In 1931, Katz and Allport asked college students how often they thought their classmates cheated on exams. The answers spread wide, which was expected. What was not expected was how tightly each answer tracked the person giving it. Students who cheated estimated that cheating was common. Students who did not cheat estimated that it was rare. Nobody in that study was describing a campus. Each of them was describing themselves, and calling it a survey.
Forty-six years later, Lee Ross, Greene and House put a name on the pattern in the Journal of Experimental Social Psychology. The false consensus effect, they wrote, is the tendency to see your own behavioral choices and judgments as relatively common — and relatively appropriate to the circumstances. That second half is doing quiet work. It is not only that you think your taste is popular. It is that you think your taste is normal, and the people who disagree with you are the ones being strange.
🧠 Everyone in the Room Was You
Three gears run this bias, and they all point the same direction. The first is availability: your own opinion is the one you have spent the most time with, so it arrives first and feels like the majority position. The second is self-serving: if your view is widespread, your choices look justified, and the justification feeds the estimate that produced it. The third is what researchers call naïve realism — the private conviction that you are seeing the world as it is, while anyone who sees it differently is seeing it through something.
Put people in a group and it tightens. When a room reaches easy agreement and rarely runs into anyone who objects, the room starts treating its own consensus as the general one. Which means the more insulated the group, the more certain it becomes that it speaks for everyone.
And the sharpest finding in the literature is about what happens when the illusion breaks. Given evidence that no consensus exists, people do not revise their estimate of the world. They revise their estimate of the dissenters, who are then assumed to be defective in some way. It is a beautiful, terrible move: the disagreement cannot be real, so the disagreeing person must be broken.
🤔 Empathy Makes It Worse
Here is where the story stops being about strangers and starts being about your job. In 2015, Hattula and his colleagues published a paper in the Journal of Marketing Research with the finding sitting right in the title: managerial empathy facilitates egocentric predictions of consumer preferences. Managers who worked hardest to imagine the customer’s point of view made predictions that were more about themselves, not less. Perspective-taking did not correct the projection. It laundered it. Six years later, Herzog and his co-authors documented the same pattern across marketing decisions: the people who build and sell things systematically project their own preferences onto the people they are building for, and it does not fade with experience.
There is a real fight about the measurement, to be fair. Dawes argued in 1990 that the way the effect is typically calculated may manufacture part of it, and the meta-analytic work since (Robbins and Krueger, 2005) treats the effect as genuine but strongly moderated by who counts as “others” — your ingroup gets projected onto far more than an outgroup does. The direction holds. The size depends on how you ask.
Culture moves the size too. Choi and Cha compared Koreans and European Americans in 2019 and found the effect stronger among Koreans for political beliefs, personal problems and behavioral choices — but not for traits and values. Their reading is that collectivist contexts put more weight on the situation, so people assume the situation is bending everyone the same way it is bending them.
🔗 The Least Representative User in the Room
Anyone who builds a product carries an invisible version of this study in their pocket. The person who designed the flow knows exactly where the button is, has already read the copy six times, and can no longer feel the friction. So they imagine what a new user would feel — and the imagining hands them back their own experience in the voice of a stranger. This is the trap underneath the standard advice to design for yourself. The instinct to build the thing you wish existed is often right. The assumption that other people want it too is the false consensus effect wearing a product roadmap.
It shows up in writing the same way. A novelist knows why a character finally walks out of the room, and knows the two earlier scenes that made it inevitable, so the draft reads as though the reader has that knowledge too. The reader has only what is on the page. Most of what looks like confusion in an early draft is not a clarity problem — it is an author’s private consensus, printed and published.
Personalized reports and readings run into the same wall. A system that tells someone what their year will hold is partly a mirror for whoever wrote its logic, which is why the detail that lands hardest is usually the specific one, and why validating a reading against its own author proves nothing at all.
🎲 The Future Agrees With Me More
There is a sequel to this bias, and it is bigger. In 2017, Rogers, Moore and Norton described belief in a favorable future: the sense that other people will eventually come around to your position. Future others cannot be surveyed, so you never get contradicted. And the future is generous, because it gives the people who disagree with you time to discover that they were wrong. Belief in a favorable future outweighs present-tense false consensus in magnitude — which is a long way of saying that “history will prove me right” is the most comfortable belief available to a human being.
So here is something to try this week. Pick one opinion you are sure “most people” hold. Say out loud what percentage of people you think agree with you. Then ask five of them, one at a time, before you explain anything. The number you get will come in lower than the number in your head, and the interesting part is not whether you were wrong. It is what you assumed about the people who disagreed.