The Backfire Effect
A reader's summary of the claim that correcting a false belief can sometimes make someone believe it more strongly — the 2010 study that first measured it, the blog post that turned it into common knowledge about how minds work, and the large 2019 replication that found the reversal itself is the rare case, not the rule.
The effect at a glance
The backfire effect names the finding that showing someone a factual correction can, under specific circumstances, increase rather than decrease their belief in the false claim being corrected — a reversal of direction, not merely a correction that fails to land. It became one of the most frequently cited reasons for pessimism about fact-checking, invoked constantly in online arguments as proof that correcting someone is worse than useless, even as the research behind it moved toward a narrower and more skeptical position than the popular version retained.
Origin
Brendan Nyhan, then at the University of Michigan, and Jason Reifler, at Georgia State University, published “When Corrections Fail: The Persistence of Political Misperceptions” in the journal Political Behavior in 2010. Their experiments presented participants with a mock news article containing a real political misperception, followed immediately by a correction, then measured belief in the original claim afterward, split by participants' ideology. On a handful of items — most notably the claim that Iraq possessed weapons of mass destruction before the 2003 invasion — conservative participants who received the correction ended up more confident in the false claim than conservative participants who never saw it.
History and context
The finding sat as one result among several in a technical political-science paper until writer David McRaney generalized it for a general audience in a 2011 post on his blog You Are Not So Smart, and later in the book of the same name — framing “the backfire effect” as a standing feature of how belief correction works, not a pattern observed on two items in one study. The term spread further through journalism and fact-checking commentary after 2016, often cited as a reason correcting misinformation might be counterproductive. That confidence came under direct pressure in 2019, when Thomas Wood and Ethan Porter published “The Elusive Backfire Effect,” testing 52 political misperceptions across large samples and finding that corrections moved belief in the accurate direction in nearly every case — the direction-reversing pattern Nyhan and Reifler had reported on a small number of items did not generalize as a common response to correction.
Main ideas
The 2010 study that named it
Political scientists Brendan Nyhan and Jason Reifler had participants read a mock news article containing a political misperception — for instance, that Iraq had weapons of mass destruction before the 2003 invasion — followed by a correction, and found that for some subgroups, notably ideologically conservative participants on the WMD item, the correction increased misperception rather than reducing it.
"Backfire" meant a direction reversal, not just a failed correction
The reported finding wasn't that corrections did nothing — a null result — but that belief in the false claim measurably strengthened after exposure to correcting information, among the specific subgroup for whom the false belief was ideologically congenial.
It only showed up on some of the tested items
Nyhan and Reifler's 2010 paper tested several political claims, and the reversal pattern appeared clearly on some — the Iraq WMD claim and a claim about the 2001 Bush tax cuts raising revenue — but not consistently across every item studied, a qualifier that dropped out of most retellings.
David McRaney's blog turned a narrow finding into a household term
Writer David McRaney popularized "the backfire effect" as a named cognitive bias in a widely shared 2011 post on his blog You Are Not So Smart, generalizing the specific experimental pattern into a broad claim that correcting anyone's false belief tends to entrench it — a framing that outran what the original study had shown.
A large 2019 replication found genuine backfire is rare
Thomas Wood and Ethan Porter tested 52 political misperceptions across large samples in a paper titled "The Elusive Backfire Effect" and found that corrections moved beliefs in the accurate direction in nearly every case tested — direction-reversing backfire essentially failed to show up as a general phenomenon.
Nyhan's own later work is more cautious than the popularization
Nyhan has since cited the Wood and Porter replication and related work acknowledging that genuine backfire effects appear to be real but uncommon under specific conditions, rather than the default response to correction the term is now used to imply.
Critique
- The original evidence base was narrow. A pattern observed on two or three politically charged claims, in one set of experiments, is a thin foundation for the sweeping claim — that correcting false beliefs generally backfires — the term came to carry in popular usage.
- The large replication points the other way. Wood and Porter's 2019 test across 52 issues found accurate- direction updating was close to universal; direction-reversing backfire, tested at scale, looks like the exception rather than a general law of how people respond to correction.
- It became an excuse not to correct anything. In practice, “it'll just backfire” got used to justify skipping corrections altogether, despite evidence now leaning toward corrections usually helping at least somewhat.
- It gets confused with a different, better-supported risk. Repeating a false claim in order to debunk it can make the claim itself more familiar, which is a real risk documented under the Illusory Truth Effect — but that is a mechanism about repetition and familiarity, not the directional belief-reversal the backfire effect specifically claimed.
Impact
The rise and partial fall of the backfire effect is now itself a case study cited in debates about how fast a narrow finding can become received wisdom once it travels outside the paper that produced it. It sharpens Brandolini's Law's point about the asymmetric cost of refuting bad claims: the backfire effect briefly supplied an argument that refutation wasn't just costly but actively counterproductive, before a larger dataset walked that argument back to something closer to “correction is worth the cost after all, most of the time.”