The backfire effect is the idea that showing someone evidence against their belief makes them hold it harder. Brendan Nyhan and Jason Reifler reported it in 2010 and it became one of the most repeated findings in popular psychology. It has since largely failed to replicate. Thomas Wood and Ethan Porter ran five experiments with more than 10,000 participants and could not produce it once.
Corrections, it turns out, mostly work. Which leaves a harder question: if facts land, why does almost nobody get talked out of a manipulative relationship?
The short version
- The backfire effect is real in the sense that it was measured once. It has not held up.
- Across 10,000+ participants, corrections moved people toward the facts.
- The original authors accepted the non-replication. Most of the internet did not notice.
- Evidence is not rejected. It is absorbed and explained. That is the real trap.
What the Backfire Effect Claimed
In When Corrections Fail (Political Behavior, 2010), Nyhan and Reifler gave participants a false claim, then a correction. In some ideological subgroups, the corrected participants came out more convinced of the falsehood than people who never saw the correction at all.
The proposed mechanism was intuitive. When a belief is fused with identity, a correction does not read as information. It reads as an attack, and the mind defends. It explained cults, conspiracies, and every argument you have ever lost at a dinner table.
It was a tidy, satisfying, deeply quotable idea. That should have been the first warning.
What Happened When People Went Looking for It
Wood and Porter set out to reproduce the finding at scale in The Elusive Backfire Effect: Mass Attitudes’ Steadfast Factual Adherence (Political Behavior, 2019). Five experiments. More than 10,000 participants. Deliberately chosen polarising issues — exactly where backfire should have been easiest to catch.
They found none. Not a weakened version. None. Across the board, corrections nudged people toward accuracy, including on issues people cared about intensely.
The most interesting part is what Nyhan and Reifler did next: they accepted it. They have since worked on and published research consistent with the view that backfire is rare. That is what the process is supposed to look like, and it is worth saying plainly, because the correction never travelled as far as the original.
Why You Believed It Anyway
Here is the part that belongs on this site.
The backfire effect spread because it was useful. It explained why the people who disagreed with you were unreachable, and it did so without requiring you to consider that your argument was weak, your evidence thin, or your delivery hostile. It handed everyone a licence to stop trying and feel scientific about it.
Then the correction arrived, and the idea survived. It is still repeated in articles, videos, and training decks written years after the replication failed.
A belief that flatters you, resists correction, and keeps circulating anyway. That is not evidence for the backfire effect. It is a demonstration of something more ordinary and much harder to escape.
The Real Mechanism Is Worse Than Backfire
Ziva Kunda’s review of motivated reasoning (Psychological Bulletin, 1990) sets out the condition that matters. People cannot simply believe whatever suits them. Motivated reasoning only works when you can construct an apparently reasonable justification.
So the evidence you present does not bounce off. It lands. It gets accepted — and then a justification gets built around it. Yes, he said that, and here is why it made sense. Nothing was rejected. Everything was digested.
This is a worse problem than backfire, because backfire at least implies a limit: stop pushing and you stop making it worse. Motivated reasoning has no such limit. The better the person’s mind, the better the justification, which is precisely why smart people fall for manipulation and stay longer once they have.
What Actually Works
If corrections work in a lab and fail in a living room, the difference is not the facts. It is what the belief is holding up.
So stop arguing the evidence and go underneath it. What need is this belief meeting? What would they lose by letting it go — not abstractly, but on Tuesday? Who would they have to become? What would have to be true for them to feel safe enough to look?
Those are the questions that move people, and they explain why cognitive dissonance resolves in whichever direction costs least. Your proof was never the obstacle. It was never the door either.
Frequently Asked Questions
Is the backfire effect real or not?
On current evidence, largely not. It was reported by Nyhan and Reifler in 2010, but Wood and Porter’s five experiments with more than 10,000 participants found no corrections that triggered it. Rare cases may exist; it is not the general rule people treat it as.
Does that mean fact-checking works?
On beliefs, generally yes. Corrections tend to move people toward accuracy. What corrections do not reliably change is behaviour, identity, or who someone trusts, which is why a factually updated person can still stay exactly where they are.
Why do people still cite the backfire effect?
Because it is a satisfying explanation that arrived first and travelled fast. Replications are less quotable than the findings they overturn, so the original keeps circulating in articles and training material written long after it stopped holding up.
Sources
- Nyhan, B. & Reifler, J. (2010). When corrections fail: The persistence of political misperceptions. Political Behavior, 32(2), 303–330. Retrieved 15 July 2026.
- Wood, T. & Porter, E. (2019). The elusive backfire effect: Mass attitudes’ steadfast factual adherence. Political Behavior, 41, 135–163. Retrieved 15 July 2026.
- Kunda, Z. (1990). The case for motivated reasoning. Psychological Bulletin, 108(3), 480–498. Retrieved 15 July 2026.
