In 1977, a team of psychologists asked college students to rate how true a series of statements seemed, such as odd trivia that few would know. Students returned for two more sessions, two weeks apart, and rated the list again. Some statements had appeared before and others were new. The repeated ones, true and false alike, tended to seem a little more accurate each time.
That experiment, by Lynn Hasher, David Goldstein and Thomas Toppino, is often cited as the start of research on what is now called the illusory truth effect. It points at something humbling: sometimes a claim feels true simply because it has been heard before.
Most of the time, believing is a sensible default
Nadia Brashier and Elizabeth Marsh proposed in 2020 that people judge truth using three kinds of information. The first is a base rate: most things we hear are true, so accepting them is usually efficient. The second is a feeling, such as how easily a claim is processed. The third is memory: we can compare a claim with what we know and with who told us. The weakness is that the first two are quick and the third takes effort, so it is often skipped.
- You meet a claim. It might be a headline, a rumour or a remark from a friend.
- You meet it again. The second time it is easier to process, because it is familiar.
- Ease is read as a clue. Easy-to-process claims can feel more reliable, even though ease says nothing about accuracy.
- A memory check may or may not happen. Without it, the claim can settle in as something you “just know”.
Later work by Lisa Fazio and colleagues (2015) found that this effect appeared even when people knew the right answer. They called this knowledge neglect: stored knowledge was available, but people did not always use it. The earlier statements were deliberately obscure, so this was a notable extension.
Identity changes how we use our skills
Facts are not only information; they are often signals of who we are. A well-known study by Dan Kahan and colleagues (2017) gave American participants a hard data-interpretation problem. When the numbers were about a skin-rash treatment, people with higher numeracy did better. When the identical numbers were framed as evidence about banning guns, answers split along political lines, and the most numerate participants were more polarised, not less. The authors read this as identity-protective cognition: skill can be used to defend a conclusion as readily as to find one.
This is one study with American participants and a specific task, and the broader literature on motivated reasoning is lively and debated. It is best taken as a pointer, not a law.
Do corrections backfire?
An early and widely repeated idea was the “backfire effect”: correcting someone could entrench the false belief. It sounds alarming, and it influenced many opinion columns. But Thomas Wood and Ethan Porter tested 52 politically charged issues in five experiments with more than 10,000 participants and reported that none of the corrections triggered a backfire. People generally moved toward the facts even when the facts were awkward for their side.
That is good news, with caveats. Moving toward a fact is not the same as changing a vote or an identity. And the continued influence effect shows that a correction does not erase the original claim from reasoning. Corrections tend to help, but they rarely finish the job.
Where the old explanations oversimplify
Popular accounts tend to cite confirmation bias, cognitive dissonance and “education versus truth” as if each were a settled switch. They are real ideas, but the research is more tangled. Education and numeracy do not make people immune, as the Kahan study suggests, but they also do not make belief hopeless. Emotional arousal, trust in a source and repeated exposure all play parts, and their relative weight depends on the claim and the person.
What this means for everyday life
None of this suggests that other people are gullible while you are clear-eyed. These are features of ordinary minds, including yours. The more useful lesson is to build small pauses where memory and source-checking can catch up with the feeling of knowing. Sharing less on impulse, noticing which claims you have mostly heard rather than checked, and looking for who benefits are modest habits. They also do not require you to distrust everything, which would be unworkable and its own form of error.
