The False Consensus Effect
A reader's summary of the bias that makes one's own views look more widely shared than they are — covering the 1977 Stanford sandwich-board study, how the effect was measured and explained, the argument that it is partly reasonable, and why an online environment that surrounds each person with similar people gives the bias plenty to feed on.
The term at a glance
When I guess what most people think, I start from what I think. The false consensus effect names the resulting skew: people who hold a view estimate it to be more common than people who do not hold it. It is a claim about estimates of others, and it says nothing about whether the view in question is right. It is regularly confused with its mirror image, pluralistic ignorance, in which people wrongly think they are in the minority.
Origin
The effect was named and demonstrated by Lee Ross, David Greene and Pamela House in 1977, in a paper titled “The false consensus effect: An egocentric bias in social perception and attribution processes.” Ross, a Stanford social psychologist, was already known for work on attribution and later became prominent for research on naive realism, the feeling that I see the world as it is and that reasonable people will agree. The false consensus effect is one of the standard experimental demonstrations of that broader tendency.
History and context
The paper arrived in the decade when social psychology was cataloguing systematic errors in everyday judgement, alongside the heuristics-and-biases programme of Daniel Kahneman and Amos Tversky. Marks and Miller's 1987 review treated it as one of the better-replicated findings in the area, while arguing about its causes. The idea is related to the confirmation bias, because both protect the beliefs one already has, and to the spotlight effect, because both let an individual perspective stand in for the group's.
Main ideas
My own view feels like the common one
The false consensus effect is the tendency to overestimate how widely one's own opinions, preferences and behaviours are shared. People who like a film, vote a certain way or skip the gym guess that a larger share of others do the same than people who chose the opposite would guess. The error is relative: it is a gap between groups of estimators, not proof that anyone is wildly wrong about the average.
The study: the sandwich board
Lee Ross, David Greene and Pamela House reported four studies in a 1977 paper in the Journal of Experimental Social Psychology. In the best known, Stanford students were asked to walk around campus for thirty minutes wearing a sandwich board that read “Eat at Joe's.” They were free to refuse. Students who agreed estimated that about 62% of others would agree; students who refused estimated that about 33% would. Each group saw its own choice as the typical one.
Not just attitudes: the guess shapes the judgement of people
The same paper found that participants drew stronger conclusions about the character of people who chose differently from them. Someone who made the opposite choice was read as revealing something about their personality, while one's own choice was read as the natural response to the situation. That pairing connects the effect to the fundamental attribution error covered elsewhere on this site.
A robust finding, with a modest size
Gary Marks and Norman Miller's 1987 review in Psychological Bulletin concluded that the effect appears reliably across many topics and samples. Later work, including Joachim Krueger and Russell Clement's 1994 analysis of the “truly false consensus” question, found that the pattern shows up even when people are given a chance to correct it. Effect sizes are typically moderate, and they vary with how the question is asked.
Why it happens: the explanations compete
Ross's original account stressed construal: people assume that their reading of a situation is the obvious one, so others would respond as they did. Other explanations include selective exposure, since we spend time with people like ourselves and our sample of others is biased; motivation, since sharing a view feels reassuring; and plain inference, since one's own opinion is a legitimate piece of evidence about others when little else is known. Researchers continue to argue about how much each contributes.
Is it actually irrational?
A line of work associated with Robyn Dawes in the late 1980s argued that using your own choice as a data point about the population is not a mistake. If you know nothing else, one observation is a reasonable estimate, and a sample of one that disagrees with a prior is better than none. The effect becomes an error mainly when other, better evidence is available and is ignored.
Online, you choose your sample of other people
In person, the people around me are partly chosen for me, by geography, work and family. On a social platform, I follow, mute and block, and a recommendation system learns what I engage with. The result is a sample of opinion that is far more like me than the population is, which is exactly the condition that makes the personal-sample explanation of the effect bite. The mechanism is the same one described under echo chambers, and it is amplified by the fact that, as the one-percent rule suggests, the loudest members are not typical of the audience.
It often shows up as surprise
The visible symptom is the line “I don't know anyone who voted for that” or “nobody actually uses that product.” Each is an honest report of one person's sample, offered as a report on everyone. Election results, product launches and viral controversies regularly produce the same reaction in the group that turned out to be the smaller one.
Critique
- Part of it may be sensible inference. Using your own view as evidence about others is rational when you have little else. The finding shows a difference between groups of estimators, which is consistent with people updating from a biased but honest sample.
- The wording of the question matters. Results vary with whether people are asked about opinions, behaviours or traits, and about their own group or a broad population. Some studies find much larger effects for in-group estimates than for estimates of the general public.
- It can be turned into an accusation. Telling an opponent that they suffer from false consensus is itself an unfalsifiable move unless there is a measurement of the real distribution. The bias is a reason to check a number, not a way to dismiss a position, and it applies to the person pointing it out as well.
Impact
The practical habit is to separate “what the people I see think” from “what people think,” and to look for a real number before treating the first as the second. A poll, an election result, sales data or the answers of someone outside my usual circle all count for more than my sense of the room. It also helps in a disagreement: assuming the other person does not share my premises, rather than that they are being obtuse, is the first step of steelmanning. And since the effect and the bandwagon effect pull in opposite directions, one making me overrate agreement and the other making me follow it, it is worth knowing which of them is doing the work.