Why Simple Science Explanations Spread Faster Online

"Sugar causes hyperactivity in kids" has circulated as accepted parental wisdom for decades, despite controlled trials, including a widely cited 1995 meta-analysis in JAMA, finding no significant behavioral effect once studies blinded parents to which children had actually consumed sugar (Wolraich, Wilson & White, JAMA, 1995). The myth survives not because the evidence is ambiguous but because it's a simpler, more shareable claim than the messier truth involving parental expectation. This article explains the cognitive science behind why simple, wrong explanations routinely outrun accurate, complicated ones online, and what that means for anyone trying to communicate real findings.

Fluency, not accuracy, drives the instinct to share

Cognitive psychologists call this processing fluency: the ease with which the brain understands, retrieves, and mentally rehearses a piece of information. Fluency is not a side detail of comprehension, it is itself an input into judgment. In a foundational 1998 study, Rolf Reber, Piotr Winkielman, and Norbert Schwarz showed that objects presented in ways that made them easier to perceive, such as being preceded by a matching visual prime, were rated as more pleasant by participants who had no idea why (Reber, Winkielman & Schwarz, Psychological Science, 1998). The brain doesn't cleanly separate "this feels easy to process" from "this feels good and true." It uses the first as a shortcut for the second.

That shortcut extends directly to belief. A claim with one clear cause, one clear effect, and no stated exceptions is processed faster than a claim hedged with confidence intervals, sample limitations, and competing explanations. That processing speed alone makes the simple version feel more trustworthy, independent of which version is actually correct. This is the mechanism, not a metaphor for it: fluency is a real, measurable variable in cognitive psychology experiments, and it correlates with judgments of truth even when the actual accuracy of the claim is held constant.

This creates a systematic bias in what spreads online: not the best-supported claim, but the most fluently processed one, and the two properties are only loosely correlated. A finding can be extremely well-supported by evidence and still lose the sharing contest to a simpler, less accurate claim that's easier to say out loud, remember, and repeat from someone else's memory a week later. Fluency doesn't just affect the first read, it compounds every time the claim gets retold, because each retelling reinforces whichever version was easiest to encode.

A large-scale 2018 study analyzing roughly 126,000 news stories on Twitter, published in *Science* by Vosoughi, Roy, and Aral, found that false stories were about 70 percent more likely to be retweeted than true stories and reached their first 1,500 people roughly six times faster. The researchers attributed much of the effect to false claims tending to be more novel and emotionally striking, properties that overlap heavily with the simplicity that makes a claim easy to process and repeat (Vosoughi, Roy & Aral, *Science*, 2018).

The illusory truth effect: repetition manufactures belief

Fluency doesn't only come from a claim's structure, it also comes from how many times a person has already encountered it. This is documented separately as the illusory truth effect: the tendency to rate a statement as more accurate simply because it has been seen or heard before, regardless of whether it's actually true. The effect was first established in a 1977 study by Lynn Hasher, David Goldstein, and Thomas Toppino, who had participants rate the accuracy of statements across three sessions spaced two weeks apart (Hasher, Goldstein & Toppino, Journal of Verbal Learning and Verbal Behavior, 1977). Statements repeated across sessions were rated as more accurate than statements seen only once, even though repetition carried no actual evidence about truth.

The effect has been replicated dozens of times since, including work by Lisa Fazio and colleagues showing that even people who possess the correct knowledge to identify a statement as false will still rate it as feeling more true after repeated exposure (Fazio et al., Journal of Experimental Psychology: General, 2015). Knowing better does not fully immunize a person against the fluency created by repetition. This matters directly for how simple claims dominate online: every retweet, every screenshot, every "did you know" caption is another repetition, and each repetition raises the felt truth of the claim slightly, independent of its actual accuracy.

Why does repetition feel like evidence?

Repetition feels like evidence because the brain conflates ease of retrieval with likelihood of truth. A statement heard three times is retrieved more smoothly than a statement heard once, and that smoothness registers as familiarity, which the mind misreads as corroboration, even when the three exposures all trace back to the same unverified source.

Online sharing turns this into a feedback loop. A simple claim gets repeated because it's fluent, and each repetition makes it feel truer, which makes it more likely to be repeated again. A complicated, accurate claim rarely gets the volume of repetitions needed to build that same felt-truth momentum, because each individual share requires more cognitive effort from the person passing it along.

Simplification strips the very details that would make a claim correct

A scientifically accurate claim about sugar and behavior requires several qualifying clauses: the effect wasn't found in properly blinded studies, and parental belief that a child had consumed sugar, regardless of whether they actually had, was itself a stronger predictor of perceived hyperactivity than the sugar itself. Every one of those clauses adds friction to sharing, since a caveat-heavy claim takes longer to state and doesn't compress into a tweet or a caption as cleanly. The version that survives online isn't the one supported by controlled trials, it's the one that fits in a single confident sentence.

This is a structural problem, not a laziness problem. Even careful communicators face a hard tradeoff between completeness and shareability, because every additional qualifying clause reduces fluency by adding retrieval load. A reader has to hold more pieces in working memory to repeat a hedged claim accurately, and working memory is a genuinely limited resource, not a matter of attention span or intelligence.

Does this mean accurate science communication should abandon simplicity?

No, but it means the simplification needs to preserve the mechanism rather than just the conclusion. "Studies find no link between sugar and hyperactivity once parents don't know which kids had sugar" is nearly as short as the myth and still fluent, but it keeps the one detail, parental expectation, that actually explains why the myth felt true to generations of parents in the first place.

The distinction that matters is between simplifying a claim's language and simplifying away its mechanism. Language can always be tightened without losing accuracy. Mechanism is what gets sacrificed when a writer optimizes purely for shareability, because the mechanism is usually the part carrying the most qualifying detail, and qualifying detail is exactly what fluency research says gets discarded first.

Social usefulness matters as much as truth

People share information partly to signal group belonging, competence, or shared values, not purely to transmit accurate facts. Marketing researcher Jonah Berger formalized this in his STEPPS framework, published in Contagious: Why Things Catch On (2013), identifying six drivers of what gets shared: social currency, triggers, emotion, public visibility, practical value, and stories. Two of these map directly onto why simple claims outperform accurate ones online.

Social currency describes how sharing information makes the sharer look good, informed, or interesting to their audience, and a simple claim is far easier to deploy socially than a hedged one, because the recipient can immediately understand and reuse it in conversation. A claim requiring the listener to hold multiple qualifications in mind is less socially rewarding to share, even when it's the more defensible one, since a mangled retelling of a nuanced claim makes the sharer look confused rather than competent. Practical value, Berger's "news you can use" driver, also rewards simplicity, because actionable advice needs to be stated plainly to be actionable at all. "Sugar makes kids hyper, so skip the birthday cake before nap time" is directly usable. "The apparent behavioral effect of sugar consumption disappears in properly blinded study conditions" is not something a parent can act on in the moment, even though it's the accurate one.

Berger's framework is a business and communication model built from his own research and consulting work, not a peer-reviewed cognitive psychology finding in the same sense as fluency or illusory truth studies. It's useful here as a structured account of *why* people share, layered on top of the fluency research that explains *what* people are more willing to believe in the first place.

Emotional arousal accelerates the same bias

Berger's STEPPS framework and the Vosoughi, Roy, and Aral Science study converge on a second variable beyond simplicity: emotional arousal. Berger's research emphasizes that high-arousal emotions, whether positive like awe and amusement or negative like anger and anxiety, physiologically prime people to share, while low-arousal emotions like sadness or contentment suppress sharing behavior. The 2018 Twitter analysis found a parallel pattern in the specific words used in replies to true versus false stories: replies to false stories were more likely to express surprise and disgust, while replies to true stories were more likely to express sadness.

Simple claims and emotionally arousing claims tend to travel together, not because one causes the other, but because both properties strip away hedging. A claim that provokes immediate outrage or surprise usually cannot survive being loaded with caveats, since qualification softens the emotional hit that makes a claim feel worth reacting to in the first place. This is why the most viral misinformation online tends to be both simple and emotionally charged at once. It's not two separate failures, it's the same underlying compression happening on two axes simultaneously.

Is emotional content inherently less accurate than neutral content?

No. Emotional framing and factual accuracy are independent variables, and a well-sourced finding can be presented with genuine emotional stakes without sacrificing precision. The risk isn't emotion itself, it's that emotional urgency creates pressure to cut the same qualifying detail that fluency pressure cuts, so the two forces often push a claim toward inaccuracy in the same direction at the same time.

The two-system architecture that makes fluency so persuasive

Fluency's outsized influence on belief isn't a quirk, it's a predictable consequence of how human cognition is built. Psychologist Daniel Kahneman, in his 2011 book Thinking, Fast and Slow, popularized a two-system framework for describing the mind's operations, drawing on decades of research he conducted with Amos Tversky. System 1 is fast, automatic, and runs constantly in the background, producing impressions, gut reactions, and snap judgments without any sense of deliberate effort. System 2 is slow, effortful, and has to be deliberately engaged, handling tasks like checking a citation, comparing two competing explanations, or working through a statistical claim step by step (Kahneman, Thinking, Fast and Slow, 2011).

The two systems are not equally available at every moment. System 2 is metabolically expensive and easily depleted, a phenomenon Kahneman and other researchers describe as limited cognitive capacity: after sustained effortful thinking, people become more likely to accept System 1's snap judgment rather than engage System 2 again. Scrolling a social feed is close to a worst-case environment for System 2 engagement, since posts arrive continuously, each one competing for a sliver of attention before the next appears. Under those conditions, a claim gets whatever judgment System 1 can produce in a fraction of a second, or it gets skipped entirely.

This is precisely the terrain where processing fluency does its work. Kahneman explicitly linked fluency to System 1, describing it as one of the cues System 1 uses to generate a feeling of cognitive ease, which in turn gets misread as a signal of validity, familiarity, or truth. A claim with one cause and one effect can be fully evaluated by System 1 in the time it takes to read it. A claim with confidence intervals, sample caveats, and competing explanations requires System 2, and System 2 rarely gets invoked for a passing post in a feed. The accurate, hedged version of a claim isn't losing on merit, it's losing because it demands a cognitive resource that the scrolling environment structurally starves.

Kahneman and Tversky's broader research program on heuristics and biases, including the fluency-related work Kahneman synthesized in *Thinking, Fast and Slow*, contributed to Kahneman receiving the 2002 Nobel Memorial Prize in Economic Sciences, the first psychologist to win the award, for integrating psychological insight into economic decision-making (Nobel Prize press release, Royal Swedish Academy of Sciences, 2002).

Why doesn't System 2 just catch fluency-based errors?

System 2 can catch them, but only when it's actively engaged, and engaging it requires either high stakes, unusual effort, or a deliberate prompt to slow down. Kahneman's research found that System 2 tends to defer to System 1's output by default because verifying every impression would be exhausting, so a feed environment optimized for rapid, low-effort scrolling produces exactly the conditions under which fluent, unverified claims pass through unchecked.

How does this connect back to why simple claims win?

Simplicity is what makes a claim processable by System 1 alone. A one-clause claim requires no System 2 engagement to feel true, while a hedged, multi-clause claim can only be fully evaluated by System 2, which the scrolling environment rarely activates. The System 1/System 2 framework doesn't add a new cause to the fluency story, it explains the cognitive machinery that makes fluency such a reliable predictor of what spreads.

Headlines carry the fluency effect before the body copy even loads

Everything above describes how fluency operates once someone is reading a claim, but headlines face a version of the same test before a reader commits to reading anything at all. A headline is evaluated by System 1 in well under a second, deciding whether the post is worth a click, a share, or neither, and that decision is made almost entirely on the headline's own fluency rather than on the accuracy of what follows.

This creates a second, earlier bottleneck on top of the body-content dynamics already described. A nuanced, accurate headline that tries to preserve a study's actual finding, complete with its qualifications, competes directly against a simple, declarative headline making the same general claim without the caveats. Newsroom and social-platform experience with headline testing has generally found that shorter, more direct headlines outperform longer, hedged ones on click-through and share rate, a pattern consistent with the fluency mechanism described above even where a specific controlled study isn't cited here. The caveat that would make a headline accurate is frequently the same caveat that makes it slower to read, and the two costs are paid together.

The practical consequence for anyone communicating science is that the headline is not a separate problem from the fluency dynamics inside the article, it's the same dynamics under tighter constraints. A hedge that fits comfortably in a 300-word body paragraph often can't survive in a ten-word headline, which is why so much science miscommunication traces back not to the body text but to a headline written by someone optimizing for clicks who trimmed the one clause that made the underlying claim correct.

What predicts viral spread: a mechanism comparison

Not every property that predicts sharing behaves the same way, and treating them as interchangeable leads to bad diagnoses of why a specific claim went viral. The table below separates four commonly cited drivers by the mechanism through which each one operates and by whether that mechanism has any built-in relationship to accuracy.

Characteristic	Mechanism	Relationship to accuracy	Key source
Processing fluency	Ease of comprehension registers as a truth signal	None — fluency is orthogonal to accuracy	Reber, Winkielman & Schwarz, Psychological Science, 1998
Repetition exposure	Familiarity from prior exposure is misread as corroboration	None — repetition can entrench a false claim as easily as a true one	Hasher, Goldstein & Toppino, Journal of Verbal Learning and Verbal Behavior, 1977
Emotional arousal	High-arousal states physiologically prime sharing behavior	None — both awe and outrage can attach to true or false claims	Berger, Contagious, 2013
Novelty/surprise	Unexpected claims draw more attention and are shared to inform others	Weak inverse — well-established truths are rarely "novel"	Vosoughi, Roy & Aral, Science, 2018

None of the four strongest predictors of viral spread has a built-in bias toward accuracy, which is the core reason simple, wrong claims can outcompete accurate, complicated ones at scale.

What this means for people writing or evaluating science claims

For anyone writing science content, the practical response isn't to add more hedges, since piling on qualifiers just increases retrieval load and guarantees the accurate version loses the fluency contest even worse. The better response is to identify the single explanatory detail that makes a claim correct, the parental-expectation clause in the sugar example, and build the simplification around preserving that one detail rather than removing it for brevity. A claim can be short and accurate at the same time. It just requires choosing which detail is load-bearing before cutting anything.

For anyone reading and evaluating claims that spread quickly, the more useful heuristic is to treat fluency and repetition as evidence of shareability, not evidence of truth. A claim that has been repeated many times and feels instantly obvious has met the bar for what travels well online. That bar has nothing to do with whether the claim survived a properly controlled study. The two filters are independent, and mistaking passage through one for passage through the other is exactly the error that keeps decades-old myths like the sugar-hyperactivity claim circulating well past their evidentiary shelf life.

The takeaway

Simple science explanations spread faster online because human cognition uses processing fluency, not evidentiary strength, as a fast proxy for truth, and repeated exposure compounds that effect independent of accuracy. Berger's STEPPS research adds that emotional arousal and social usefulness reward the same kind of compression that fluency rewards, which is why simple and viral so often travel together. The fix isn't making science communication more complicated, it's identifying the one detail that makes a claim actually correct and refusing to cut that detail even when everything else gets trimmed for length.

---

Read next:

• Why headlines overstate breakthrough science
• How to spot AI-written science fluff
• Why the best explainers start with the answer

FAQ

Q: Why does the "sugar causes hyperactivity" myth persist despite contrary evidence?

Controlled, blinded studies, including a widely cited 1995 JAMA meta-analysis, found no significant effect of sugar on children's behavior once parental expectation was accounted for. The myth persists because the simple claim is easier to process and repeat than the more accurate explanation involving parental bias.

Q: What is processing fluency, and how does it affect what people believe?

Processing fluency describes how easily a piece of information is mentally understood and rehearsed. A 1998 study by Reber, Winkielman, and Schwarz found that fluency increases how positively and truthfully a statement feels to people, independent of whether the statement is actually accurate, which biases sharing toward simpler claims.

Q: What is the illusory truth effect?

The illusory truth effect is the tendency to rate a statement as more accurate simply because it has been encountered before, regardless of its actual truth. First documented by Hasher, Goldstein, and Toppino in 1977 and replicated many times since, it means repetition alone can manufacture a feeling of truth.

Q: Did research find that false claims spread faster than true ones online?

Yes. A 2018 study published in Science analyzing roughly 126,000 Twitter news stories found false stories were about 70 percent more likely to be retweeted than true stories and spread significantly faster, an effect researchers linked partly to false claims being more novel and emotionally striking.

Q: What is Jonah Berger's STEPPS framework?

STEPPS is a framework from Jonah Berger's 2013 book Contagious describing six drivers of shareability: social currency, triggers, emotion, public visibility, practical value, and stories. It's a communication and marketing model, not a peer-reviewed accuracy study, but it explains why simple, emotionally resonant, and immediately usable claims tend to spread further than hedged ones.

Q: Can a scientifically accurate explanation still be simple enough to spread well?

Yes, as long as the simplification preserves the core mechanism rather than just the conclusion. A short claim that keeps the key explanatory detail, like parental expectation in the sugar-hyperactivity example, can be nearly as shareable as the inaccurate myth while remaining correct.

Q: What is the System 1/System 2 framework, and why does it matter for misinformation?

System 1 and System 2 are Daniel Kahneman's terms, from his 2011 book Thinking, Fast and Slow, for the mind's two modes of thought: System 1 is fast, automatic, and effortless, while System 2 is slow, deliberate, and requires conscious effort. Fluent, simple claims can be fully processed by System 1 alone, while accurate, hedged claims require System 2, which social-feed environments rarely activate, giving simple claims a structural advantage in belief and spread.

Q: Does headline wording affect how far a science claim spreads, separately from the article body?

Yes. A headline is judged by System 1 in under a second, before a reader engages with any supporting detail, so a short, declarative headline tends to outperform a longer, hedged one on clicks and shares. Because qualifying clauses are often the first thing cut to shorten a headline, headline-level simplification can strip the exact detail that made the underlying claim accurate.