Social Cognition
Group Conformity Task
Also called: conformity paradigm, line judgment conformity task, Asch paradigm, Asch-type line judgment task, group pressure task
Measure how often people match a group's wrong answer: participants match lines alone, then after seeing a simulated group's answers, some of them wrong, scored by the share matched.
4-7 minUntimed
Free to start: 10 free participants per account, then $0.99 per participant for academic research, or any yearly plan. Try it here first, with no account.
- Papers, 2021 to 2025
- 31
- Evidence grade
- D
- Registered clinical studies
- 0
What it measures
Measures how often people match a group's wrong answer. Participants match a line to one of three, 12 times alone and then 18 times after seeing a simulated group's answers, which are wrong on 12 rounds. The main score is the share of those wrong answers the participant matches.
Constructs
- Social influence on judgment
- Weighing a group's answer against one's own perception
- Normative and informational influence
- Perceptual judgment under uncertainty
Research uses
- Social psychology research on conformity
- Studies of social influence and decision making
- Research on trust in machine and group advice
- Group comparison within a study
How the paradigm works
The Group Conformity Task measures how often people give a group's wrong answer on a matter they can check with their own eyes. In the studies of Asch (1951, 1956), a person matched a standard line to one of three numbered lines after a group of instructed students had answered aloud. Over 18 trials the group gave the same wrong answer on 12. Judging alone, people erred on under 1% of estimates. With the group, 36.8% of estimates on those 12 trials were errors, and about a quarter of people made none. In a replication with five confederates, 33% of answers on such trials were wrong, and 25% when right answers were paid (Franzen and Mader, 2023). Across 133 studies from 17 countries, conformity on this task had fallen in the United States since the 1950s and was higher in collectivist countries (Bond and Smith, 1996), and how it changes with group size depends on the setting (Bond, 2005).
More about the paradigm
Here the group is simulated, and the standard instructions say so; a neutral wording leaves that line out, and the last screen always says the group was a computer program and that its wrong answers were set in advance. No screen shows a name or a face: each answer is a letter and a line number. A group known to be a machine may move people less: adults resisted a group of small humanoid robots on the line task (Vollmer, Read, Trippas and Belpaeme, 2018), yet people matched a group described as computers on about a third of critical trials on arithmetic and emotion judgments from a brief display (Hertz and Wiese, 2018). So half the rounds use clear differences, of 20% and 35% from the standard, and half use close ones, of 6% and 12%, where a person alone is often unsure. Franzen and Mader mixed easy and hard trials too. Each line's foot sits up to 8 px above or below the others, so comparing tops does not give the answer away; the lines themselves are drawn by our code from the session's seed.
The session has 2 practice rounds with feedback, 12 rounds alone and 18 rounds with the group, the first two of them right and clear, as in Asch. A group of 5 (2 to 8) answers first, one answer every 350 ms, and the participant answers last, with no time limit. Each right answer after the practice earns 10 points and no feedback follows, so a matched wrong answer costs points; points are a setting of the study, and with a hard task accuracy incentives can raise conformity (Baron, Vandello and Brunsman, 1996). One check question with feedback comes first, and it does not mention the group.
The main score is the share of the 12 wrong group answers the participant matches, reported for clear and close rounds and as a count. The rounds alone are built the same way, and each sets aside a wrong line, so the same share can be read with no group: the difference between the two is what the group added. On close rounds, people will sometimes choose the group's line on their own, so read that difference, not the raw share. Accuracy alone and on the rounds where the group was right, and the rounds where the participant chose the other wrong line, are also reported. These are research measures of answers in a line task. They do not describe a person, and no result calls anyone a conformist or a follower. Rates vary with culture and period, and they may run lower than with a human group, so compare people who saw the same instructions. Studies usually explain the purpose afterwards.
Scores
- Wrong Answers Matched%
- Share of the 12 rounds with a wrong group answer in which the participant gave that same answer.
- Read across a group; no better direction
- Wrong Answers Matched, Clear Lines%
- The same share over the 6 rounds whose wrong lines differ from the standard by 20% and 35%.
- Informational
- Wrong Answers Matched, Close Lines%
- The same share over the 6 rounds whose wrong lines differ from the standard by 6% and 12%.
- Informational
- Rounds Matchedcount
- How many of the 12 wrong group answers the participant matched.
- Read across a group
- Same Choice Alone%
- Share of the 12 rounds alone in which the participant chose the wrong line set aside for that round: the same share with no group.
- Baseline for the wrong answers matched
- Shift Toward the Group%
- Wrong answers matched minus the same choice alone: what the group added.
- 0 means no change from judging alone
- Other Wrong Line%
- Share of the rounds with a wrong group answer in which the participant chose the other wrong line.
- Informational
- Accuracy, Group Right%
- Share right on the 6 rounds where the group gave the right answer.
- Informational
- Accuracy Alone%
- Share right on the 12 rounds alone.
- Informational
- Accuracy Alone, Clear Lines%
- Share right on the 6 clear rounds alone.
- Close to 1 expected; under two thirds is flagged for review
- Accuracy Alone, Close Lines%
- Share right on the 6 close rounds alone.
- Informational
- Points Wonpoints
- Points for right answers after the practice.
- Informational
- Check Correctflag
- 1 when the check question on which line to choose was answered correctly, 0 when not.
- 1 supports interpretation
- Median Answer Timems
- Median time from the group's last answer to the participant's, over the rounds with the group.
- Informational
- Answers Under 100 mscount
- Answers given less than 100 ms after the round's screen appeared.
- 0 expected
- Roundscount
- Rounds answered, alone and with the group.
- Informational
Use and evidence
31 papers from 2021 to 2025 name this task in their title or abstract, up 48% compared with 2016 to 2020. That ranks it 354 of 495 test families in our census. See the counts for every test.
Show the numbers
| Year | Papers |
|---|---|
| 2025 | 13 |
| 2024 | 4 |
| 2023 | 7 |
| 2022 | 4 |
| 2021 | 3 |
| 2020 | 4 |
| 2019 | 6 |
| 2018 | 4 |
| 2017 | 5 |
| 2016 | 2 |
| 2015 | 6 |
| 2014 | 6 |
| 2013 | 8 |
| 2012 | 3 |
| 2011 | 4 |
| 2010 | 3 |
| 2009 | 1 |
| 2008 | 3 |
| 2007 | 0 |
| 2006 | 1 |
| 2005 | 1 |
| 2004 | 1 |
| 2003 | 2 |
| 2002 | 1 |
| 2001 | 2 |
| 2000 | 0 |
Use grade D
30 to 99 papers from 2021 to 2025.
Evidence grade D
Little published measurement evidence found.
- Papers on test-retest reliability
- 0
- Papers on reliability or validity
- 7
- Registered clinical studies using it as an outcome
- 0
How the grades work
Use: A is 1,000 or more papers from 2021 to 2025, B 300 to 999, C 100 to 299, D 30 to 99 and E fewer than 30. Evidence: A means at least 25 papers report its test-retest reliability and at least 25 registered clinical studies use it as an outcome; B, at least 8 test-retest papers, and either 5 registered clinical studies or 150 reliability or validity papers; C, at least 2 test-retest papers, 30 reliability or validity papers or 3 registered clinical studies; D, little published measurement evidence found. The counts measure use. They include every paper that names the task in its title or abstract. Grades describe the published research on this task, not ConductCognition's version of it.
Source: ConductCognition test census 0.22.0, counted 2 November 2026 from OpenAlex and ClinicalTrials.gov.
Task design
The trial sequence as the test runs in ConductCognition.
- Each round
- A standard line and three numbered lines, one of them the standard's length, stay on screen until the answer. Each line's foot sits up to 8 px above or below the others, so comparing tops does not give the answer away. There is no time limit.
- Lines
- Drawn by our code from the session's seed, from a standard of 100 to 150 px. On clear rounds the wrong lines differ from it by 20% and 35%, on close rounds by 6% and 12%: half the rounds each. Both wrong lines can be longer, both shorter, or one of each.
- Rounds
- 2 practice rounds with feedback, 12 rounds alone, then 18 rounds with the group: the group is wrong on 12 and right on 6, and the first 2 are right and clear. No feedback follows the practice.
- Group
- Simulated. 5 answers (standard; 2 to 8), all the same, appear one every 350 ms as a letter and a line number, with no name or face, and the participant answers last. The standard instructions say a computer program gives the answers; a neutral wording leaves that line out. The last screen always says the group was a program and that its wrong answers were set in advance.
- Baseline
- The 12 rounds alone are built like the 12 wrong-answer rounds, one for each level, shape and wrong line, and each sets aside the wrong line a group would have chosen.
- Points
- 10 for each right answer after the practice, shown only at the end. Points are hypothetical by default; a setting changes the instruction line to say the study pays.
- Check
- One question with feedback before the practice: which line to choose. It does not mention the group.
- Scores
- The share of wrong group answers matched, overall and for clear and close lines; the same share alone and the difference between them; accuracy alone and on the rounds where the group was right.
Asch (1956) seated each person among 6 to 8 instructed students who answered aloud. Here the group is simulated and the participant sees its answers on screen, and rounds alone stand in for a separate control group. Adults resisted a robot group on clear lines (Vollmer and colleagues, 2018), so rates may run lower than with a human group, and half the rounds use close lines, where people are less sure of their own judgment. On those rounds people sometimes choose the group's line on their own, so read the difference from the rounds alone. The percentages are ConductScience choices that no person has yet run. Studies usually explain the purpose afterwards. These are research measures of answers in a line task: no result calls anyone a conformist or a follower, and they do not diagnose.
Quality checks
Each session is marked valid, needs review or invalid, with the reason, before you see its scores.
Scored rows: the scored run's rounds alone and with the group. Accuracy is judged against chance (one in three) over the rounds alone and the rounds where the group was right; the rounds with a wrong group answer are left out, so matching the group never counts against a session. Accuracy below chance, one line number on at least 90% of rounds without accuracy above chance, and answers under 200 ms on at least half the rounds invalidate. A run short of its planned rounds, a wrong check answer and under two thirds right on the clear rounds alone are flagged for review.
Why these rules
Matching a group's wrong answer is what this test measures, so it can neither fail nor flag a session. The rounds alone and the rounds where the group was right still have a right answer that people find, so they show whether the lines were judged at all. Accuracy significantly below chance (one-sided exact test, p < .05) cannot come from guessing; it means the response mapping was reversed or the task was not followed (Slick, Sherman and Iverson 1999 treat below-chance forced-choice performance as definite response bias). Chance-level accuracy can be a real result in impaired participants, so it is only flagged. One option on at least 90% of answered trials, when the correct options are balanced and accuracy is not above chance, means the answers did not depend on the items (the task counterpart of long-string responding in surveys; Curran 2016). An answer under 200 ms is faster than a simple visual reaction, so it cannot reflect the item. Following response-time effort screening for untimed tests (Wise and Kong 2005), the session is invalid only when at least half the answers are that fast.
No percentile. Scores and quality flags only.
Running this test
Settings you can change
| Setting | Standard | Allowed |
|---|---|---|
| Group size | 5 | 2 to 8 |
| Group wording | Simulated by a computer program (standard) | Simulated by a computer program (standard), Neutral: a group |
| Payment line | Hypothetical points (standard) | Hypothetical points (standard), The study pays for points |
Yearly plans can change settings within these limits. Free and pay-as-you-go studies use the standard settings.
Need a setting this test does not have?
Devices and languages
- Untimed test: runs on desktops, laptops, tablets and phones. Participant devices
- Languages: English, Spanish, French, German, Italian, Portuguese (Brazil), Chinese (Simplified) and Japanese.
- Available on: Every plan, including the free plan.
- No percentile. Scores and quality flags only. How norms work
- Recruiting online? See the setup guides for Prolific, Sona Systems, CloudResearch Connect and Qualtrics.
Cite and describe
Participants completed the Group Conformity Task in a web browser (ConductCognition task version 1.0.0, scoring version 1.0.0; Asch SE, 1951, 1956), with the standard settings (group size: 5; group wording: Simulated by a computer program; payment line: Hypothetical points). The main score was wrong answers matched (share of the 12 rounds with a wrong group answer in which the participant gave that same answer).
ConductCognition versions: task version 1.0.0, scoring version 1.0.0. Every export records the versions each session used, for a methods section.
- 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. The first report of the line judgment paradigm.
- 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. https://doi.org/10.1037/h009371818 trials, 12 of them critical, and a control group judging alone.
- Franzen, A., & Mader, S. (2023). The power of social influence: A replication and extension of the Asch experiment. PLOS ONE, 18(11), e0294325. https://doi.org/10.1371/journal.pone.0294325A recent replication with easy and hard trials and paid right answers.
- 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. https://doi.org/10.1037/0033-2909.119.1.111Change over time and across cultures.
- Bond, R. (2005). Group size and conformity. Group Processes & Intergroup Relations, 8(4), 331-354. https://doi.org/10.1177/1368430205056464Why group size is a setting.
- Vollmer, A.-L., Read, R., Trippas, D., & Belpaeme, T. (2018). Children conform, adults resist: A robot group induced peer pressure on normative social conformity. Science Robotics, 3(21), eaat7111. https://doi.org/10.1126/scirobotics.aat7111Adults resisted a robot group on the line task.
- Hertz, N., & Wiese, E. (2018). Under pressure: Examining social conformity with computer and robot groups. Human Factors, 60(8), 1207-1218. https://doi.org/10.1177/0018720818788473Conformity to a group described as computers.
- Baron, R. S., Vandello, J. A., & Brunsman, B. (1996). The forgotten variable in conformity research: Impact of task importance on social influence. Journal of Personality and Social Psychology, 71(5), 915-927. https://doi.org/10.1037/0022-3514.71.5.915Accuracy incentives with easy and hard tasks.
Source papers
- Asch (1956): 123 male college students among instructed groups of 6 to 8 made errors on 36.8% of 12 critical estimates (mean 4.41 errors); 29 made none. 37 people judging alone in writing erred on under 1%. Franzen and Mader (2023): 33% wrong on 6 critical trials with five confederates, 25% with pay for right answers. All of these used human groups. No published rate applies to a group known to be simulated or to these lines.Paradigm source, Adults
Questions
How long does the Group Conformity Task take?
About 4 to 7 minutes, as listed in the test library.
What device does a participant need?
Untimed test: runs on desktops, laptops, tablets and phones. The browser must be Chrome, Edge, Firefox or Safari.
What scores does it produce?
Wrong Answers Matched, Wrong Answers Matched, Clear Lines, Wrong Answers Matched, Close Lines, Rounds Matched, Same Choice Alone, Shift Toward the Group, Other Wrong Line, Accuracy, Group Right, Accuracy Alone, Accuracy Alone, Clear Lines, Accuracy Alone, Close Lines, Points Won, Check Correct, Median Answer Time, Answers Under 100 ms and Rounds. Each session is also marked valid, needs review or invalid, with the reason.
How widely is the Group Conformity Task used in research?
31 papers from 2021 to 2025 name it in their title or abstract, up 48% compared with 2016 to 2020. That ranks it 354 of 495 test families in the ConductCognition test census (use grade D). Source: ConductCognition test census 0.22.0, counted 2 November 2026 from OpenAlex and ClinicalTrials.gov.
How much reliability evidence does the Group Conformity Task have?
In the published research, 0 papers report its test-retest reliability and 0 registered clinical studies use it as an outcome (evidence grade D in the ConductCognition test census).
Does it report a percentile?
No. Scores and quality flags only.
Which settings can a researcher change?
Group size (2 to 8), group wording (Simulated by a computer program (standard), Neutral: a group) and payment line (Hypothetical points (standard), The study pays for points). Settings can be changed on yearly plans; free and pay-as-you-go studies use the standard settings.
Which languages can participants use?
Participants can take this test in English, Spanish, French, German, Italian, Portuguese (Brazil), Chinese (Simplified) and Japanese. English is on every plan, and the other languages come with yearly plans. Have a native speaker read the machine-translated text before launch.
Is this a clinical test?
No. ConductCognition is for research use only. It runs, scores and exports the test; clinical interpretation is outside the platform.
Every test: the full test list.
Run this test in your next study
Start free with 10 free participants on any test, then use participant credits or a yearly plan.
Related tests