Anyone who wants to run a psychology experiment today first has to get approval from an ethics committee. Participants must volunteer, know what they are signing up for, be able to stop at any time and be told afterwards what the study was really about.
In the middle of the 20th century, none of that existed. Researchers ran experiments that still appear in every introductory textbook, but that no ethics board would approve today. And for several of them, it has since turned out that the textbook version is not quite what happened.
10. The Stanford Prison Experiment
In 1971, Philip Zimbardo built a mock prison in the basement of Stanford’s psychology building. Two dozen students were randomly assigned to be guards or prisoners. Within days, the guards were humiliating the prisoners, and the experiment was stopped after six days instead of the planned two weeks.
That is the textbook story. But in 2018, researchers who examined the original recordings found that the guards had been encouraged to be tough, and one prisoner whose breakdown became famous later said he had been acting so he could get out. What remains is less a proof that roles take over our behavior, and more a demonstration of how far people go when an authority makes clear what is expected of them.
9. Blue eyes, brown eyes
The day after Martin Luther King was murdered in April 1968, teacher Jane Elliott divided her class of eight-year-olds in Riceville, Iowa, by eye color. Blue-eyed children were told they were smarter and got extra privileges. Brown-eyed children had to wear collars.
Within a day, the “superior” children became arrogant and cruel, and the others’ schoolwork suffered. The next day, she reversed the roles, with the same result. It was not a scientific experiment, and parents were not asked for permission, but it became a famous lesson about discrimination that is still used in training today.
8. The Monster Study
In 1939, speech researcher Wendell Johnson at the University of Iowa had a graduate student test his theory that stuttering is caused by labeling. Twenty-two orphans took part. Some were praised for their speech, while others, who spoke normally, were constantly told they were stuttering and should not speak until they could do it right.
They did not become stutterers, but several children in the negative group became anxious and withdrawn, and some suffered for years. Colleagues later nicknamed it the “Monster Study”. In 2007, the state of Iowa paid 925,000 dollars to former participants.
7. The Robbers Cave Experiment
In 1954, Muzafer Sherif took 22 eleven-year-old boys to a summer camp in Oklahoma, divided into two groups that did not know about each other. As soon as the researchers organized competitions, hostility broke out within days: name-calling, flag burning and raids on each other’s cabins.
Then Sherif created problems that could only be solved together, such as a broken water supply. Afterwards, the boys wanted to ride home on the same bus. The catch: a year earlier, a similar attempt had failed, because the boys refused to hate each other and turned against the researchers instead. That earlier study was never published.
6. Learned helplessness
In 1967, Martin Seligman and Steven Maier gave dogs electric shocks they could not escape. Later, the same dogs were placed in a box where a simple step over a low barrier would end the shocks. Many did not even try, while dogs that had not been through the first phase quickly jumped away.
“Learned helplessness” became one of the most influential ideas in psychology and a model for depression. But in 2016, Maier and Seligman revised their own theory: passivity turns out to be the brain’s default response to prolonged stress, and what animals learn is actually a sense of control.
5. Harlow’s monkeys
In the 1950s, many psychologists believed that a baby loves its mother mainly as a source of food, and that too much cuddling was bad for children. Harry Harlow proved them wrong with experiments that would be unthinkable today.
He separated baby rhesus monkeys from their mothers and gave them two substitutes: a bare wire “mother” with a bottle of milk and a soft cloth “mother” without food. The babies drank from the wire one but spent most of the day clinging to the cloth one. Later, Harlow isolated young monkeys for months, and they never fully recovered. His work changed how children were treated in hospitals and orphanages, but also helped spark the animal rights movement.
4. The Milgram experiment
In 1961, Stanley Milgram wanted to understand how ordinary people could take part in the Holocaust. Participants were told to give increasingly strong electric shocks to a “learner” in the next room whenever he made a mistake. The learner was an actor, and the shocks were fake.
In the best-known version, 65 percent of participants went all the way to the maximum of 450 volts. But Milgram ran more than 20 versions, and obedience was often much lower. Later research also showed that the experimenter often went off script to pressure people, and some participants suspected the shocks were not real. What stands is that almost no one simply got up and walked out.
3. The bystander effect
In 1968, John Darley and Bibb Latané had participants fill in a questionnaire while smoke slowly filled the room. Those who were alone usually reported it quickly. When other people in the room, secretly actors, ignored it, most participants just kept writing.
The research was inspired by the 1964 murder of Kitty Genovese, after The New York Times reported that 38 neighbors watched and did nothing. That story turned out to be largely false: several people called the police, and one neighbor held her as she was dying. And a 2019 study of real CCTV footage found that in nine out of ten conflicts, at least one bystander did intervene.
2. The Asch conformity experiment
Around 1951, Solomon Asch created a task nobody could get wrong: which of three lines is the same length as a fourth? Alone, almost everyone answered correctly.
But when seated with a group of actors who all gave the same obviously wrong answer, about three-quarters of participants went along with the group at least once. Interestingly, when just one actor gave the right answer, conformity almost completely disappeared. One ally was enough.
1. Little Albert
In 1920, John B. Watson wanted to prove that fear is learned. His subject was a baby of about nine months, known as Albert, who was not afraid of anything and happily played with a white rat.
Every time Albert reached for the rat, Watson and his assistant Rosalie Rayner struck a steel bar with a hammer right behind his head. Soon, Albert cried at the sight of the rat, and the fear spread to a rabbit, a dog, a fur coat and even a Santa Claus mask. The fear was never undone. Who Albert really was is still debated, and the experiment has never been successfully replicated.
