Milgram’s fourth prod: the order people refused most

Stanley Milgram’s obedience experiments. · published 2 September 2026

People can be led a very long way by someone who seems to have legitimate authority, but the famous number describes one arrangement out of two dozen, with a script the experimenter did not keep to — and the plainer the order became, the more people refused.

The angle

The famous study is remembered as proof of blind obedience, but its own transcript points the other way: three prods that framed the shocks as reasonable science kept people going, while the fourth — “you have no other choice, you must go on” — produced more refusals than any other, suggesting people were being persuaded rather than commanded, and that plain authority is weaker than authority dressed as a worthwhile cause.

The notes this episode was written from

These are the working notes, unedited. Everything the video states as fact comes from here, and nothing else may be stated as fact. The sections on limits and on what the finding does not show are part of the brief, not an afterthought — see the method.

Run at Yale between 1961 and 1962. First published as “Behavioral Study of Obedience” (Journal of Abnormal and Social Psychology, 1963), then as the book Obedience to Authority (1974).

The setup: the volunteer takes the role of “teacher” and is told to give an increasingly severe electric shock to a “learner” for every wrong answer, from 15 volts up to 450, the last switches marked XXX. The learner is an actor and the shocks are not real. When the volunteer hesitates, an experimenter in a grey coat presses them to continue with a set of four scripted prods.

The number everyone quotes is 65 per cent: in the best-known condition, 26 of 40 volunteers went all the way to 450 volts.

What that number leaves out

There was not one experiment. Milgram ran around two dozen variations, and obedience moved enormously between them — near total compliance in some, and none at all in others. Moving the learner into the same room, moving the experimenter out of it, or setting the study in an ordinary office rather than at Yale all changed the result substantially. The famous figure is one condition standing in for a range.

Gina Perry’s Behind the Shock Machine (2012) worked from Milgram’s archive at Yale, including the recordings. The experimenter, John Williams, frequently departed from the four scripted prods, improvising and pressing volunteers well beyond what the published method described. Obedience was in part produced by pressure that was never written down.

Many volunteers did not believe the shocks were real, and said so afterwards. The exact proportion is disputed between Milgram’s own post-experiment figures and later readings of the archive, so it should be described as a substantial number and a live disagreement, not given as a percentage.

The most revealing detail is the fourth prod. “You have no other choice, you must go on” is the only one phrased as a direct order, and it produced more refusals than any other — the more it sounded like an order, the more people stopped. Haslam and Reicher have argued from this that what the study shows is not blind obedience but persuasion: people continued while they were convinced they were helping a worthwhile piece of science.

The phenomenon is not fabricated. Jerry Burger’s partial replication (American Psychologist, 2009), stopping at 150 volts for ethical reasons, found rates comparable to Milgram’s.