Misinformation Correction Experiment

The backfire effect is when a factual correction increases, rather than decreases, belief in the falsehood among a group for whom the fact is ideologically threatening. Nyhan and Reifler reported such effects for certain conservative misperceptions, suggesting corrections could be counterproductive. The phenomenon is group-specific, so it is detected by estimating correction effects separately by partisanship, and a positive effect among the threatened group is the signature of backfire.

It appears to be rare. Wood and Porter, replicating across many issues with large samples, found that corrections generally move beliefs toward the truth and that backfire is hard to reproduce. The original findings were not fabricated, but they appear fragile and issue-specific. The current consensus is that corrections usually reduce misperceptions, though they often fail to change the attitudes and behavior linked to those beliefs, which is a different and important limitation.

Factual beliefs and political attitudes are distinct. A respondent can accept a corrected fact yet retain the attitude it was supposed to support, by reinterpreting the fact, attributing it differently, or weighting other considerations, a pattern consistent with partisan motivated reasoning. This is why correction experiments often measure downstream attitudes and behavior separately, finding that getting the facts right is necessary but not sufficient to change political opinions.