When false information about a company spreads online, managers face an uncomfortable choice. Should they issue a public correction and risk drawing even more attention to the false claim? Or should they stay quiet and hope it fades away?
A survey of 301 managers found that 69% had heard about the possibility that corrections could backfire, and on average, they rated their concern about it at 67 on a 100-point scale. That fear, it turns out, may be misplaced. A series of five experiments involving 4,337 participants found no evidence that fact-checking labels on false brand claims made people believe those claims more. The research, set to appear in the International Journal of Research in Marketing, offers a systematic look at how repetition and correction compete when consumers encounter misinformation about companies.
The fear behind the question
The concern has a name: the familiarity backfire effect. The idea is straightforward. When someone corrects a false claim, they typically have to repeat or reference that claim in the process. This repetition can make the claim feel more familiar. And familiar information, research has shown, tends to feel more true. Psychologists call this the “illusory truth effect,” a well-documented pattern in which people rate statements they have seen before as more accurate than statements they are encountering for the first time, regardless of whether those statements are actually true.
A familiarity backfire effect would occur if the boost in believability from repetition outweighed the reduction in believability from the correction itself. In other words, the correction would do more harm than good. Some earlier studies found hints of this effect, particularly with brief or vague corrections. Others found the opposite. The evidence was mixed, and researchers had called for more direct comparisons of repeated exposure and correction.
Ipek N. Nibat of Sabanci University in Istanbul, along with colleagues at Grenoble Ecole de Management, Kiel University, and Audencia Business School, designed a series of experiments to pull apart these two competing forces and measure them separately. The team used real misinformation about actual companies, sourced from the fact-checking platform Snopes.com, as well as fabricated brands in some conditions.
How the experiments worked
The studies followed a three-stage structure adapted from established research protocols. In the first stage, participants viewed a mix of true and false news headlines about brands, presented without any labels. This served as the “familiarization” phase. In the second stage, participants completed unrelated tasks, such as identifying patterns in abstract paintings, to create a psychological break. In the third stage, participants saw the same headlines again, along with a set of new ones they had not seen before.
At this final assessment stage, the false headlines either carried no label, a “disputed by 3rd party fact-checkers” tag, or a “false based on 3rd party fact-checkers” tag. Participants rated how accurate they believed each headline to be and how they felt about the brand mentioned in it. By comparing responses to headlines participants had already seen (repeated) against headlines they were seeing for the first time (novel), the researchers could measure the repetition effect. By comparing responses across the different labeling conditions, they could measure the correction effect.
The difference between these two effects is what the researchers called the “net effect.” If repetition boosted believability more than correction reduced it, that would count as a backfire. If correction reduced believability more than repetition boosted it, that would count as a reversal, meaning the correction was strong enough to overcome the familiarity boost.
Repetition made false claims feel truer, but correction won out
Across all five studies, the illusory truth effect appeared as expected. Participants consistently rated repeated false claims as more accurate than false claims they were seeing for the first time. The effect sizes ranged from small to medium.
But in every study, the correction effect of the “false” label was larger than the repetition effect. In Study 1, the “false” label reduced belief in misinformation by 0.59 points on a four-point accuracy scale, while the repetition effect only increased belief by 0.15 points. The correction was nearly four times stronger than the damage caused by repetition. Similar patterns appeared across the remaining studies.
The “disputed” label told a different story. It consistently reduced belief in misinformation, but its effect was weaker. In most studies, it only managed to offset the repetition effect, meaning it neutralized the damage without producing any additional benefit. It never backfired, but it also never reversed the repetition effect the way the “false” label did.
Brand perceptions are harder to repair
While corrections moved the needle on whether people believed the false claims, the picture was less encouraging when it came to how people felt about the brands themselves. Across most studies, repetition did not significantly change brand evaluations. Corrections sometimes improved them, but the effects were small and inconsistent.
Study 2 included a baseline condition in which participants evaluated brands without ever being exposed to any misinformation. This allowed the researchers to measure the full impact of encountering false claims. Brand evaluations in every misinformation condition, even those with strong corrections, fell below this baseline. In practical terms, even the best correction did not fully undo the reputational damage from exposure to a false claim. The misinformation left a residual mark on how consumers felt about the brand, even when they no longer believed the claim to be accurate.
This gap between belief and evaluation is consistent with a phenomenon psychologists call the “continued influence effect.” Even after people accept that something is false, the original claim can still color their judgments and decisions. A classic example from earlier research found that consumers rated eating at McDonald’s less positively after hearing (and then rejecting) the false claim that the company’s meat contained worms.
Unknown brands benefit more from corrections
Study 2 also compared well-known brands against fictitious, unknown brands. Repetition increased belief in misinformation for both types, and the effect was slightly stronger for well-known brands. The researchers suggest this may be because consumers have richer mental associations with familiar brands, which can make allegations feel more plausible.
Corrections reduced belief for both brand types, but the downstream effects on brand evaluations differed. For well-known brands, neither the “false” nor the “disputed” label significantly improved evaluations. For unknown brands, however, the “false” label produced a measurable improvement. This suggests that lesser-known brands, which lack the anchoring of strong prior associations, are more responsive to corrective information.
Corrections hold up over time and for new audiences
Study 3 tested a harder scenario. Participants saw the correction label during a second exposure but then encountered the misinformation a third time without any label. They had to rely on memory alone. Even under these conditions, the “false” label still reduced belief more than repetition increased it.
Study 4 flipped the timing. Instead of correcting misinformation after participants had already seen it, the correction appeared during the very first exposure. This simulated what happens when a company’s correction reaches people who have not yet encountered the false claim. Once again, the “false” label reduced belief without backfiring. A follow-up study added a three-day delay between correction and assessment. The results held. Corrections left a durable imprint in memory, even when time had passed.
What this means for companies
The findings point to several actionable takeaways. First, companies worried about making things worse by issuing corrections can take some reassurance from these results. In none of the tested conditions did a correction increase belief in false claims more than the repetition effect alone. Second, the wording of the correction matters. Clear, unambiguous language like “false” consistently outperformed the softer “disputed” label. Third, corrections that reach new audiences, people who have not yet seen the misinformation, also appear safe and effective.
At the same time, companies should temper their expectations. Corrections can reduce how much consumers believe a false claim, but they do not fully restore how consumers feel about the brand. Rebuilding brand perception after a misinformation episode likely requires more than a label. The researchers note that companies do not need to wait for social media platforms to act. Instead, firms can partner with independent fact-checking organizations and communicate corrections directly to their audiences.
The study has some boundaries worth noting. All experiments used self-reported measures of belief and brand evaluation rather than actual purchasing behavior. The misinformation tested involved factual claims that could be verified, not more ambiguous or emotionally charged rumors. And the time delay tested was only three days. Whether these patterns hold over weeks or months, or under conditions of high emotional intensity, remains an open question.
