The Illusory Truth Effect

What it is

The original design was simple. In the 1977 paper, participants judged whether plausible but obscure statements were true or false and rated their confidence in each judgement. They did this across three sessions spaced two weeks apart. Some statements returned in every session and others appeared only once. Confidence rose for the repeated statements, whether they were true or false, and stayed flat for the statements that appeared once. The authors concluded that how often a statement is encountered is one of the factors that decide whether it seems valid.

Two features of that design have shaped the field since. The statements were ones participants could not easily check, so the effect was shown in the space where prior knowledge does not settle the question. And the comparison was between repeated and unrepeated items, which turns out to matter for how large the effect looks. Dechêne, Stahl, Hansen and Wänke reviewed 51 studies in 2010 and found that the between-items effect, repeated statements set against new ones, was larger than the within-items effect, in which the same statement is rated twice. They read that difference as evidence of two different comparison processes.

The mechanism most often proposed is processing fluency. A repeated statement is easier to process, and ease is read as a sign of truth. Kahneman gives repeated experience as the first of four sources of cognitive ease, and the illusory truth effect is the clearest laboratory form of that source. Reber and Schwarz reported a different route to the same judgement in 1999: statements shown in colours that stood out against white were more often judged true, with no repetition involved.

In effect

The 2026 meta-analysis by Ye, Attali, Ghazi, Cachia, Cassotti and Borst covers 182 studies, 366 effect sizes and 31,184 participants from 1977 to 2025. After correcting for small-study effects, the pooled estimate is small but robust, at 0.37 on the standardised scale with a range of about 0.30 to 0.44, and the studies differ considerably from one another. A correction of that kind is a demanding test for a result that began as a single early paper, and the effect remained after it.

Work on the shape of the effect has added two practical points. Hassan and Barber (2021) reported that gains in rated truth follow a logarithmic pattern, so most of the increase comes early and further repetition adds less. Fazio, Rand and Pennycook (2019) reported that the effect occurs across the range of plausibility, so statements that sound unlikely are not exempt.

Readers meet the effect wherever claims circulate: repeated headlines, familiar slogans, a rumour heard from several people. The laboratory studies use short statements, and the pooled estimate describes that kind of material. How much of the same size carries over to repeated claims in the open is a separate question, and the studies above do not measure it.

What it does not say

The illusory truth effect does not say that repetition makes people believe whatever they are told. The pooled effect is small, the studies vary, and the outcome measured is a rating of how true a statement seems. It is not a measure of what a person goes on to do.

It also does not settle the mechanism. Processing fluency is the explanation most often offered, and the fluency literature has its own record of results that held and one that did not, the reasoning benefit of hard-to-read fonts. The illusory truth effect is a specific, heavily replicated outcome, and fluency is a general proposal about why such outcomes occur. A reader who cites the effect does not have to endorse every claim made in the name of fluency.

Finally, the effect does not say that checking is pointless. The 1977 statements were deliberately obscure. The finding concerns what repetition adds when little else is known, and a statement a person can verify is a different case.

---