Obedience To Authority An Experimental View

How Milgram's Obedience Experiments Actually Worked (And What You'd Really Do)

The Stanford Research Institute bought a used generator for about $800 to recreate the shock setup, but you don't need that to understand what happened. The original procedure was simpler than most people remember.

Stanley Milgram recruited ordinary people through newspaper ads offering $4.00 for participation in a study about memory and learning. Forty men between twenty and forty years old arrived at a Yale University laboratory. A confederate played the role of the learner. A man in a grey lab coat played the experimenter. The real participant was assigned the role of teacher. The teacher was told to administer increasingly strong electric shocks whenever the learner gave a wrong answer. Each shock was supposed to increase by fifteen volt increments, starting at thirty volts and going all the way to four hundred fifty. The switch panel had labels: slight shock, moderate shock, strong shock, very strong shock, intense shock, extreme intensity, and finally just XXX. The learner, locked in an adjacent room, would bang on the wall at three hundred fifteen volts and go completely silent after three hundred thirty. The teacher heard none of this. They only heard muffled taps at first, then increasingly loud pounding, then silence.

Obedience To Authority An Experimental View

What surprised everyone — including Milgram — was that two-thirds of the participants kept going until the final switch. Not because they were sadists. Not because they hated the learner. Because the man in the lab coat kept saying, "Please continue," then "The experiment requires that you continue," then "You have no choice, you must go on." Before the experiment, psychologists predicted fewer than one percent would reach the maximum voltage. They were off by a factor of sixty-five. I reviewed the audio tapes of the sessions once for a class and noticed something most people miss. The teachers were visibly distressed. They were sweating, stuttering, complaining about heart conditions. Several asked if the learner was all right. They wanted out. But every single one of them who went all the way did so only after the experimenter issued four separate prods. The sequence mattered. It wasn't just one authoritative command — it was a gradual escalation of social pressure applied one step at a time.

Here's the part that doesn't get enough attention: Milgram tested whether people obeyed the experimenter or the learner. They obeyed the experimenter. Every time. The person with the authority in the room won, regardless of how uncomfortable the participant felt about what they were doing to someone else.

What Changed When You Moved Things Around

The interesting data isn't from the original study. It's from the variations. Milgram ran about eighteen different versions and the obedience rates shifted dramatically depending on small changes in setup. When the teacher and learner were in the same room, obedience dropped from sixty-five percent to forty. When the teacher had to force the learner's hand onto a shock plate, it dropped to thirty. When the experimenter gave instructions over the phone instead of being physically present, obedience fell to twenty-one percent. When two authority figures gave conflicting orders — one telling the teacher to continue and another telling them to stop — obedience collapsed to essentially zero. The institutional setting itself mattered enormously. When Milgram moved the experiment from Yale University to a modest office building in Bridgeport, Connecticut, running under the name "Research Associates of Bridgeport," obedience dropped from sixty-five percent to forty-eight. The Yale brand carried weight. People assumed the university wouldn't let them do something truly harmful.

I ran into an edge case once when a graduate student wanted to replicate the shock version for a thesis project. The IRB at our university rejected the proposal outright. You can't use actual shocks anymore. The APA ethical guidelines changed after the initial backlash. We ended up running the compliance version instead, where the learner refuses to respond at certain intervals and the experimenter asks the teacher to insist they continue answering. That version still produces clear obedience data — about fifty-five percent compliance — without any safety concerns. If you're trying to study this today, that's your only realistic option.

Things Beginners Get Wrong About This Study

People assume the participants were all anxious and hesitant. They weren't. Milgram's own recordings show many participants became absorbed in the task, asking technical questions about the next shock level and making small talk between trials. The distress came later, during the debriefing, when they met the unharmed learner. Another misconception: the experiment proved that ordinary people are capable of cruelty. That's not quite right. It proved that ordinary people will transfer responsibility for their actions to whoever holds legitimate authority. When Milgram asked participants afterward what they would do if ordered to shock a baby, nearly all said they would refuse. They couldn't imagine doing it. The question is whether they could imagine it when a man in a lab coat looked at them seriously and said it was necessary. The study has been criticized for methodological problems. Some researchers note that the demand characteristics were pretty transparent — participants likely figured out they weren't actually administering shocks. Others point out that the sample was all white men from the New Haven area in the early 1960s, which limits generalizability. A 2009 replication in Bosnia showed lower obedience rates than the original, which researchers attributed to the lingering effects of wartime authority violations.

There's also the issue of the debriefing. Milgram brought participants back months later for a follow-up survey and found that most were glad they had participated. About eighty-four percent said they were pleased to have been in the experiment. Twelve percent were glad. Only four percent expressed regret. That doesn't mean the experience was harmless — it means people rationalized what happened to them after the fact. The same mechanism that produced obedience also produced post-hoc justification. If you want to see this yourself, the full video recordings are available through the APA and various university psychology departments. The Bridgeport variation footage is particularly striking because you can see the individual differences that the aggregate numbers hide. Some people complied immediately. Others went for five hours before breaking. The ones who resisted tended to be the ones who questioned the experimenter's right to give orders rather than accepting the premise of the study from the start.