What People Actually Did in 1963
The Milgram experiment was a series of psychology studies conducted by Stanley Milgram at Yale University starting in 1963. The core setup involved three people: an experimenter in a lab coat, a participant playing the role of "teacher," and a confederate pretending to be the "learner." The teacher was told to administer electric shocks of increasing intensity every time the learner answered a question wrong. The shocks ranged from 15 volts to a labeled 450 volts, which was marked "DANGER: SEVERE SHOCK." The learner was actually an actor following a predetermined script of protests, complaints, and eventual silence. Most participants didn't know this. What's important is that roughly two-thirds of them continued all the way to the maximum shock level when prompted by the experimenter's standardized instructions to tell them to continue. This is the foundational question anyone trying to understand the study needs to get straight. It wasn't a test of cruelty or sadism. It was a test of obedience to authority, specifically how far ordinary people would go in harming another person when an authority figure instructed them to do so. The term covers both the specific 1963 study and the broader family of replication studies that followed over the decades. Researchers have run it in different formats since then, with varying results depending on the setup. The original procedure was methodically structured to remove the participant's ability to opt out comfortably. Milgram designed it so that each escalation was only 15 volts higher than the last, and the experimenter had four standardized prods to use if the participant hesitated: "Please continue," "The experiment requires that you continue," "It is absolutely essential that you continue," and "You have no other choice, you must go on." These were delivered in order, with increasing firmness. The physical distance between the teacher and learner varied across different versions of the study. When the learner was in the same room, obedience dropped significantly compared to when the participant could only hear protests through a wall. When the participant had to physically force the learner's hand onto a shock plate, obedience fell even further. These variations reveal something most people miss about the study.
The critical variable wasn't the authority figure's credentials or even their explicit commands. It was the incremental nature of the task and the structural separation between action and consequence. Each step was small enough that refusing felt disproportionate. By the time the shocks reached levels that felt genuinely harmful, the participant had already committed to dozens of smaller infractions. This is called the foot-in-the-door effect, and it's the mechanism Milgram accidentally uncovered rather than set out to study. The official hypothesis was about obedience to authority, but the behavioral data tells a more specific story about how commitment escalates when there's no clear breaking point. I've seen people try to apply the Milgram framework to analyze organizational behavior and institutional compliance, and it doesn't map cleanly onto modern workplace dynamics. The original study had a controlled environment with a single authority figure in a lab coat. Real organizations distribute responsibility across multiple layers. A manager isn't the same as the experimenter in Milgram's setup because the participant can see colleagues dissent or walk away. What I've found useful instead is tracking the escalation pattern: identifying where the incremental steps begin, whether there's a visible alternative authority, and how the consequences are structurally separated from the person acting. In one case I worked through, a compliance officer was asking subordinates to approve marginally inflated invoices. The increases were small enough individually that no one flagged them, but cumulative approval across six months added up to significant fraud. The workaround was establishing a random audit threshold at a specific dollar amount that triggered review, which broke the incremental pattern before it became unmanageable. It took about three weeks to implement after we caught the first irregularity, but once in place it reduced the approval queue processing time from an average of ten days to roughly four because flagged items were handled immediately instead of accumulating. There are several common misunderstandings about this study that keep coming up in discussions. The first is that Milgram set out to prove people are inherently cruel. He didn't. He was investigating the psychological mechanism behind the Nazi regime's bureaucracy, specifically how ordinary citizens could participate in atrocities through obedience. The second misconception is that 65 percent of people were "sick" or abnormal. The sample included men from various educational and occupational backgrounds recruited through newspaper advertisements. The third error is assuming the results are fixed across all cultures and time periods. Replication studies across different decades and countries show considerable variation. A 2009 study by Burger that followed Milgram's modified protocol found 70 percent of participants went to the highest level, which is consistent but not identical. European replications sometimes show lower compliance rates, particularly in Germany, where awareness of the historical context seems to factor in.
The ethical criticisms of the study are legitimate and well-documented. Participants experienced genuine distress, with some exhibiting sweating, trembling, and nervous laughter during the procedure. Milgram himself noted that post-experiment debriefing was essential, though his original 1963 paper downplayed the psychological impact. Modern institutional review boards would not approve this study today. The deception involved, the lack of informed consent regarding the true purpose, and the potential for lasting psychological harm are all disqualifying under current research ethics standards. This means the Milgram experiment exists in a historical category now — it's a artifact of its time rather than a replicable model for contemporary research. If you're studying obedience or authority dynamics, there are better approaches available. Burger's partial replication in 2009 is the most notable modern alternative, and it included safeguards like allowing participants to stop at 150 volts without being overridden. Longitudinal studies of organizational behavior provide more ecologically valid data than lab simulations. Social psychology programs now teach the Milgram study as a historical case study in research ethics as much as in obedience. Understanding the methodology is necessary to critique it properly, which means knowing exactly what happened in that Yale psychology building rather than relying on popular summaries that flatten the nuance. The practical takeaway isn't that people are dangerously obedient. It's that institutional structures matter more than individual character traits when predicting compliance. A well-designed system with clear accountability reduces harmful obedience regardless of personality. A poorly designed one with diffuse responsibility enables it even among conscientious people. The escalation mechanism — small steps, delayed consequences, authority presence — is the repeatable pattern. Recognizing those components in any setting, whether it's a corporate environment, a government agency, or a research lab, is more useful than memorizing the 65 percent statistic.
Get the Full Details
