“Woman’s Butt Implants Explode While Doing Squats for an Instagram Workout Video.” This headline appeared on a fake-news Web site in July, only to be heedlessly picked up by other media outlets, thereby initiating one of the year’s most cringe-inducing urban legends. What leads a piece of modern folklore to gain such traction?

A number of studies have shown that humans tend to remember certain kinds of information better than others, such as knowledge that might keep us alive or help us find a mate. In one study, subjects were asked to read an urban legend, rewrite it from memory, and then pass their version to the next person (a sequence resembling a game of telephone). At the end of the chain, the legends whose themes could have social or survival-related utility (nudity, spiders, that kind of thing) were recalled most accurately—as evolutionary theory might predict [1].

Many popular myths carry implicit warnings. When researchers analyzed 220 urban legends, they found that the stories were much more likely to mention hazards than benefits. This makes sense: Because believing in a fake hazard is less harmful than failing to believe in a real one, evolutionarily speaking, we should err on the side of being overcredulous about threats. And indeed, the researchers reported, test subjects found statements about topics ranging from German shepherds to Lasik surgery more believable when they mentioned risks, like mauling or double vision [2].

Other research suggests that we may have an incentive not just to believe fearsome legends but to pass them on—sharing information about threats can make you seem more reliable. When people in one study read two descriptions of the same product, one of which mentioned a threat (“If you press control keys during installation, the software may damage your hard disk”), they rated the writer of the threatening description as more competent [3].

One theory of cultural transmission argues that stories, myths, and religious concepts are most likely to endure when they have enough familiar elements to feel plausible, but also have two to three “counterintuitive” elements that make them memorable—a phenomenon known as minimally counterintuitive (MCI) bias. One study analyzed 45 online versions of the ubiquitous Bloody Mary story—say her name to a mirror three times, and she comes out to kill you—and found that the average number of counterintuitive elements was 2.36 [4].

Knowing that an urban legend isn’t true won’t necessarily inoculate you against its virality. A recent study, to be published in the Journal of Experimental Psychology: General, found that reading a false statement (say, “The Atlantic Ocean is the largest ocean on Earth”) made people more likely to rate the statement as true when they encountered it a second time—even if they were told on both readings that it might be false, and even if they later demonstrated that they knew the Pacific was in fact the largest ocean. Exposure breeds familiarity, which fosters credulity—even when you know better [5]. Which is to say, stories about exploding implants might be with us for a while.

[1] Stubbersfield et al., “Serial Killers, Spiders and Cybersex” (British Journal of Psychology, May 2015)

[2] Fessler et al., “Negatively-Biased Credulity and the Cultural Evolution of Beliefs” (PLOS One, April 2014)

[3] Boyer and Parren, “Threat-Related Information Suggests Competence” (PLOS One, June 2015)

[4] Stubbersfield and Tehrani, “Expect the Unexpected? Testing for Minimally Counterintuitive (MCI) Bias in the Transmission of Contemporary Legends” (Social Science Computer Review, Feb. 2013)

[5] Fazio et al., “Knowledge Does Not Protect Against Illusory Truth” (Journal of Experimental Psychology: General, forthcoming)