neur.ro

Asch conformity experiments

type: effect · origin: study · evidence: supported · themes: how we judge others

Claim assessed: When a group unanimously gives an obviously wrong answer, many people repeat it, even though they can see it is wrong.

Published · how we work

In short

In Solomon Asch’s experiments in the 1950s, a participant sat at a table with several other “participants” who were in fact the experimenter’s confederates. Each had to say aloud, in turn, which of three lines was as long as a reference line. The task was very easy, but on some trials the confederates all gave the same wrong answer. About a third of the real participants’ answers followed the group. The result has been repeated many times, including in replications from 2019 and 2023, but it varies with culture, with the era and with who makes up the majority. The less-known part: most participants stayed independent on most trials.

What it says

The experiment. The real participant answered near the end, after hearing the others. On ordinary trials the confederates answered correctly; on the “critical” ones they all gave the same wrong answer. According to Franzen and Mäder (2023), Asch reported an average of 36.8% wrong answers on the critical trials across 123 participants. According to Ušto et al. (2019) and Capuano and Chekroun (2024), about three quarters of participants gave in at least once, and the rest never did. We haven’t read Asch’s studies; we describe them from these sources.

The meta-analyses. Bond and Smith (1996) collected 133 studies from 17 countries. In 44 strict replications, conforming answers averaged 25%. Conformity declined in US studies after the 1950s and was higher in collectivist than in individualist cultures. Bond (2005), across 125 Asch-type studies, found an average effect size of 0.89, that is, a large effect, and showed that group size matters differently depending on whether answers are public or private. We describe both from the sources we read and from the abstracts; the articles are available only behind a paywall.

Recent replications.

And failed replications. According to Ušto et al., Perrin and Spencer (1981) obtained, with British students, a single conforming answer out of 396, and Lalancette and Standing (1990), with Canadian students, none. Perrin and Spencer called the effect “a child of its time”. Later replications in Britain, Kuwait, Japan, Bosnia and Switzerland nonetheless found conformity.

Percentage of answers that followed the group on critical trials: Asch, 1950s, 36.8%; average of 44 strict replications, 25%; Franzen and Mäder 2023, 33%, and with pay for the correct answer, 25%; Ušto et al. 2019, Bosnia, 35.4%; Perrin and Spencer 1981, Britain, 1 in 396. A third, for 70 years, with exceptions Answers that followed the group on trials where the group was wrong Asch, 1950s 36.8% 44 strict replications (Bond and Smith) 25% Switzerland, 2023 33% Switzerland, 2023, paid for correct answers 25% Bosnia, 2019 35.4% Britain, 1981 1 in 396
Asch, from Franzen and Mäder (2023); strict replications, from Bond and Smith (1996) as cited by Franzen and Mäder; Bosnia, Ušto et al. (2019); Britain, Perrin and Spencer (1981), from Ušto et al. The studies differ in the number of trials and confederates.

Why do we conform? The classic explanation distinguishes two motives: we want to be accepted by the group (normative influence) or we think the group knows better (informational influence). Bond (2005) found that group size matters differently depending on whether answers are given aloud or in private, which suggests both motives are at work. Capuano and Chekroun (2024) note that people conform less readily when a moral norm is at stake than a social convention. Hodges and Geyer (2006), as summarised in the same review, read the results differently: participants don’t simply submit, but try to balance truth with respect for the group, alternating correct and conforming answers.

Example

In a meeting, every colleague says in turn that the proposed plan is realistic, although one of them can clearly see the deadline is impossible. When their turn comes, they also say “yes, it works”. A single colleague saying “I don’t think so” before them would have made their answer much easier. The example is an illustration, not a study cited here.

How to apply it

The steps below are a practical way to apply it that we propose, based on the studies cited.

  1. Ask for opinions in writing, before the discussion. The pressure is greater when answers are given aloud, in turn. A vote or a written note beforehand lets people say what they see.
  2. Let the least influential people speak first. Asch’s participant answered after hearing the group. Order matters.
  3. Be the voice that disagrees. The pressure in Asch’s experiment comes from a unanimous group. A different answer said aloud breaks the unanimity for those who come after.
  4. Say what is at stake. Paying for the correct answer reduced conformity, and moral norms hold up better than conventions. When a decision matters, say so explicitly.

Limits and nuances

Sources

See also: Bystander effect, Milgram experiment, Stanford prison experiment