks to a “learner” (a confederate, Mr Wallace) for every wrong answer, rising from 15V to a clearly labelled 450V. An experimenter in a grey lab coat gave standardised prods such as “the experiment requires that you continue” when the teacher hesitated.
Every participant went to at least 300V, and 65% continued to the maximum 450V. Many showed extreme distress — sweating, trembling, nervous laughter — yet still obeyed. Milgram concluded that ordinary people will obey a legitimate authority figure even when ordered to harm another person.
Later variations showed obedience is shaped by the situation. Proximity mattered: when teacher and learner were in the same room obedience fell to 40%, and to 30% when the teacher had to force the learner's hand onto a shock plate. Location mattered: in a run-down office (rather than prestigious Yale) obedience dropped to 47.5%. Uniform mattered: when the experimenter was replaced by an “ordinary member of the public” in everyday clothes, obedience fell to 20%.
Two explanations link to this work. The agentic state is when a person acts as an instrument carrying out another's wishes, shifting responsibility onto the authority. Legitimacy of authority is the idea that we obey people we perceive to have justified power, often signalled by visible cues like a uniform or institution.
There is supporting research. Hofling et al. (1966) found 21 of 22 nurses obeyed a doctor's phone order to give an unsafe drug dose, showing obedience to authority in a real setting.
Internal validity has been questioned. Orne & Holland argued participants may not have believed the shocks were real, responding to demand characteristics. However, Milgram's participants showed genuine distress, and Sheridan & King found high obedience even when the shocks (to a puppy) were real.
The study raises serious ethical issues — deception, questionable informed consent, psychological harm and an obscured right to withdraw. In Milgram's defence, participants were debriefed and 84% later said they were glad to have taken part. An alternative reading (Haslam & Reicher) is social identity theory: participants obeyed because they identified with the scientific goal, not through blind obedience.