The spread of true and false news online (Science, 2018)
Analyzing Twitter rumor cascades, false stories diffused farther, faster, deeper, and more broadly than true stories, helping explain why conspiratorial content can go viral.
All topics15 peer-reviewed studies for, 8 against
The best evidence supports a nuanced “yes”: conspiratorial and false narratives can go viral in the sense of forming unusually fast, wide diffusion cascades, but that virality is often localized and heavily concentrated. Large-scale Twitter rumor-cascade analysis shows false stories spread farther and faster than true ones, consistent with the idea that sensational and conspiratorial claims are well-suited to rapid sharing dynamics. Work on misinformation diffusion and echo-chamber structure further suggests that polarization and social reinforcement can amplify these narratives within communities, enabling bursts of virality through complex contagion processes. The strongest counter-evidence is about scope: multiple studies find that exposure and sharing of low-quality/fake sources are highly skewed, with a small minority accounting for a large share of engagement, and that average exposure can be a small slice of people’s overall information diets. Taken together, the evidence implies conspiracies are “viral” in bursty, network-dependent pockets rather than uniformly viral across society. Platform design, recommendation systems, and event salience likely determine whether a conspiracy remains niche or breaks into broader visibility.
Ranked by citation count.
Analyzing Twitter rumor cascades, false stories diffused farther, faster, deeper, and more broadly than true stories, helping explain why conspiratorial content can go viral.
Finds engagement with fake news sources was extremely concentrated—small fractions of users accounted for most exposures and shares—suggesting “virality” is not uniform across the population.
Analyzing Twitter rumor cascades, false stories diffused farther, faster, deeper, and more broadly than true stories, helping explain why conspiratorial content can go viral.
Messages embedding moral emotion see higher retweet rates within ideological networks, indicating content cues that turbocharge spread.
Modeling suggests echo-chamber structure and social reinforcement can enable misinformation to become “viral,” behaving like complex contagion rather than simple contagion.
Study outlines how the pandemic disrupted cognitive and social structures, driving people online where social media radicalizes conspiracy beliefs through rapid spread and stickiness. Real-world conspiracy actions posted back on social media further amplify and reinforce beliefs in a self-reinforcing cycle.
van Prooijen's analysis demonstrates that conspiracy theories function as intergroup phenomena driven by group-oriented motivations—both symbolic identity-based and realistic harm-based. The research shows how conspiracy beliefs facilitate ingroup cohesion and provide urgent signals for group members to prepare for potential outgroup threats, explaining why they spread so effectively through social networks.
Facebook data show polarized communities and diffusion dynamics that allow misinformation (including conspiratorial narratives) to propagate widely within echo chambers.
Tests three motives driving conspiracy theory sharing: belief-bolstering, alarm-sounding, and need for chaos. Using N=3336 US adults, finds motivated sharing and need for chaos strongly predict willingness to share conspiracy theories on social media, identifying key psychological drivers of virality.
Study of 783 participants showing that higher anxiety and lack of control predict adoption of conspiracy theories. Anxiety-provoking crises increase conspiracy belief adoption as psychological sense-making mechanism, relevant to understanding why conspiracies spread during health emergencies.
Analysis of ~126k rumor cascades shows false news—often novel/affect-laden—diffuses more broadly, rapidly, and deeply than truth, especially in politics.
Across multiple datasets, moral-emotional wording remains positively associated with message diffusion, reinforcing earlier findings.
Review argues social uncertainty and low institutional trust heighten susceptibility and sharing of conspiratorial narratives.
This epidemiological network analysis study of 166 Russian towns found that fragmented social network structures significantly amplified misinformation spread, with positive associations between network modularity and knowledge of fake statements about COVID-19, demonstrating how network topology facilitates conspiracy virality.
Applies epidemiological SIR modeling to analyze the 5G-Coronavirus conspiracy theory on Twitter, demonstrating that conspiracy theories spread through social networks with dynamics similar to infectious disease transmission, showing fact-checking is most effective early in diffusion.
Uses nationally representative surveys (N=30,271) to identify key social, political, and economic factors driving conspiracy theory engagement across recognition, belief, and diffusion stages. Shows that early-life economic capital and employment stability significantly influence whether and how individuals spread conspiracy theories.
Bayesian analysis of 15 million tweets from Polish X (Twitter) users shows that the fraction of conspiracy-spreading neighbors in one's immediate network is the strongest predictor of CT dissemination. Even minimal exposure (1-5%) to conspiracy theorists significantly increases sharing likelihood, with single-contact ties proving more influential than repeated contacts.
Finds engagement with fake news sources was extremely concentrated—small fractions of users accounted for most exposures and shares—suggesting “virality” is not uniform across the population.
Simple accuracy prompts reliably decrease willingness to share false headlines, indicating virality is malleable.
Large preregistered studies show robust, generalizable reductions in misinformation sharing via accuracy nudges.
Hybrid field–lab evidence links higher cognitive reflection to less sharing of low-quality/misinfo content on Twitter.
Reports limited reach of fake news sources on Twitter in that context, implying that platform-wide “viral” narratives may be bounded and event-dependent.
Shows visits to untrustworthy sites were a small share of overall media diets on average and concentrated among a subset of people, tempering claims of broad viral reach.
Linking Twitter behavior to voter files shows exposure/sharing extremely concentrated; most users see little fake news.
Recent work identifies outsized contribution from a small group to fake-news diffusion, tempering claims of mass virality.
Research links rapid spread to novel, emotional and moral-emotional content, along with echo chambers where people see the same claims repeated by their network. Having even a small share of conspiracy-spreading contacts has been found to raise the likelihood of sharing.
An analysis of Twitter rumor cascades found false stories diffused farther, faster, deeper and more broadly than true stories. Other studies note that most users see little fake news, because exposure is concentrated among a small minority.
Studies identify a small cadre of supersharers who drive much of false-news spread. Motivated sharing, a need for chaos, anxiety and lack of control are also associated with greater willingness to share conspiracy theories.
Accuracy prompts that shift attention to whether a headline is true have reduced misinformation sharing intentions in large, preregistered studies. Modeling of one COVID-19 conspiracy theory on Twitter suggests fact-checking is most effective early in its spread.