The Asch Conformity Experiments and a 2023 Replication

Key Points

  • A 2023 PLOS ONE paper by Franzen and Mader replicated the Asch line task with 210 subjects recruited in Bern and reported a 33 percent error rate in its condition without incentives, close to the 36.8 percent it gives for Asch.
  • In that paper, small monetary incentives of one Swiss franc per correct answer lowered the error rate to 25 percent, and the authors wrote that incentives do not eliminate the effect.
  • A 2024 systematic review of conformity research since 2004 reported that recent work confirms conformity across diverse contexts, and described continuing debate about interpretation and about culture.
  • The Bern authors listed limits, including a student sample, strangers as confederates, small incentives and moderate political statements.

Published .

The symmetrical stone front of the University of Bern main building, with a central dome, two corner towers with domed roofs, three tall arched windows over the entrance and a wide green lawn in front under a pale cloudy sky.
The main building of the University of Bern, 2019. The replication recruited its subjects on the University of Bern campus. Photo: Marcel.photo, CC BY-SA 4.0, via Wikimedia Commons, resized, photo source.

In a 2023 paper in PLOS ONE, Axel Franzen and Sebastian Mader replicated the line-judgment experiment of Solomon Asch with 210 subjects recruited on the University of Bern campus, about half of them in a condition without incentives, and found an error rate of 33 percent in that condition, which they described as replicating Asch's original findings1. Their paper reports Asch's own mean error rate as 36.8 percent1. This article reads that paper and a 2024 systematic review to say what the Asch result is, what later work found, and where the questions remain.

What the Asch task was

In the Franzen and Mader description of the original experiment, 6 to 8 confederates gathered in a room and gave false answers when matching a line to one of three reference lines, while one uninstructed subject judged as well1. The 2024 systematic review describes the task as simple and unambiguous, with the confederates giving unanimous incorrect responses on certain trials2.

Franzen and Mader report that Asch found a mean error rate of 36.8 percent among 123 real subjects in the critical trials, the ones in which the group gave the wrong answer1.

The 2023 replication in Bern

The authors recruited 210 subjects on the campus of the University of Bern and placed five confederates in the room with one naive subject, who sat in seat number 51. The task had ten line tasks, and the confederates pointed to the correct line in four of them (the first, second, fourth and eighth), leaving six critical trials1.

After exclusions, 202 valid cases remained1. Almost none of the naive subjects gave a wrong answer when the group gave the correct one1. In the condition without incentives, the average share of incorrect answers on critical trials was 33 percent, which the authors called a perfect replication of Asch's 36.8 percent (t(16) = 0.59, p = 0.57)1. By trial the share was 10 percent in the third task and 44 percent and 47 percent in the sixth and seventh1.

Three extensions

In a randomized comparison, correct answers earned one Swiss franc each. The error rate in that incentivized condition fell to 25 percent, and the authors wrote that monetary incentives do not eliminate the effect of group pressure1. They also noted that one franc per answer is a small incentive, so they left open whether larger rewards would reduce errors further1.

The same paper tested five political statements, with the group's answer randomized, and found an average conformity rate of 38 percent1. The authors described the statements as relatively moderate and general1. On personality, apart from openness, none of the traits they measured (intelligence, self-esteem, the need for social approval and the Big Five) were convincingly related to susceptibility to group pressure1.

What the wider literature says

Franzen and Mader report that a meta-analysis by Bond and Smith covering 44 strict replications (the authors' term) found an average error rate of 25 percent, and that most of those replications used male university students in the United States1. The 2024 systematic review covers conformity research since 2004, identified 48 articles under PRISMA guidelines, and reported that recent work confirms the prevalence of conformity across diverse contexts, echoing Asch's findings2. It also called for more interdisciplinary and intercultural work, comprehensive meta-analyses and replications2.

The review summarizes the 1996 meta-analysis by Bond and Smith, which included 133 studies from 17 countries, as finding conformity higher in collectivist cultures than in individualist ones2. It also reports that a later three-country comparison by Tu and Fischbach (China, Korea and the United States) found no difference in conformity, and that the review suggested overly homogeneous student samples as a possible explanation2. The review describes the two findings as contradictory and says the three-country comparison is limited in what it can show about culture.

How the results are read

The review notes continuing debate about how to interpret Asch's findings. It says Spillane and JoulliƩ applied a theory of authority and read the laboratory conformity as a response to the experimenter's perceived authority, not necessarily behavior seen in real life. It says Hodges and Geyer argued that participants do not conform consistently and instead balance truth against consensus2.

Where this leaves the Asch result

The 2023 Bern paper found an error rate of 33 percent, close to the 36.8 percent it reports for Asch. Its authors listed limits: subjects were students, they were strangers to the confederates, the incentives were small and the political statements were moderate1. The 2024 review says recent work confirms conformity across diverse contexts, echoing Asch's findings, while the explanations and the role of culture remain debated. This article describes published findings and gives no advice about anyone's own decisions.

Trivia

  1. In the 2023 Bern replication, how many confederates sat with each naive subject? Five1.
  2. What average error rate did Franzen and Mader report for the standard line task without incentives? 33 percent1.
  3. How much did the Bern replication pay for each correct answer in its incentivized condition? One Swiss franc1.
  4. According to the 2024 systematic review, how many studies from 17 countries did Bond and Smith's 1996 meta-analysis include? 1332.
  5. How many articles did the 2024 systematic review identify for its review of conformity research since 2004? 482.

Sources

  1. Franzen A, Mader S. The power of social influence: A replication and extension of the Asch experiment. PLOS ONE, 2023. Accessed October 28, 2026.
  2. Capuano C, Chekroun P. A Systematic Review of Research on Conformity. International Review of Social Psychology, 2024. Accessed October 28, 2026.