150 volts: the moment of choice 

Read the transcript of a session and the shape of the thing is not a slow slide. It is a single fork in the road that arrives about a third of the way along the panel.

Up to about 120 volts the learner's scripted reactions are grunts and complaints of discomfort. At 150 volts the script changes in kind rather than degree. The learner says the shocks are becoming painful, refuses to go on, and asks to be released, invoking his heart. This is the first point at which continuing means overriding an explicit withdrawal of consent by the person being harmed, and it is the point at which the largest single group of defiant participants broke off.

That distribution matters for how the study is quoted. The famous figure from the best known condition is that 26 of 40 men proceeded to the final 450 volt switch, roughly sixty-five percent. Stated bluntly it sounds as though people are indifferent to suffering. Look at where the refusals cluster and a different picture appears. Almost everyone who was going to defy the experimenter did so at or shortly after the moment the learner asked to stop, and almost nobody who continued past that point stopped later. Once a participant had talked himself past the withdrawal of consent, the remaining twenty switches were, behaviourally, one decision already made.

Two consequences follow. The first is analytical. If the psychologically decisive event happens at 150 volts, then the interesting variable is not tolerance for cruelty but the way a person resolves a conflict of authority and conscience at first contact. Jerry Burger's partial replication, published in 2009, took exactly this view and stopped sessions at 150 volts on ethical grounds, arguing that the point of divergence had already been passed. His compliance rates at that threshold were broadly comparable to Milgram's, which is modest but real evidence that the effect was not a peculiarity of 1961.

The second is practical, and this is where the study earns its place outside the laboratory. Participants who resisted did not usually deliver a speech. They asked questions, checked who was responsible, requested to see the learner, and physically turned away from the panel. Refusal was made of small procedural moves rather than moral courage in a single stroke. Whatever we take from Milgram, that is the transferable part.

What kept the others going was not force, because the experimenter never once threatened anybody. He had four sentences, and that was all.