Disinformation

Abstract

Disinformation is a form of deception: false or misleading information that is deliberately created and spread in order to mislead. This article treats it as a problem in cognition—why false claims are believed, why corrections so often fail, and what reduces susceptibility. It develops the main threads: the continued influence effect, in which retracted misinformation keeps shaping inference; the illusory truth effect, in which mere repetition raises perceived accuracy even for implausible or known-false claims; the account of fake-news susceptibility as a failure of analytic thinking rather than of partisan motivation; the empirical shape of how false news spreads and how prevalent it actually is; and the two constructive responses—careful debunking and inoculation, or prebunking, which confers resistance before exposure. A worked example and three interactive demonstrations make each manipulable.

Keywords: disinformation, deception, misinformation

Disinformation is among the most consequential applications of deception, and one of the hardest to counter. Where an individual lie targets one listener, disinformation is false information produced and distributed to move the beliefs of many—about elections, vaccines, climate, or an unfolding crisis—and the digital information environment has made its production cheap and its spread fast. The scientific question is not merely how to fact-check a given claim but why false claims take hold in the first place, why telling people the truth so often fails to dislodge them, and what, if anything, builds durable resistance.

What cognitive and behavioural science contributes is a shift from treating the problem as one of bad actors and gullible audiences to one of ordinary memory and reasoning operating as they normally do. The same processes that make the mind efficient—relying on familiarity as a cue to truth, building a coherent mental model and resisting gaps in it, and conserving effort rather than scrutinising every claim—are the processes disinformation exploits. The sections below take these in turn—the continued influence of corrected misinformation, the illusory truth of repetition, the role of analytic thinking, the spread and prevalence of false news, and the correction and inoculation responses—after first placing disinformation within the broader category of deception under which it is classified.

What Disinformation Is

Disinformation is false or misleading information that is deliberately created and disseminated in order to deceive. Two distinctions fix the concept. The first is intent: disinformation is spread with the knowledge or purpose that it is false, whereas misinformation is the broader category of inaccurate information regardless of intent—the person sharing it may sincerely believe it. The same false claim can be disinformation in the mouth of its author and misinformation in the feed of someone who innocently passes it on. The second is scale and target: where an ordinary lie aims to instill a false belief in a particular person, disinformation is typically produced for distribution—to influence public opinion, discredit an opponent, or pollute the information environment—so that its defining relation holds between a source and an audience.

Because the sharer’s intent is often unknowable, the cognitive literature frequently studies misinformation as the superordinate phenomenon—how false claims are believed, remembered, and corrected—and treats disinformation as the deliberately produced subset whose psychology of reception is the same (Ecker et al., 2022). This article follows that usage: the mechanisms of belief and correction described below apply to false information however it originated, while the term disinformation marks the deliberate, usually organised, production that places it squarely within deception.

In the Medical Subject Headings vocabulary, Disinformation (D000087862) is a descriptor filed as a narrower heading beneath Deception (D008192), itself a form of social behavior. The placement encodes a substantive claim: disinformation is deception raised to the level of publics and institutions, sharing the parent construct’s defining feature—the deliberate instilling of a belief known to be false—while adding organised production and mass distribution. The remainder of this article takes the cognition of disinformation in turn: the continued influence of corrected claims, the illusory truth of repetition, the role of analytic thinking, the spread and prevalence of false news, and the correction and inoculation responses.

Perspectives on Disinformation

The science of disinformation is organised less by one grand theory than by a set of perspectives, each tied to a distinct question. The first asks why corrections fail and answers with the continued influence effect. A second asks why repetition works and answers with the illusory truth effect. A third asks who falls for false claims and why and locates susceptibility in a lack of analytic thinking rather than in partisan motivation. A fourth asks how false information actually spreads and how common it is, replacing anecdote with measurement. A fifth asks what can be done and distinguishes debunking after the fact from inoculation before it. These perspectives are complementary rather than rival: a single false claim may spread widely because it is novel, be believed because it is repeated, resist correction because of continued influence, and be blunted—if at all—only by reasoning or by prior inoculation. Table 1 summarises them, and the sections that follow develop each in turn.

Table 1

Perspectives on Disinformation

The continued influence effect

The most stubborn fact about misinformation is that correcting it does not undo it. In the founding demonstration, Hollyn Johnson and Colleen Seifert had people read a report of a warehouse fire in which an early message implied that negligently stored paint and gas cans had caused the blaze. When a later message explicitly retracted that detail—stating the closet had in fact been empty—readers accepted the correction yet still drew on the discredited cause when asked to explain the fire, invoking the paint and gas in their inferences (Johnson & Seifert, 1994). The retracted information continued to influence reasoning even though participants could report that it had been retracted. This is the continued influence effect: a correction updates what people can recite but not the mental model they reason from.

Stephan Lewandowsky and colleagues synthesised the mechanisms in an influential review (Lewandowsky et al., 2012). People build a coherent causal model of an event, and a retraction leaves a gap that the discredited detail had filled; unless the correction supplies a plausible alternative to take its place, the mind prefers the complete-but-wrong model to the accurate-but-incomplete one. Corrections are also undercut by familiarity—repeating the myth in order to debunk it can itself strengthen it—and by the effort of retrieving the correction rather than the more fluent original. The practical lesson, developed in a later review (Ecker et al., 2022), is that effective corrections are explicit, repeated, and provide a causal alternative rather than a mere negation. Demo 1 lets the reader set how strongly a retraction reduces reliance and read off how much of the original claim’s influence persists.

The illusory truth effect

A second mechanism needs no deception at the point of reception at all: it is enough to repeat a claim. The illusory truth effect is the finding that statements encountered more than once are judged more likely to be true than novel statements, because prior exposure makes them easier to process and that fluency is misread as a signal of accuracy. Lisa Fazio and colleagues showed how deep the effect runs: repetition increased belief even for statements that contradicted participants’ own stored knowledge, so that knowing the correct answer did not protect people from the pull of familiarity (Fazio et al., 2015). Truth, in the moment of judgment, is inferred partly from how fluent a claim feels, and repetition manufactures fluency.

The effect transfers directly to disinformation. Gordon Pennycook, Tyler Cannon, and David Rand found that a single prior exposure to a fake-news headline raised its perceived accuracy on a later encounter—the effect held even for headlines that were implausible, even when the headlines were accompanied by a fact-check warning, and even when they ran against the participant’s political stance (Pennycook, Cannon & Rand, 2018). A claim does not have to be believed on first sight to gain ground; it only has to be seen, because each repetition buys familiarity that later masquerades as truth. Demo 2 lets the reader repeat a set of claims and watch perceived truth climb with exposure.

Analytic thinking and why people fall for fake news

Why are some people more taken in than others? A natural hypothesis is motivated reasoning: people believe fake news that flatters their politics because they are motivated to, and reasoning ability only helps them rationalise what they already want to believe. Pennycook and Rand tested this against a simpler alternative and found the simpler one better supported. Across studies, the ability to tell true headlines from false ones was predicted by performance on the Cognitive Reflection Test—a measure of the disposition to override an intuitive-but-wrong answer with deliberate thought—and more reflective reasoners were better at discernment regardless of whether a headline aligned with their politics (Pennycook & Rand, 2019). Falling for fake news looked like “lazy, not biased” cognition: a failure to stop and think, not a triumph of partisan motivation over reason.

The broader review of this programme frames susceptibility as the normal operation of an efficient mind that leans on intuition and familiarity and scrutinises little of what it sees (Pennycook & Rand, 2021). The account is consequential for intervention: if the problem is insufficient reflection rather than fixed partisanship, then prompts that merely redirect attention toward accuracy should help—a prediction the correction-and-inoculation section takes up. It also tempers pessimism, because a disposition to reflect is more malleable than a political identity.

How false news spreads and how prevalent it is

Measurement has replaced anecdote about the reach of false information, and the findings cut two ways. Soroush Vosoughi, Deb Roy, and Sinan Aral analysed the diffusion of verified true and false stories on Twitter over more than a decade and found that falsehood spread significantly farther, faster, deeper, and more broadly than the truth across every category of information, and most markedly for political news. Falsehood’s advantage was not driven by bots, which spread true and false stories at similar rates, but by people: false stories were more novel and evoked more surprise and disgust, and novelty is what people pass on (Vosoughi, Roy & Aral, 2018). The structure of human attention, not merely automated amplification, gives false news its edge.

Yet the share of most people’s information diet that is outright fake is smaller than the alarm suggests. Andrew Guess, Jonathan Nagler, and Joshua Tucker matched survey respondents to their Facebook sharing during the 2016 US campaign and found that sharing fake-news articles was rare—the vast majority of users shared none—and heavily concentrated: the strongest predictor of sharing was age, with users over 65 sharing nearly seven times as many fake-news articles as the youngest group, independent of education and ideology (Guess et al., 2019). The two findings are complementary: false news enjoys a structural advantage in how far it travels, but its production and spread are concentrated in a small, identifiable minority—which is precisely the condition under which targeted interventions can matter.

Correction and inoculation

If corrections after the fact are weakened by continued influence, one response is to act before exposure. Inoculation theory borrows the logic of vaccination: a weakened dose of a persuasive attack, together with a forewarning and a refutation, builds resistance to the full attack later. Sander van der Linden and colleagues showed that pre-emptively exposing people to the techniques of climate-change misinformation—here, the false-consensus tactic of a bogus petition—protected their accurate beliefs against a subsequent dose of that misinformation, neutralising much of its effect (van der Linden et al., 2017). Inoculation shifts the target from the specific false claim to the manipulation technique, so that one “dose” can generalise across many falsehoods built from the same trick.

Jon Roozenbeek and van der Linden scaled the idea into an active intervention: a browser game, Bad News, in which players take the role of a disinformation producer and learn its tactics—impersonation, emotional manipulation, conspiracy, trolling—from the inside. Playing the game reduced players’ rated reliability of manipulative content afterward, a prebunking effect that held across the techniques taught (Roozenbeek & van der Linden, 2019). A lighter-touch intervention targets the moment of sharing rather than belief: Pennycook and colleagues found that people often share misinformation not because they believe it but because the social-media context distracts them from accuracy, and that a simple nudge drawing attention to accuracy—asking people to rate the accuracy of a single headline—improved the quality of what they subsequently shared (Pennycook et al., 2020). Demo 3 lets the reader apply inoculation and watch the perceived reliability of manipulative content fall.

Worked Example

The mechanisms become concrete in three small exercises, each matching one of the demonstrations. The indices below are illustrative—they capture the qualitative shape of published findings rather than reproducing any single reported statistic.

Continued influence (how much of a correction fails). Suppose reliance on a discredited cause is scored on a 0–10 scale of how often people invoke it in their explanations. A group never told the misinformation scores a baseline of 2 (some invoke the idea spontaneously). A group told the misinformation but never corrected scores 6. A group told the misinformation and then given a retraction scores 4. The correction removes 6 − 4 = 2 points of the 6 − 2 = 4 points the misinformation added, so it undoes 2 ÷ 4 = 50% of the effect and leaves 50% as continued influence. A bare negation often clears less; supplying a causal alternative clears more. Demo 1 lets the reader set the correction’s strength and read off the persisting share.

Illusory truth (repetition as evidence). Represent belief as the probability of judging a statement “true.” Suppose a novel claim is called true with probability 0.40, and each prior exposure adds a fluency increment of 0.06, capped where fluency saturates. After one repetition the claim is judged true with probability 0.40 + 0.06 = 0.46; after three, 0.40 + 3 × 0.06 = 0.58. Repetition has moved a minority-believed claim into majority-believed territory without adding one shred of evidence—the increase is pure familiarity. Demo 2 lets the reader vary the increment and the number of exposures.

Inoculation (resistance before exposure). Suppose, without any intervention, a manipulative headline is rated reliable by a proportion 0.60 of viewers. An inoculation that teaches the underlying technique reduces that by a protection factor; take it to cut perceived reliability by 0.25 in relative terms. The inoculated proportion is 0.60 × (1 − 0.25) = 0.45, an absolute drop of 0.15. Because the inoculation targets the technique rather than the single claim, the same 25% relative protection applies to other headlines built on the same trick—which is what lets one dose generalise. Demo 3 lets the reader set the baseline and the protection factor and watch the reliability rating fall.

Discussion

Read across its perspectives, the science of disinformation tells a coherent story about an ordinary mind in a hostile information environment. False claims spread because novelty and emotion drive human sharing (Vosoughi, Roy & Aral, 2018); they are believed because repetition manufactures the fluency the mind reads as truth (Fazio et al., 2015; Pennycook, Cannon & Rand, 2018); they take hold more readily in people who reason less reflectively, rather than in people of one political stripe (Pennycook & Rand, 2019, 2021); and once lodged, they resist correction because the mind prefers a complete causal model to an accurate but gap-ridden one (Johnson & Seifert, 1994; Lewandowsky et al., 2012). None of these mechanisms requires an unusually credulous audience; each is a side effect of normally adaptive cognition.

The constructive implications follow from taking those mechanisms seriously. Because correction is weakened by continued influence, it must be done well—explicit, repeated, and furnished with a causal alternative (Ecker et al., 2022)—or, better, pre-empted: inoculation builds resistance by exposing the manipulation technique before the falsehood arrives, and generalises across claims that share the trick (van der Linden et al., 2017; Roozenbeek & van der Linden, 2019). Because much sharing reflects inattention to accuracy rather than genuine belief, lightweight accuracy prompts can improve what circulates (Pennycook et al., 2020). And because both the production of fake news and its sharing are concentrated in small minorities (Guess et al., 2019), targeted interventions are more tractable than the scale of the problem first suggests. The honest summary is that no single fix dissolves disinformation, but a layered response—slowing spread, prompting accuracy, inoculating in advance, and correcting carefully when prevention fails—engages each mechanism at the point where it operates.

Current Directions

The most active applied front is scaling inoculation. Where the earliest demonstrations used bespoke messages or a browser game, recent work tests short “prebunking” videos and technique-based messages that can be delivered through ordinary advertising slots to millions, and asks how long the conferred resistance lasts and how often it must be renewed (Roozenbeek & van der Linden, 2019; van der Linden et al., 2017). A second front probes the boundary conditions of the accuracy-nudge effect—how large it is in the field as opposed to the lab, for whom it works, and whether it fades with repeated exposure—as part of a broader reckoning with which interventions survive contact with real platforms (Pennycook et al., 2020; Pennycook & Rand, 2021). A third extends correction research to generative artificial intelligence, which lowers the cost of producing fluent, individually tailored falsehoods and so stresses every mechanism at once—spread, repetition, and the plausibility that feeds continued influence (Ecker et al., 2022). The common thread is a move from demonstrating mechanisms in the lab to engineering durable, scalable defences for the environments where disinformation actually circulates.

Common Misconceptions

Correcting a false claim undoes its effect.

It usually does not. The continued influence effect means a retraction updates what people can recite but not the mental model they reason from; the discredited detail keeps shaping inference unless the correction is explicit, repeated, and supplies a plausible alternative cause (Johnson & Seifert, 1994; Lewandowsky et al., 2012).

Only the ignorant or the uneducated fall for fake news.

Susceptibility tracks a failure to stop and think—low analytic reasoning—far better than it tracks education or intelligence, and repetition raises belief even in people who know the correct answer. Discernment improves with reflection regardless of political alignment (Fazio et al., 2015; Pennycook & Rand, 2019).

People share misinformation because they believe it.

Often they share it without having stopped to consider whether it is true at all. The social-media context diverts attention from accuracy, and a simple prompt to think about accuracy improves the quality of what people subsequently share (Pennycook et al., 2020).

Bots are the main reason false news goes viral.

Automated accounts spread true and false stories at similar rates; the asymmetry in how far falsehood travels is driven by people, who pass on false stories more because they are more novel and emotionally provocative (Vosoughi, Roy & Aral, 2018).

Glossary

Accuracy nudge.

A lightweight intervention that briefly directs attention to the concept of accuracy—for example, asking a person to rate one headline’s truth—which improves the quality of what they subsequently share without changing their beliefs.

Analytic thinking.

The disposition to override an intuitive but possibly wrong response with deliberate reasoning, often indexed by the Cognitive Reflection Test. Higher analytic thinking predicts better discernment of true from false news.

Cognitive Reflection Test (CRT).

A short set of problems whose intuitive answers are wrong, used to measure the disposition to override a first impulse with deliberate reasoning; higher scores predict better discernment of true from false news.

Continued influence effect.

The tendency for misinformation to keep affecting beliefs and inferences even after it has been clearly retracted and the retraction remembered.

Debunking.

Correcting a false claim after exposure. Effective debunking is explicit, repeated, and supplies a causal alternative rather than a bare negation, which otherwise leaves a gap the myth refills.

Deception.

The deliberate attempt to create in another a belief the deceiver holds to be false; the MeSH parent category under which disinformation is filed.

Disinformation.

False or misleading information deliberately created and spread to deceive, typically at scale; in MeSH a narrower descriptor beneath deception.

Fake news.

Fabricated information presented in the format of a news report, lacking the editorial norms of legitimate journalism; a common vehicle for disinformation.

Fluency.

The subjective ease with which information is processed. Because fluent processing is misread as a signal of truth, anything that increases fluency—notably repetition—can raise perceived accuracy.

Illusory truth effect.

The increase in perceived accuracy of a statement that results from prior exposure to it; repeated claims feel more true, even when implausible or contradicted by stored knowledge.

Inoculation (prebunking).

A pre-emptive intervention that forewarns of, and refutes, a manipulation technique before exposure, conferring resistance to later misinformation built on that technique.

Misinformation.

Inaccurate information regardless of intent to deceive. Disinformation is the deliberately produced subset; the psychology of how each is believed and corrected is the same.

Motivated reasoning.

Reasoning directed toward a desired conclusion, such as accepting congenial falsehoods. Evidence suggests fake-news susceptibility owes more to insufficient reflection than to motivated reasoning.

Prior exposure effect.

The specific finding that a single earlier encounter with a fake-news headline raises its later perceived accuracy, even for implausible headlines and against the viewer’s politics—the illusory truth effect applied to disinformation.

Key Researchers

Sinan Aral

(Massachusetts Institute of Technology). Social scientist whose large-scale analysis of diffusion on social media established that false news spreads farther and faster than the truth, driven by human sharing of novel content rather than by bots. ORCID · Google Scholar · Faculty page

Ullrich K. H. Ecker

(University of Western Australia). Cognitive psychologist whose work on the continued influence effect and the psychology of misinformation correction maps why retractions fail and what makes them work. ORCID · Google Scholar · Faculty page

Lisa K. Fazio

(Vanderbilt University). Cognitive psychologist whose research on the illusory truth effect showed that repetition raises belief even for statements that contradict a person’s own knowledge. ORCID · Google Scholar · Faculty page

Stephan Lewandowsky

(University of Bristol). Cognitive scientist whose syntheses of misinformation, its correction, and the continued influence effect frame the modern study of belief in false information. ORCID · Google Scholar · Wikipedia

Sander van der Linden

(University of Cambridge). Social psychologist whose development of inoculation theory for misinformation—prebunking—shows that pre-emptively exposing manipulation techniques builds durable resistance. ORCID · Google Scholar · Wikipedia

Gordon Pennycook

(Cornell University). Psychologist whose work on analytic thinking, the prior-exposure effect, and the “lazy, not biased” account of fake-news susceptibility reframed who falls for disinformation and why. ORCID · Google Scholar · Wikipedia

David G. Rand

(Massachusetts Institute of Technology). Computational social scientist whose work on the accuracy-nudge intervention and the reasoning basis of fake-news belief links the cognition of misinformation to scalable responses. ORCID · Google Scholar · Wikipedia

Jon Roozenbeek

(King’s College London). Psychologist whose Bad News inoculation game demonstrated that teaching manipulation techniques from the inside confers psychological resistance to online misinformation. ORCID · Google Scholar · Faculty page

Frequently Asked Questions

What is disinformation?

Disinformation is false or misleading information that is deliberately created and spread in order to deceive, usually at the scale of publics rather than a single person. It is distinguished from misinformation, which is inaccurate information regardless of intent—the same false claim is disinformation in the mouth of its author and misinformation when innocently passed on.

How is disinformation different from misinformation?

The difference is intent. Disinformation is spread with the knowledge or purpose that it is false; misinformation is the broader category of inaccurate information whether or not the person spreading it means to deceive. Because the sharer’s intent is often unknowable, researchers frequently study misinformation as the umbrella phenomenon and treat disinformation as its deliberately produced subset.

Why don’t corrections work?

Because of the continued influence effect: a retraction updates what people can recite but not the mental model they reason from, so the discredited claim keeps shaping inference. Corrections work better when they are explicit, repeated, and supply a plausible alternative cause rather than a bare denial.

Why does repeating a false claim make it more believable?

Because of the illusory truth effect. Repetition makes a claim easier to process, and that fluency is misread as a sign of truth. Repeated statements are judged more accurate even when they are implausible and even when the person knows the correct answer.

Who is most likely to fall for fake news?

Susceptibility tracks a lack of analytic thinking—a failure to stop and reflect—more than it tracks education, intelligence, or political affiliation. More reflective reasoners are better at telling true from false headlines regardless of whether a headline flatters their politics.

Does false news really spread faster than true news?

Yes. A large analysis of social-media diffusion found that false stories spread farther, faster, and more broadly than true ones, especially for political news. The asymmetry is driven by people rather than bots, because false stories tend to be more novel and emotionally provocative, and novelty is what gets shared.

What is inoculation or prebunking?

It is a pre-emptive defence modelled on vaccination: forewarning people about a manipulation technique and refuting it before they encounter the real thing builds resistance. Because inoculation targets the technique rather than a single claim, one “dose” can generalise to many falsehoods built on the same trick.

Can anything reduce the sharing of misinformation?

Yes. People often share false claims without stopping to consider whether they are true, because the social-media context distracts from accuracy. A simple nudge that briefly directs attention to accuracy—such as asking someone to rate one headline’s truthfulness—improves the quality of what they share afterward.

References

Ecker, U. K. H., Lewandowsky, S., Cook, J., Schmid, P., Fazio, L. K., Brashier, N., Kendeou, P., Vraga, E. K., & Amazeen, M. A. (2022). The psychological drivers of misinformation belief and its resistance to correction. Nature Reviews Psychology, 1(1), 13–29. https://doi.org/10.1038/s44159-021-00006-y

Fazio, L. K., Brashier, N. M., Payne, B. K., & Marsh, E. J. (2015). Knowledge does not protect against illusory truth. Journal of Experimental Psychology: General, 144(5), 993–1002. https://doi.org/10.1037/xge0000098

Guess, A., Nagler, J., & Tucker, J. (2019). Less than you think: Prevalence and predictors of fake news dissemination on Facebook. Science Advances, 5(1), eaau4586. https://doi.org/10.1126/sciadv.aau4586

Johnson, H. M., & Seifert, C. M. (1994). Sources of the continued influence effect: When misinformation in memory affects later inferences. Journal of Experimental Psychology: Learning, Memory, and Cognition, 20(6), 1420–1436. https://doi.org/10.1037/0278-7393.20.6.1420

Lewandowsky, S., Ecker, U. K. H., Seifert, C. M., Schwarz, N., & Cook, J. (2012). Misinformation and its correction: Continued influence and successful debiasing. Psychological Science in the Public Interest, 13(3), 106–131. https://doi.org/10.1177/1529100612451018

Pennycook, G., Cannon, T. D., & Rand, D. G. (2018). Prior exposure increases perceived accuracy of fake news. Journal of Experimental Psychology: General, 147(12), 1865–1880. https://doi.org/10.1037/xge0000465

Pennycook, G., McPhetres, J., Zhang, Y., Lu, J. G., & Rand, D. G. (2020). Fighting COVID-19 misinformation on social media: Experimental evidence for a scalable accuracy-nudge intervention. Psychological Science, 31(7), 770–780. https://doi.org/10.1177/0956797620939054

Pennycook, G., & Rand, D. G. (2019). Lazy, not biased: Susceptibility to partisan fake news is better explained by lack of reasoning than by motivated reasoning. Cognition, 188, 39–50. https://doi.org/10.1016/j.cognition.2018.06.011

Pennycook, G., & Rand, D. G. (2021). The psychology of fake news. Trends in Cognitive Sciences, 25(5), 388–402. https://doi.org/10.1016/j.tics.2021.02.007

Roozenbeek, J., & van der Linden, S. (2019). Fake news game confers psychological resistance against online misinformation. Palgrave Communications, 5, 65. https://doi.org/10.1057/s41599-019-0279-9

van der Linden, S., Leiserowitz, A., Rosenthal, S., & Maibach, E. (2017). Inoculating the public against misinformation about climate change. Global Challenges, 1(2), 1600008. https://doi.org/10.1002/gch2.201600008

Vosoughi, S., Roy, D., & Aral, S. (2018). The spread of true and false news online. Science, 359(6380), 1146–1151. https://doi.org/10.1126/science.aap9559