Repeat a moderately plausible statement of unknown truth and people rate it likelier to be true than one they have not heard, an effect that even works on statements first labeled false. Most studies used two or three repetitions. DiFonzo and colleagues ran it out much further, with rumors embedded in narratives, and asked what shape the curve takes.
It keeps climbing, but with diminishing returns. Across three experiments with rumors repeated up to six or nine times, validity judgments rose with repetitions: each additional hearing added belief, and each added less than the one before. Self-reported familiarity mediated the relationship, and moderated it only slightly.
The basic effect is robust and unflattering. Moderately plausible statements of unknown truth value, presented repeatedly, get rated as more likely to be true than non-repeated statements. This is the effect. Even statements initially identified as false are later misremembered as true at a higher rate than new ones, with one null finding in the literature. That carries awkward implications for anything that repeats a myth so as to deny it.
What had not been mapped was the shape. Prior work mostly used two or three repetitions, with a few multiple-repetition designs but none going as far as this. Three experiments had participants read narratives in which different rumors appeared between zero and six times, or zero and nine, while validity estimates, and dispositional skepticism were all recorded.
The relationship turned out logarithmic. Belief rose with repetitions, each additional hearing adding less than the last, and incremental repetitions produced incremental increases in validity estimates. The authors read that as support for the processing fluency contrast account, and as evidence that people register fine gradations of fluency rather than a crude familiar or unfamiliar split. Self-reported familiarity statistically mediated the effect.
Skepticism came out weak. It did moderate the effect, making this the only study beyond one dissertation to test the trait. Its influence on belief was small next to repetition. In the face of repeated hearings of a plausible unverified statement, even the skeptic grows more confident. The authors also note that people with stronger overall belief in rumors responded more strongly to repetition.
wat. Each repetition buys a little more belief, and the buying never quite stops. Being a skeptic slows it slightly.
wat wat
Skepticism as a trait barely protected anyone. In the face of repeated hearings even the skeptic grew more confident, which the authors read as the power of the situation over personality.
How do we know?
Three experiments in which participants read narratives containing different rumors repeated zero to six or zero to nine times, with validity estimates, processing fluency measured as rumor familiarity, and dispositional skepticism all recorded.
What is still unknown
The authors measure fluency as self-reported familiarity, which is a proxy rather than a direct read of processing. The statements were moderately plausible rumors, so this does not say what repetition does to the implausible.
Learn something?