Conformity
People will endorse an obviously wrong answer rather than stand alone against a unanimous group.
Conformity is adjusting a judgement or a statement to match a group. Asch separated it from genuine persuasion: participants often knew the group was wrong and agreed anyway to avoid standing out. Unanimity is the critical condition, and a single dissenter collapses the pressure.
How it shows up in software
Wherever a product shows the group's answer before collecting the user's own, it is running a conformity experiment. Visible vote counts, aggregate poll results, prefilled defaults endorsed by others, and public comment threads all pull individual judgement toward the visible majority. It corrupts exactly the products whose value depends on independent signal, such as reviews, estimates, and research.
Using it well
- Collect independent input before showing the aggregate, in polls, planning estimates, and research sessions.
- Use anonymous submission where a judgement is socially risky, since anonymity cuts normative pressure specifically.
- Surface a credible minority view alongside the majority, because a single visible dissenter restores independent judgement.
- Reveal all responses at once in group estimation, which is why planning poker hides cards until everyone commits.
Where it turns manipulative
- Showing a running tally during voting to steer the outcome is manipulation of a process presented as a measurement.
- Framing a default as what most users chose, when it is what the company prefers, uses conformity pressure to carry a false claim.
- Suppressing dissenting reviews or comments removes the one thing that reliably frees other users to disagree.
Where you have seen it
Planning Poker tools
Hide every estimate until all participants have submitted, then reveal simultaneously, which removes anchoring on the first senior voice.
Polis
Shows participants statements to agree or disagree with before clustering opinion, rather than displaying current totals during voting.
Stack Overflow
Lets any user post a competing answer that is scored separately, so a dissenting solution can persist next to the accepted one.
What the research says
- Asch, 1951 and 1956Well evidenced
Participants judging which of three lines matched a reference gave the group's wrong answer on about a third of critical trials, and most conformed at least once. Adding one confederate who dissented cut conformity sharply.
Replicated many times. The often-quoted figure describes trials, not people, and the popular retelling frequently overstates how many participants conformed consistently.
- Bond and Smith, 1996Well evidenced
Meta-analysis of 133 Asch-style studies across seventeen countries found conformity declined in the United States over time and was higher in more collectivist cultures.
The effect is real but historically and culturally variable, which argues against treating any single number as the rate.
- Deutsch and Gerard, 1955Well evidenced
Separated normative influence, conforming to be accepted, from informational influence, conforming because others seem to know more, and showed anonymity reduced the normative kind.
Grades are a judgement about the evidence, not about how useful the idea is. Plenty of contested effects are still worth knowing, as long as you do not cite them as settled.