Milgram revisited: the obedience study at sixty

In 1961, in a basement lab at Yale, Stanley Milgram instructed ordinary American men to deliver electric shocks of increasing intensity to a stranger in another room as part of what they believed was a memory experiment. The stranger — an actor — was not actually being shocked, but screamed convincingly. Sixty-five percent of subjects continued past the "danger: severe shock" marker, delivering what they believed was 450 volts (Milgram, 1963).

The result has been the most-cited finding in twentieth-century social psychology. It has also been steadily reinterpreted as the underlying records have been re-examined.

1. What Milgram actually found

The headline 65% figure refers to one experimental condition — voice-feedback only, with the experimenter physically present, in a Yale laboratory. Across the eighteen experimental variations Milgram ran, full obedience varied from 0% to 92% depending on conditions. The headline was the highest-compliance variant, not the average.

The condition variations are themselves the most informative findings:

• When the experimenter was physically absent and gave instructions by phone, compliance dropped to about 21%
• When the learner was in the same room, compliance dropped to about 40%
• When the experimenter wore plain clothes rather than a lab coat, compliance dropped further

The original study, properly read, is about contextual obedience — not a fixed feature of human nature.

2. The Australian re-analysis

In the 2010s, Australian researchers led by Gina Perry obtained access to Milgram's original archive and published Behind the Shock Machine (2013). Her review of the unpublished interviews, transcripts, and methodological notes complicated the picture substantially:

• Many subjects suspected the shocks weren't real and continued partly on that basis
• Several subjects who "obeyed" did so while expressing escalating distress that the published data doesn't fully capture
• The experimenter's prods varied across sessions in ways that influenced compliance more than Milgram acknowledged

This doesn't refute the findings. It does mean the "ordinary people will torture strangers if told to" framing oversimplifies what the records actually show.

3. The 2017 partial replication

Burger and colleagues' modern partial replication (2009-2017) ran the paradigm with modern ethical constraints — stopping at 150 volts instead of 450. They reported obedience rates within a few percentage points of Milgram's original at the same intensity (Burger, 2009). The basic phenomenon replicates. The interpretation continues to be debated.

4. The reinterpretation

The most sophisticated current reading, advanced by Alex Haslam and Stephen Reicher: Milgram's subjects weren't simply obeying authority. They were engaging with a cause the experimenter framed as scientifically important. Subjects who saw themselves as collaborators in a worthwhile project complied; subjects who saw themselves as victims of a coercive authority resisted.

This shifts the moral implication. The dangerous thing isn't blind obedience; it's identification with a frame that legitimizes harm. The frame is something humans can examine and reject. Mere obedience is harder to fight than identification with the wrong cause.

5. The honest summary

Milgram's experiments showed that ordinary people, in carefully constructed authority-laden contexts, will go further than they expected. The strong popular framing — that everyone is a potential torturer — overshoots. The careful framing — that institutional context can produce harm even from unwilling participants, and that which institutional cause people identify with matters more than the abstract authority — fits the evidence better.

The experiments remain consequential. The interpretation has matured.