The Obedience Files 

Stanley Milgram was 27 and newly hired at Yale when he built a machine that did not work. The shock generator had thirty switches labelled from 15 to 450 volts, with descriptions running from Slight Shock to a final XXX. It was wired to nothing.

Volunteers answered a newspaper ad offering four dollars and fifty cents for an hour of research on memory and learning. Each arrived, drew a rigged lot that made him the teacher, and watched a mild middle-aged man called Mr Wallace be strapped into a chair in the next room. Wallace was an actor. So, in effect, was the experimenter in the grey technician's coat, whose entire script consisted of four escalating prompts, ending with: You have no other choice, you must go on.

In the best known version, 26 of 40 men went to the maximum switch. That number gets quoted as if it were a fixed property of human nature, and it is not. It is the result of one configuration of a room. Milgram ran more than twenty variations, and the obedience rate moves enormously with distance, authority and the presence of anyone who refuses.

The findings arrived a year after the Eichmann trial and were read straight into that argument, which is partly why they landed so hard. They also triggered a decade of ethical objection about deception and distress that helped produce the institutional review boards every psychology department now runs on. The collection treats all three layers: what happened, what it means, and what it cost.