The famous study does not say what you think
The Milgram obedience experiments are among the most cited — and most misrepresented — studies in psychology. Later examination of the archives showed the experimenter improvised and pressured participants well beyond the script, results varied enormously across conditions, and many participants expressed doubt the shocks were real. The tidy lesson — "ordinary people blindly obey" — is not what the data actually supports.
What is usually claimed
That around 65% of ordinary people would deliver a lethal electric shock to a stranger simply because a man in a lab coat told them to — proving we are all latent monsters, one instruction away from atrocity.
What the archives actually show
- The 65% figure is one condition of many. Obedience varied dramatically across variations — in some it collapsed almost entirely. Reporting a single headline number is a choice, not a finding.
- The experimenter went off-script. Recordings show prods far beyond the four standardised lines — coaxing, improvising, pressuring. That is not "obedience to authority"; it is a negotiation.
- Many participants doubted the shocks. A substantial share expressed suspicion that it was not real — and those who believed it was real were less likely to comply fully.
- Participants were not blank obeyers. They argued, resisted, checked, and agonised. The recordings are full of people struggling, not marching.
A leading modern reading is "engaged followership": people did not obey blindly — they were persuaded to identify with the scientific project and came to see the harm as a regrettable cost of something worthwhile.
That is a more disturbing conclusion, not a comforting one. It means the danger is not a person who stops thinking. It is a person who keeps thinking — and has been given a cause.
What authority genuinely does
Setting Milgram aside, the ordinary influence of authority is well attested: we defer to symbols of expertise — titles, uniforms, credentials, confident delivery — often without checking whether the expertise is real or relevant. The doctor selling supplements, the "as seen on TV" badge, the actor in a white coat: all borrow the signal without the substance.
Two questions, every time: "Is this person actually an expert in this specific thing?" and "What do they gain if I comply?" A Nobel laureate in physics has no special authority on nutrition — and expertise plus an incentive is not expertise.
Keep going
More from this wing: all 8 pages · related: manipulation tactics, the bias explorer, friction.
Frequently asked
Is the Milgram experiment valid?
Do people blindly obey authority?
How do you resist authority-based persuasion?
Sources & further reading
- Meta-Milgram: an empirical synthesis of the obedience experiments (note: 65% is one condition of many). PLOS ONE / PMC ↗
- Perry & colleagues, archival reanalysis: the experimenter improvised beyond the four prods, and many participants doubted the shocks were real. Perry et al., 2020 ↗
- Haslam, Reicher & colleagues, the Yale archive and “engaged followership” interpretation. Textbook-coverage review, author copy ↗
Educational content, not professional advice. Last reviewed July 2026.