The ethics fight that changed research forever 

The objection came fast. In 1964, the year after Milgram's first paper, the developmental psychologist Diana Baumrind published a critique in American Psychologist. Her argument was not that the findings were uninteresting. It was that Milgram's own descriptions of his participants, sweating, trembling, laughing uncontrollably, one seized by a convulsive seizure, documented serious distress that he reported as data rather than treated as harm. She questioned whether the debriefing could undo the experience of having discovered something intolerable about yourself, and whether the knowledge gained justified the cost.

Milgram replied the same year and did not concede much. He described his dehoaxing procedure, in which every participant was reconciled with the unharmed learner and given an explanation of the deception, and pointed to follow-up questionnaires in which a large majority said they were glad to have taken part and very few reported regret. He also had a psychiatrist interview a sample of participants a year later and reported no evidence of lasting injury. His deeper claim was that the discomfort came from what people learned about themselves, not from anything done to them, and that no researcher had a right to suppress an unwelcome result.

Both positions have aged unevenly. Milgram's follow-up data are real and are also exactly the sort of self-report a satisfied institution collects about itself. Archival work by later scholars, Gina Perry among them, indicates that debriefing was less thorough and less prompt for many participants than the published account suggests, with some not learning for months that the shocks had been fake. Set against that, participants were volunteers who could leave, and the study produced knowledge that has genuinely shaped how institutions think about complicity.

The consequences were structural rather than personal. Milgram was briefly suspended from the American Psychological Association pending review and then admitted to membership, and his book received an award. Across the following decade, the obedience studies became the standard case in every discussion of deception, informed consent, the right to withdraw and mandatory debriefing. Together with the medical scandals of the same period, they helped produce the institutional review boards and ethics committees that now examine research protocols before a single participant is recruited. A modern committee would not approve the original design, which is why Burger's 2009 replication stopped at 150 volts.

That is the honest summary of the whole file. A study that could not be run today, whose numbers should be quoted with care and whose theory has been reasonably contested, still told us something true about how situations work on people, and the argument it started did more good than the result itself.

The Obedience Files ends there, in a committee room, which is a stranger and better ending than a laboratory.