Audio tapes reveal mass rule-breaking in Milgram's obedience experiments

Reassessing one of the most famous psychological experiments in history, a recent analysis of audio recordings reveals that subjects who seemingly obeyed orders to administer severe electric shocks actually broke the rules of the scientific study most of the time. The authors suggest that this routine violation of experimental procedures transformed the laboratory into a scene of unauthorized violence, altering our understanding of compliance and coercion. The research was published in the journal Political Psychology.

In the early 1960s, American social psychologist Stanley Milgram conducted a series of experiments to understand how ordinary people could be directed to commit violent acts. Volunteers were recruited for what they were told was a study on memory and learning at Yale University. Upon arriving, they were assigned the role of a teacher and introduced to a learner, who was actually an actor working for the researchers.

The teacher was instructed to read a list of word pairs to the learner and test their memory. For every incorrect answer, the teacher had to administer an electric shock, increasing the voltage steadily up to a supposedly lethal level. As the shocks grew stronger, the learner would begin to grunt, shout protests, and eventually scream in simulated agony.

For decades, psychologists have generally accepted that the participants who went all the way to the maximum voltage did so because they believed in the scientific validity of the enterprise. The assumption has been that the presence of a lab-coated authority figure gave the violent actions a sense of legitimacy. Theoretical explanations for the high rates of obedience rely heavily on the idea that the volunteers willingly participated in a structured, orderly scientific protocol.

Lead author David Kaposi, a researcher at The Open University in the United Kingdom, and his colleague David Sumeghy wanted to test whether this assumption matched the reality of the sessions. Kaposi and Sumeghy questioned whether the obedient participants actually followed the specific instructions that made up the memory test cover story. If the volunteers ignored the scientific procedures, the theoretical justification for their violence would be drawn into question.

To investigate this, the researchers turned to the original audio tapes preserved at the Yale University Library. They secured recordings from four experimental conditions that closely resembled the standard baseline setup. After excluding sessions with missing data or technical irregularities, the sample yielded 136 full audio recordings of individual sessions.

The research team broke down the participants into two populations based on their ultimate actions in the lab. Obedient individuals were those who administered all the shocks up to the theoretical maximum. Disobedient individuals were those who refused to continue at some point and formally ended their participation.

Kaposi and Sumeghy then evaluated how well each participant adhered to the explicit rules of the memory and learning study. According to the original directions, the teacher was required to complete a strict five-step sequence for every single shock. This cycle involved reading a test question, evaluating the learner’s answer, announcing the shock voltage, pressing the shock lever, and finally reading the correct answer aloud.

A failure to complete any portion of this sequence was labeled as a procedural violation. The researchers categorized these deviations into two types based on how they occurred. An omission took place when a subject completely skipped a step, such as forgetting to announce the voltage level or failing to read the correct answer.

A commission occurred when the subject technically performed a step but did so in a way that ruined the premise of the memory study. For example, some participants read the next test question aloud while the learner was actively screaming in protest. Under those conditions, it would be impossible for the learner to hear the question, rendering the educational facade meaningless.

The audio analysis revealed a striking pattern of rule-breaking across the board. Out of all the participants traditionally classified as completely obedient, not a single one actually followed the full five-step procedure from start to finish. While every obedient participant reliably pressed the shock lever, they regularly neglected or ruined the other steps required to justify the shock.

In fact, nearly half of the shock sequences administered by obedient participants contained one or more procedural violations. On average, these individuals violated the experimental rules in 48.4 percent of their actions. The act of pressing the shock lever was consistently completed, but the scientific framework surrounding it was continuously broken.

The researchers noted that the sessions generally unfolded in three distinct phases. In the early stages, where the learner remained relatively quiet, procedural violations were low across both groups. As the session progressed and the learner’s recorded protests intensified, the rate of rule-breaking spiked dramatically.

During the final phase of the maximum-voltage sessions, the learner’s protests ceased completely. Despite this silence, the surviving subjects did not return to following the proper scientific procedures. Their violation rates remained consistently high until the very last shock was delivered.

The researchers compared these overall rates with the rule-breaking behavior of the disobedient subjects. To ensure a fair comparison, they only analyzed the portion of the disobedient sessions where those participants were actively complying and administering shocks. The data showed that disobedient subjects committed significantly fewer procedural violations during their compliant phases.

On average, eventual quitters violated the experimental protocol in 30.6 percent of their active sequences. Unlike the fully compliant group, several of the disobedient individuals followed the procedure flawlessly right up until the point they refused to continue. The statistical difference highlights that the people who eventually quit were actually better at following the scientific protocol than those who went to the end.

The most frequent violation in obedient sessions involved reading the memory test questions over the simulated screams of the learner. Doing this effectively guaranteed that the learner would fail the test and receive another shock. By talking over the protests, the obedient subjects abandoned the goal of testing memory and simply facilitated continuous shocks.

Kaposi and Sumeghy interpret these patterns as a complete breakdown of the supposedly legitimate scientific environment. The subjects were not committing violence for the sake of an orderly memory study. With the scientific elements either forgotten or rushed, the laboratory changed into a setting for unauthorized and senseless violence.

The study authors propose that the experimenter played a major, passive role in establishing this dynamic. When the participants broke the rules and skipped steps, the authority figure rarely intervened to correct them or pause the session. By staying silent and letting the memory study fall apart, the experimenter allowed an atmosphere of illegitimate violence to flourish.

This silent approval of the deteriorating conditions may have functioned as a form of coercive control. Participants found themselves trapped in a situation where the stated rules no longer applied, yet the expectation to deliver shocks remained constant. The authors suggest that this environment compromised the volunteers’ freedom to choose, bringing into doubt the idea that they were acting out of willing obedience.