The question
Earlier studies of “suggestion” had shown that people shift their judgments toward those of others. Asch thought that was not the whole picture: on a vague question, leaning on others can be perfectly reasonable. Solomon Asch wanted to know something sharper: what happens when the group is plainly, visibly wrong? Would people trust their own eyes, or the room?
How it was done
A group of seven to nine young men gathered in a classroom for what they were told was a test of visual judgment. On each round they saw two cards: one with a single line, and one with three lines of different lengths. Each person said aloud which of the three matched the single line.
Only one person at the table was a real participant. Everyone else was working with Asch and had been told what to say. The real participant was seated next to last, so he heard almost everyone else answer before his turn.
A series had 18 rounds. On 6 of them the group answered correctly. On the other 12, the “critical” rounds, the group unanimously chose the same wrong line. The differences were not subtle: between about 2 and 4.5 centimetres (¾ to 1¾ inches) in the main series. When people did the same task alone, they got it wrong less than 1% of the time.
The main series included 123 participants from three colleges. Asch first reported the experiments in 1951, described them for a general audience in Scientific American in 1955, and published the full programme in a 1956 monograph.
What came out
Faced with a unanimous wrong majority, participants went along with it on 36.8% of the critical rounds. The other 63.2% of their answers stayed correct.
The averages hide large differences between people:
- About one in four participants never went along with the group, not even once.
- About three in four went along at least once.
- Only about one in twenty (6 of 123) went along on all 12 critical rounds.
- People were fairly consistent: those who resisted early tended to keep resisting, and those who yielded early tended to keep yielding.
Asch then changed the conditions, one at a time:
- Size of the majority. One person contradicting the participant had little effect. Two produced errors on 13.6% of rounds, and three on 31.8%. Beyond that, adding more people changed little.
- One ally. When a single other person gave the correct answer, the majority lost most of its pull, down to about a quarter of its usual effect.
- Losing the ally. When that ally switched to the majority after six rounds, errors jumped back up to roughly the level of a unanimous group.
- Any dissent helps. Even a dissenter who gave a different wrong answer weakened the majority. When he picked a line even further from the correct one, errors fell to 9%, and the few that remained were mild.
In interviews after the session, the people who held out did not all feel confident. Some were sure of themselves; others doubted their own eyes but felt they had to report what they saw. Among those who yielded, some came to believe the group must be right, while others knew it was wrong and went along anyway. A few said they had actually come to see the lines the way the group did.
What really happened (replications and critiques)
The “conformity” label is half the story. Asch titled his 1956 monograph Studies of independence and conformity, with independence first, and he was at least as interested in what lets people resist. Yet most retellings focus on the 36.8% and the “three in four”. In 1990 Friend, Rafferty and Bramel showed that textbooks had increasingly presented the study as proof of conformity. In 2015 Richard Griggs looked at 20 popular introductory textbooks: 16 cited the “75% conformed at least once” figure, and none mentioned that about 95% of participants also defied the group at least once. The data support both readings. Most people bent a little, and most also stood their ground most of the time.
The effect replicates, but its size moves. A 1996 meta-analysis by Bond and Smith pooled 133 Asch-style studies from 17 countries:
- In the United States, conformity declined from the 1950s onward.
- Conformity was higher in more collectivist cultures than in more individualist ones.
- It was higher when the majority was larger, when the lines were harder to tell apart and in samples with more women, and lower when the majority were outsiders rather than peers.
Earlier, British researchers Perrin and Spencer had failed to reproduce the effect with British students and called it “a child of its time” (1980). In a 1981 follow-up with other groups in Britain, including unemployed young men of West Indian origin, they did find high conformity, which suggests the setting and the people matter.
A modern replication. In 2021, Axel Franzen and Sebastian Mader re-ran the experiment at the University of Bern with 210 students (202 valid), using five confederates. Without rewards, participants went along with the wrong answer on 33% of critical rounds (in a shorter, 10-round version of the task), very close to Asch’s figure. When each correct answer earned one Swiss franc, errors fell to 25%: lower, but far from zero. On political statements, the same students showed a conformity rate of about 38%. Of the personality traits measured, only openness showed a link (less conformity), and the authors say that result needs to be replicated.
Without actors. Some critics wonder whether the confederates’ acting gave the game away. In 2010 Kazuo Mori and Miho Arai, working in Japan with students who knew each other, used special glasses so that participants physically saw different lines, with no actors at all. They found conformity among women but not among men, though the number of people in the minority position was small (16 women and 10 men, out of 104 students).
Ethics
The design depended on deception. Participants believed they were in a vision test with other volunteers. They did not know the purpose, and they did not know the others had been told what to say. Being contradicted by a unanimous group, round after round, was stressful for some of them. Asch interviewed every participant after the session.
By today’s standards the core design could still be approved. Clear consent wording, a full explanation afterwards and an explicit right to stop would be required, and the 2021 Bern study, for example, obtained verbal consent and debriefed participants by email.
What it means today
The line task is artificial, but the mechanism shows up wherever people answer in front of others: in meetings where the most senior person speaks first, in comment sections, and in group chats where nobody wants to be the one who disagrees.
The most useful finding is also the most hopeful one. A single dissenting voice breaks most of the pressure. That suggests some practical habits, though they go beyond what Asch himself tested:
- If you disagree, saying so may free others who silently agree with you.
- If you lead a group, collecting answers privately or in writing before the discussion reduces the pull of the first voices.
- Being alone in a minority is hard. Finding one ally, even one who doesn’t fully agree with you, changes the situation.
Would you have passed?
Coming soon. See the cards, hear the “group” answer, and choose. At the end you’ll see how often you went along, next to Asch’s 36.8% and the 33% found in Bern in 2021. One honest caveat: on a screen you know the group isn’t real, so the pressure is much weaker than in a room full of people looking at you.
Questions people ask
What percentage of people conformed in the Asch experiment?
Across the 12 critical trials, 36.8% of answers matched the wrong majority. About three in four participants went along at least once, but about one in four never did, and only about one in twenty went along every time.
Has the Asch experiment been replicated?
Yes, many times. A 1996 meta-analysis pooled 133 studies from 17 countries. A replication run in Switzerland in 2021 (published in 2023) found an error rate of 33%, very close to Asch's original figure.
What reduces conformity in the Asch experiment?
Above all, one ally. When a single other person gave the correct answer, the majority's influence fell to about a quarter of its usual strength. Rewards for correct answers also reduced errors.
Was the Asch experiment ethical?
It relied on deception: participants believed they were in a vision test with other volunteers. Many participants found it uncomfortable. Asch interviewed every participant afterwards and then told them what the study was really about.
Sources
- Asch, S. E. (1951). Effects of group pressure upon the modification and distortion of judgments. In H. Guetzkow (Ed.), Groups, leadership and men (pp. 177–190). Carnegie Press.
- Asch, S. E. (1955). Opinions and social pressure. Scientific American, 193(5), 31–35. doi:10.1038/scientificamerican1155-31
- Asch, S. E. (1956). Studies of independence and conformity: I. A minority of one against a unanimous majority. Psychological Monographs: General and Applied, 70(9), 1–70. doi:10.1037/h0093718
- Bond, R., & Smith, P. B. (1996). Culture and conformity: A meta-analysis of studies using Asch's (1952b, 1956) line judgment task. Psychological Bulletin, 119(1), 111–137. doi:10.1037/0033-2909.119.1.111
- Franzen, A., & Mader, S. (2023). The power of social influence: A replication and extension of the Asch experiment. PLOS ONE, 18(11), e0294325. doi:10.1371/journal.pone.0294325
- Griggs, R. A. (2015). The disappearance of independence in textbook coverage of Asch's social pressure experiments. Teaching of Psychology, 42(2), 137–142. doi:10.1177/0098628315569939
- Friend, R., Rafferty, Y., & Bramel, D. (1990). A puzzling misinterpretation of the Asch 'conformity' study. European Journal of Social Psychology, 20(1), 29–44. doi:10.1002/ejsp.2420200104
- Mori, K., & Arai, M. (2010). No need to fake it: Reproduction of the Asch experiment without confederates. International Journal of Psychology, 45(5), 390–397. doi:10.1080/00207591003774485
- Perrin, S., & Spencer, C. (1980). The Asch effect: A child of its time? Bulletin of the British Psychological Society, 33, 405–406.
- Perrin, S., & Spencer, C. (1981). Independence or conformity in the Asch experiment as a reflection of cultural and situational factors. British Journal of Social Psychology, 20(3), 205–210. doi:10.1111/j.2044-8309.1981.tb00533.x