There is nothing to play here, deliberately. The original procedure worked only
because participants did not know its purpose, and it caused real distress — so this
entry is a description, not a reenactment. What you will find instead: what actually
happened in the Yale laboratory in 1961–62, what changed the numbers, what the
archives later revealed, and what this famous study can and cannot tell you.
What is happening
In the original 1963 study at Yale,
volunteers were told to punish a stranger with stronger and stronger electric shocks
each time he failed a memory test, using a machine labeled from 15 to 450 volts. The
shocks were fake and the "learner" was an actor, but 26 of the 40 men (65%) went all
the way to the final 450-volt switch, even while many of them sweated, trembled, and
protested.Historical / descriptive
Why did they continue? Milgram's own answer — people slip into an "agentic state" and
hand responsibility to the authority — is no longer the only reading.
A newer reading argues participants
were not blindly obeying orders but actively cooperating with what they believed was
important science: people tended to keep going after appeals like "the experiment
requires that you continue," while the one prod that sounds like a direct command —
"you have no other choice, you must go on" — mostly produced refusal. This "engaged
followership" account is influential but remains one interpretation among
several.Mixed / contested
What this shows
How many people obeyed depended heavily on details of
the situation, not on unusual cruelty in the volunteers. In Milgram's own follow-up
conditions, full obedience fell from about two-thirds when the victim was unseen in
another room to 40% with the victim in the same room and 30% when the teacher had to
press the victim's hand onto a shock plate; and when the experimenter left and gave
orders by telephone, fully obedient subjects dropped from 26 out of 40 to 9 out of
40 — with some participants secretly giving only the weakest shock while assuring the
experimenter they were following instructions.Historical / descriptive The study's durable lesson is
situational: small changes in proximity, authority and peer behavior moved obedience
across almost its entire possible range.
What this does not show
The study does not show that about two-thirds of
people will blindly follow destructive orders in everyday life. Obedience ranged from
close to zero (when two experimenters disagreed) to over 90% (when someone else
pressed the switch) across the variations, many participants doubted the shocks were
real, and the once-common leap from this laboratory to real-world atrocities such as
the Holocaust is now widely contested.Supported, with limits
Research and limitations
The closest modern test, run in 2006 and
published in 2009, repeated Milgram's best-known setup but stopped everyone at the
150-volt mark — the first point where the learner demands to be released — for
ethical reasons. 70% of participants were still willing to continue past that point,
compared with 82.5% at the same point in Milgram's version, a difference small enough
to be due to chance; because the study ended there, it can only estimate, not prove,
how many people would go to 450 volts today.Supported, with limits
The archives complicate the textbook story further.
Research in the Yale archives has shown that most
participants left the laboratory without being told the shocks were fake (a written
explanation came much later, and Perry estimates hundreds went home believing they
had shocked a man), the experimenter often improvised beyond his four scripted
prods, and unpublished data show many participants doubted anyone was really being
hurt. In 18 of 23 variations, people who fully believed the shocks were real gave
lower shocks on average than doubters — suggesting skepticism inflated the famous
obedience numbers.Supported, with limits
The experiments triggered one of
psychology's defining ethics debates, starting with Diana Baumrind's 1964 critique of
the stress inflicted on participants, and Milgram's full procedure has been
considered ethically out of bounds for decades — no study using similar procedures
had been published for over thirty years before Burger's modified version. That is
why all modern evidence comes from partial or altered replications rather than a full
repeat.Historical / descriptive
Research drawer — claims, sources and limits
Every highlighted claim above is recorded here with its evidence grade and sources. Grades are per claim; this site never computes an overall truth score.
↩ In the original 1963 study at Yale, volunteers were told to punish a stranger with stronger and stronger electric shocks each time he failed a memory test, using a machine labeled from 15 to 450 volts. The shocks were fake and the "learner" was an actor, but 26 of the 40 men (65%) went all the way to the final 450-volt switch, even while many of them sweated, trembled, and protested.
Milgram, S. (1963). Behavioral study of obedience. Journal of Abnormal and Social Psychology, 67(4), 371-378. https://doi.org/10.1037/h0040525 · accessed 2026-08-15 Support note: Reports the baseline procedure (teacher/learner cover story, shock generator from 15 to 450 volts in 15-volt steps, standardized verbal prods) and the headline result: 26 of 40 male subjects (65%) administered the maximum 450-volt shock despite the learner's protests.
Milgram, S. (1965). Some conditions of obedience and disobedience to authority. Human Relations, 18(1), 57-76. https://doi.org/10.1177/001872676501800105 · accessed 2026-08-15 Support note: Documents participants' extreme tension: subjects were observed to sweat, tremble, stutter and groan; 71 of 160 subjects in the first four conditions showed definite nervous laughter or smiling, with 15 showing full-blown uncontrollable laughing fits — yet many still obeyed fully.
↩ How many people obeyed depended heavily on details of the situation, not on unusual cruelty in the volunteers. In Milgram's own follow-up conditions, full obedience fell from about two-thirds when the victim was unseen in another room to 40% with the victim in the same room and 30% when the teacher had to press the victim's hand onto a shock plate; and when the experimenter left and gave orders by telephone, fully obedient subjects dropped from 26 out of 40 to 9 out of 40 — with some participants secretly giving only the weakest shock while assuring the experimenter they were following instructions.
Milgram, S. (1965). Some conditions of obedience and disobedience to authority. Human Relations, 18(1), 57-76. https://doi.org/10.1177/001872676501800105 · accessed 2026-08-15 Support note: Reports the proximity series with 40 fresh subjects per condition: 34% defied in Remote, 37.5% in Voice Feedback, 60% in Proximity, 70% in Touch-Proximity (i.e., obedience roughly 66%/62.5%/40%/30%); reports 26 fully obedient subjects with the experimenter present versus 9 when orders came by telephone (p<.001), including subjects who covertly used the lowest shock while claiming compliance.
Milgram, S. (1974). Obedience to Authority: An Experimental View. New York: Harper & Row. https://en.wikipedia.org/wiki/Obedience_to_Authority:_An_Experimental_View · citation only Support note: The book compiles all ~18+ numbered variations (including women as subjects at 65%, peer rebellion at 10%, two disagreeing experimenters near 0%, and a subsidiary-act condition at 92.5%), which is the standard citation for the full range of variation results.
↩ The closest modern test, run in 2006 and published in 2009, repeated Milgram's best-known setup but stopped everyone at the 150-volt mark — the first point where the learner demands to be released — for ethical reasons. 70% of participants were still willing to continue past that point, compared with 82.5% at the same point in Milgram's version, a difference small enough to be due to chance; because the study ended there, it can only estimate, not prove, how many people would go to 450 volts today.
Burger, J. M. (2009). Replicating Milgram: Would people still obey today? American Psychologist, 64(1), 1-11. https://doi.org/10.1037/a0010932 · accessed 2026-08-15 Support note: Reports that 70% (28/40) of base-condition participants (men and women, mean age 42.9) continued past the 150-volt switch versus 82.5% (33/40) at the comparable point in Milgram's Experiment 5, a non-significant difference (p=.29); details the ethical safeguards (two-stage clinical screening that excluded about 38% at the second stage, repeated withdrawal reminders, 15-volt sample shock, immediate debriefing) and the rationale that 79% of Milgram's subjects who passed 150 volts went to the end.
↩ A newer reading argues participants were not blindly obeying orders but actively cooperating with what they believed was important science: people tended to keep going after appeals like "the experiment requires that you continue," while the one prod that sounds like a direct command — "you have no other choice, you must go on" — mostly produced refusal. This "engaged followership" account is influential but remains one interpretation among several.
Haslam, S. A., & Reicher, S. D. (2012). Contesting the "nature" of conformity: What Milgram and Zimbardo's studies really show. PLoS Biology, 10(11), e1001426. https://doi.org/10.1371/journal.pbio.1001426 · accessed 2026-08-15 Support note: Argues from the prod evidence (scientific justifications produced compliance while the direct order typically prompted refusal), from variation-by-variation differences in identification with the scientific enterprise, and from post-study participant responses, that Milgram's findings reflect engaged, identification-based followership rather than blind obedience.
Haslam, S. A., Reicher, S. D., Millard, K., & McDonald, R. (2015). "Happy to have been of service": The Yale archive as a window into the engaged followership of participants in Milgram's "obedience" experiments. British Journal of Social Psychology, 54(1), 55-83. https://doi.org/10.1111/bjso.12074 · citation only Support note: Analyzes participants' post-experiment letters in the Yale archive and reports that many expressed happiness at having contributed to science, offered as direct archival evidence for the engaged-followership account.
↩ Research in the Yale archives has shown that most participants left the laboratory without being told the shocks were fake (a written explanation came much later, and Perry estimates hundreds went home believing they had shocked a man), the experimenter often improvised beyond his four scripted prods, and unpublished data show many participants doubted anyone was really being hurt. In 18 of 23 variations, people who fully believed the shocks were real gave lower shocks on average than doubters — suggesting skepticism inflated the famous obedience numbers.
Perry, G., Brannigan, A., Wanner, R. A., & Stam, H. J. (2020). Credibility and incredulity in Milgram's obedience experiments: A reanalysis of an unpublished test. Social Psychology Quarterly, 83(1), 88-106. https://doi.org/10.1177/0190272519861952 · accessed 2026-08-15 Support note: Reanalyzes an unpublished internal study by Milgram's assistant Taketo Murata: in 18 of 23 variations, mean shock levels were lower for subjects who fully believed the learner was receiving painful shocks than for those who did not fully believe; also documents that a written debrief 'had not been provided to the majority of subjects in the laboratory,' that archival work has revealed unstandardized experimental protocols and misreported results, and that in Hollander & Turowetz's tape analysis 72% of obedient participants offered 'the learner wasn't really being harmed' as an explanation.
Perry, G. (2013). Behind the Shock Machine: The Untold Story of the Notorious Milgram Psychology Experiments. New York: The New Press. https://thenewpress.org/books/behind-the-shock-machine/ · citation only Support note: Perry's book-length archival and interview investigation is the origin of the debriefing findings (her estimate that roughly 600 participants left believing the shocks were real, and a participant who learned the truth decades later), of the documentation of the experimenter's departures from script, and of an unpublished 24th 'relationship' variation using pairs of friends and relatives.
↩ The study does not show that about two-thirds of people will blindly follow destructive orders in everyday life. Obedience ranged from close to zero (when two experimenters disagreed) to over 90% (when someone else pressed the switch) across the variations, many participants doubted the shocks were real, and the once-common leap from this laboratory to real-world atrocities such as the Holocaust is now widely contested.
Burger, J. M. (2009). Replicating Milgram: Would people still obey today? American Psychologist, 64(1), 1-11. https://doi.org/10.1037/a0010932 · accessed 2026-08-15 Support note: Summarizes the variation range from Milgram (1974): obedience 'virtually disappeared' when two experimenters gave conflicting instructions, fell to 10% after two peers rebelled (Experiment 17), and rose to 37 of 40 (92.5%) when another person pressed the switches (Experiment 18); Burger also warns that attributing the behavior to obedient personalities rather than situational forces is an example of the fundamental attribution error.
Perry, G., Brannigan, A., Wanner, R. A., & Stam, H. J. (2020). Credibility and incredulity in Milgram's obedience experiments: A reanalysis of an unpublished test. Social Psychology Quarterly, 83(1), 88-106. https://doi.org/10.1177/0190272519861952 · accessed 2026-08-15 Support note: States explicitly that the belief data 'revise our perception of the classical interpretation of the experiment and its putative relevance to the explanation of state atrocities, such as the Holocaust,' and criticizes narratives claiming the experiments show we are 'all capable of torture and murder at the behest of an authority figure.'
↩ The experiments triggered one of psychology's defining ethics debates, starting with Diana Baumrind's 1964 critique of the stress inflicted on participants, and Milgram's full procedure has been considered ethically out of bounds for decades — no study using similar procedures had been published for over thirty years before Burger's modified version. That is why all modern evidence comes from partial or altered replications rather than a full repeat.
Burger, J. M. (2009). Replicating Milgram: Would people still obey today? American Psychologist, 64(1), 1-11. https://doi.org/10.1037/a0010932 · accessed 2026-08-15 Support note: Reviews the ethical controversy (citing Baumrind 1964 among others), notes Milgram's defense from follow-up questionnaires (83.7% of subjects glad to have participated, per Milgram's data), states that current ethical standards 'clearly place Milgram's studies out of bounds' (citing Elms 1995), and that no study using similar procedures had been published in more than three decades (citing Blass 2000).
Baumrind, D. (1964). Some thoughts on ethics of research: After reading Milgram's "Behavioral Study of Obedience." American Psychologist, 19(6), 421-423. https://doi.org/10.1037/h0040128 · citation only Support note: The founding critique of the experiments' ethics, arguing the emotional harm and breach of trust inflicted on participants could not be justified — the standard citation for the start of the ethical debate.