Why Conspiracy Theories Spread Online

A false story about a Twitter user's post reached 1,500 people roughly six times faster than a true one, according to the largest study ever done on the spread of news online, an MIT analysis of 126,000 stories shared more than 4.5 million times between 2006 and 2017. That gap is not an accident of bad luck. It is the product of two systems working together: a human brain that evolved to find patterns and threats fast, and a ranking algorithm that rewards whatever keeps people scrolling. Understanding both is the only way to understand why conspiracy theories move the way they do.

What a conspiracy theory actually is

A conspiracy theory is an explanation for an event that attributes it to a secret, coordinated plot by powerful actors, usually in place of a more mundane, distributed, or accidental explanation. Not every conspiracy theory is false: real conspiracies exist, from Watergate to the tobacco industry's decades-long cover-up of smoking research. What defines conspiracy thinking as a psychological pattern is not that the claim turns out to be wrong. It is that the claim is treated as unfalsifiable, that contrary evidence is read as proof of a deeper cover-up, and that the explanation is adopted with a confidence the evidence does not support.

Karen Douglas, a professor of social psychology at the University of Kent who has spent close to two decades studying this territory, and her colleagues describe belief in conspiracy theories as driven by three overlapping motives: epistemic (the need for knowledge and certainty), existential (the need to feel safe and in control), and social (the need to feel good about oneself and one's group). None of those needs is irrational on its own. Wanting to understand a confusing event, wanting to feel less powerless, and wanting to belong to a group that shares your values are all normal human drives. Conspiracy theories exploit all three at once, which is part of why they are so sticky.

A 2018 MIT Media Lab study of 126,000 news stories found falsehoods were 70 percent more likely to be retweeted than true stories, and reached their first 1,500 people about six times faster. The effect was strongest for false political stories, not for false claims about science, disasters, or finance.

The psychological mechanism: three needs, one story

The reason a conspiracy theory can feel more convincing than the boring, verified explanation is that it answers more than one question at the same time. A single narrative supplies a cause, a villain, a sense of control, and a group of people who already agree with you. Think of it like a vending machine that dispenses four different snacks for the price of one: the "purchase" (belief) feels efficient even when the snacks are stale.

Epistemic need: filling the gap uncertainty leaves behind

Humans are pattern-detection machines by design. A landmark study by Jennifer Whitson and Adam Galinsky, published in Science in 2008, found that people put into a state of reduced control were significantly more likely to perceive patterns in random noise, including seeing images in static and detecting nonexistent connections between unrelated events. Uncertainty does not just make people uncomfortable; it makes the brain actively hunt for a structure to attach itself to, even when no structure exists. A sudden, unexplained, high-stakes event, like a pandemic, an assassination, or a plane disappearing, creates exactly the kind of information vacuum this mechanism is built to fill.

Existential need: trading helplessness for a villain

Feeling unsafe is aversive, but feeling unsafe and not knowing why is worse. A conspiracy theory converts a diffuse, uncontrollable threat, like a virus with no single cause, into a concentrated, controllable one, like a group of people making deliberate choices. That trade matters even when it makes the world sound scarier, because a villain can in principle be resisted, exposed, or fought. Randomness cannot. Douglas's research group has repeatedly found that conspiracy belief rises during periods of societal crisis and among people who report feeling powerless, which tracks with this existential function rather than a general dislike of evidence.

Social need: identity, belonging, and the "need for uniqueness"

Believing something most people do not believe can function as a badge. Research by Roland Imhoff and Pia Lamberty on what they call the "need for uniqueness" found that people who score higher on wanting to see themselves as special and different from the mainstream are more likely to endorse conspiracy theories, independent of the specific content of the theory. Joining an online community built around a shared secret knowledge also supplies something many people are short on: a defined in-group, a shared vocabulary, and status that comes from being an early believer rather than a "sheep" who was fooled.

How platforms turn a psychological tendency into a mass phenomenon

None of the three needs above require the internet. Conspiracy theories about assassinations, moon landings, and secret societies circulated for decades through pamphlets, radio, and word of mouth. What changed with social platforms is the speed, the scale, and, most importantly, the selection pressure applied to which explanations survive.

Novelty and emotion beat accuracy in the feed

The MIT study cited above did not just measure speed. It measured why false stories moved faster, and the answer was two-fold: false stories were significantly more novel than true stories, and they provoked stronger emotional reactions, specifically fear, disgust, and surprise, compared to the anticipation, sadness, joy, and trust that accompanied true stories. Ranking systems built to maximize engagement do not have a truth filter. They have an attention filter, and novelty plus strong emotion is a more reliable predictor of a click, a share, or a comment than accuracy is.

Recommendation systems as an accidental funnel

Renée DiResta, a researcher who has studied disinformation networks at Stanford's Internet Observatory, has described modern recommendation systems as effectively "computational propaganda" machines even when no human at the company intends that outcome: an algorithm optimized purely for watch time or session length will, left alone, learn that more extreme or more emotionally charged content keeps a subset of users engaged longer, and will serve more of it. The mechanism does not require the platform to endorse the conspiracy theory. It only requires the system to notice that people who watch one conspiracy video tend to watch a second one.

Deplatforming does not delete belief, it disperses it

A 2023 study of conspiracy communities found that groups banned from mainstream platforms like Reddit and Facebook frequently reconstitute themselves on smaller, less moderated platforms such as Telegram and niche forums, and that these communities can become more resilient and more ideologically intense after the move, not less, because the remaining members are the most committed ones and the community loses exposure to outside pushback. This matters for anyone assuming that removing a conspiracy community from a large platform ends the problem: it often relocates and concentrates it instead.

QAnon began as a single anonymous post on the message board 4chan in October 2017 and grew into a movement referenced by candidates for the U.S. Congress within three years, illustrating how quickly a fringe claim can move from a niche forum to mainstream political discourse once it is picked up and amplified across larger platforms.

Why smart, educated people are not protected

One of the most persistent myths about conspiracy belief is that it is a problem of low intelligence or poor education. The research does not support that framing, and treating it as true tends to backfire, because it makes debunking feel condescending rather than informative.

Analytical intelligence and education correlate weakly, and inconsistently, with conspiracy belief once other factors are controlled for. What predicts belief more reliably is a cluster of dispositional traits: a general distrust of official information sources, a tendency toward what researchers call "teleological thinking" (assuming events happen for a purpose rather than by chance), feelings of social or political powerlessness, and belonging to a social network where the theory is already normalized. Smart people are not immune to any of those factors. If anything, verbal fluency and confidence can make a person a more persuasive advocate for a belief they adopted for non-rational reasons, because they are better equipped to construct a defense for it after the fact.

This is why fact-only corrections tend to underperform. Telling someone their belief is wrong, without addressing the epistemic, existential, or social need the belief was serving, leaves the underlying need unmet. Something else, often another unverified claim, tends to fill the space the correction opened up.

What actually reduces belief: prebunking over debunking

If correcting a conspiracy theory after it spreads is inefficient, the research points to a different strategy: intervening before exposure rather than after. This is called inoculation theory, or "prebunking," and it borrows its logic directly from vaccination. Sander van der Linden, who leads the Cambridge Social Decision-Making Lab, and his collaborators have run field experiments showing that briefly exposing people to a weakened, clearly labeled version of a manipulation technique, such as a fake headline built around a false dichotomy or an appeal to emotion, makes them measurably more resistant to the real thing when they encounter it later.

A set of short animated videos built on this model, tested across seven experiments involving close to 30,000 participants and published in Science Advances, increased viewers' ability to spot manipulative content after a single viewing, and the effect held up when the videos were served through YouTube's own recommendation system rather than in a lab. Google's research unit Jigsaw has since run prebunking ad campaigns aimed at populations at high risk of exposure to specific disinformation narratives, including a 2022 campaign in Poland, Slovakia, and the Czech Republic targeting false claims about Ukrainian refugees.

The mechanism behind prebunking is closer to how a vaccine works than to how a fact-check works. A fact-check tries to remove a belief that already exists, competing against everything that made the belief attractive in the first place. Prebunking tries to arrive first, giving people a mental category, "this is the kind of claim designed to make you angry before you can check it," that they can apply the next time they encounter something built the same way. It is like teaching someone to recognize a specific magic trick's setup rather than debunking one particular performance of it: the skill transfers to tricks they have not seen yet.

Case studies: how the mechanism plays out in practice

Abstract mechanisms are easier to evaluate against real events. Three cases, spanning different platforms and different eras, show the same pattern of epistemic, existential, and social needs meeting algorithmic amplification.

Pizzagate (2016). A false claim that a Washington, D.C. pizzeria was the hub of a child-trafficking ring run by prominent political figures spread across Twitter, Reddit, and 4chan in the weeks before and after the 2016 U.S. presidential election. It combined a real existential fear (child safety), a search for hidden causes behind a bitterly contested election, and an in-group of self-styled investigators piecing together "clues." In December 2016, a man drove from North Carolina to the restaurant and fired a rifle inside it while searching for captives that did not exist, illustrating that online belief does not stay confined to the online space.

QAnon (2017 onward). What began as a single anonymous forum post claiming to be from a government insider grew, through a mix of YouTube explainer videos, Facebook groups, and dedicated forums, into a decentralized movement with adherents on multiple continents within a few years. Its structure rewarded exactly the social need Imhoff and Lamberty describe: followers were framed as "digital soldiers" doing important, secret research, which supplied status and purpose that ordinary information consumption does not.

COVID-19 misinformation (2020–2022). During the pandemic, the same three-need framework mapped closely onto real behavior. Douglas and colleagues found that people higher in conspiracy belief were more reluctant to take precautionary measures, including vaccination, a pattern consistent across multiple countries and consistent with the pattern that uncertainty and threat increase demand for a coherent, controllable explanation, even a false one.

Common misconceptions about why conspiracy theories spread

"Only uneducated or unintelligent people believe them." This is not supported by the research on predictors of conspiracy belief, which points instead to distrust, felt powerlessness, and social network effects. Education can reduce exposure to some misinformation formats but does not reliably reduce belief once the underlying need is present.

"Algorithms are the whole explanation." Ranking systems accelerate and select for engaging content, but they did not invent pattern-seeking, distrust, or in-group identity. Conspiracy theories about assassinations and secret societies predate social media by decades. The honest description is that platforms are an amplifier layered on top of a pre-existing psychological tendency, not the sole cause.

"Debunking with facts fixes it." Corrections that only address the factual claim, without addressing the emotional or social need the belief was meeting, are frequently ignored or produce a backfire effect where the person doubles down. The stronger evidence base is behind prebunking, which intervenes before exposure rather than after belief has formed.

"Banning a community online ends the problem." Deplatforming reduces reach on the platform doing the banning, but research on displaced communities shows they often reconstitute on smaller platforms and can become more concentrated and more resistant to outside correction, not less.

Comparing debunking and prebunking

Approach	When it's applied	Mechanism	Evidence of effectiveness
Fact-check / debunking	After a claim has already spread and been believed	Attempts to replace a false belief with a correct one directly	Mixed; can reduce belief modestly but sometimes triggers backfire effects, especially when identity is involved
Prebunking / inoculation	Before exposure to the specific claim	Teaches recognition of manipulation techniques (false dichotomies, emotional appeals, scapegoating) so they generalize	Field-tested at scale via YouTube (Cambridge/Jigsaw, ~30,000 participants across 7 experiments); measurable resistance gains
Deplatforming	After a community has formed around a claim	Removes reach and discovery on a given platform	Reduces mainstream visibility but can concentrate and radicalize the remaining community on smaller platforms
Algorithmic friction (labels, slower shares)	At the point of sharing	Adds a pause before amplification, reducing the novelty/emotion advantage false content has	Platform-reported reductions in resharing of labeled content; independent replication is more limited

Table: Four common responses to conspiracy content and what the research says about each.

Frequently Asked Questions

Q: Why do conspiracy theories spread faster than accurate information online? A: A large-scale MIT analysis of over 126,000 news stories found false stories were 70 percent more likely to be retweeted than true ones and reached 1,500 people about six times faster, largely because false content tends to be more novel and provokes stronger emotional reactions, both of which ranking algorithms reward.

Q: Are conspiracy theory believers less intelligent than other people? A: No. Research on predictors of conspiracy belief points to distrust of institutions, feelings of powerlessness, and social network effects, not intelligence or education level, as the stronger and more consistent predictors.

Q: What psychological needs do conspiracy theories satisfy? A: Researchers, notably Karen Douglas at the University of Kent, describe three overlapping motives: epistemic (wanting certainty and understanding), existential (wanting to feel safe and in control), and social (wanting to feel good about oneself and one's group).

Q: Does correcting a conspiracy theory with facts make people change their mind? A: Sometimes, but the effect is often modest, and corrections can backfire, especially when the belief is tied to identity or group membership. Interventions that work before exposure, called prebunking, generally show stronger and more durable effects.

Q: What is prebunking and how is it different from fact-checking? A: Prebunking exposes people to a weakened, labeled example of a manipulation technique before they encounter the real version, similar to how a vaccine works. Cambridge University research involving nearly 30,000 participants across seven experiments found this builds a transferable skill for spotting manipulation, unlike fact-checking, which addresses one specific false claim after the fact.

Q: Do algorithms cause conspiracy theories or just spread them? A: Algorithms did not invent conspiracy thinking, which predates the internet by centuries, but engagement-optimized ranking systems reliably amplify novel, emotionally charged content, giving conspiracy theories a structural advantage in the feed over duller, accurate explanations.

Q: Does banning a conspiracy community from a platform stop it from spreading? A: It reduces the community's reach on that specific platform, but research on deplatformed groups shows they frequently relocate to smaller, less moderated platforms where the remaining, most committed members can become more resistant to outside correction rather than less.

Q: Can real conspiracies make people more likely to believe false ones? A: There is no direct causal study proving this, but researchers note that documented real conspiracies, such as Watergate or the tobacco industry's concealment of smoking research, make the general premise "powerful groups sometimes coordinate in secret" plausible, which lowers the threshold for accepting unverified claims that share that structure.

Q: Why do some people need to believe things others do not? A: Roland Imhoff and Pia Lamberty's research on the "need for uniqueness" found that people who want to see themselves as different from the mainstream are more likely to endorse conspiracy theories regardless of the specific claim, because holding an unusual belief itself supplies a sense of distinctiveness.

The takeaway

Conspiracy theories spread online because they solve three problems at once, filling in uncertainty, restoring a feeling of control, and supplying identity and belonging, and because the platforms carrying them are built to reward exactly the novelty and emotional intensity that false, dramatic explanations tend to have in excess. Neither piece explains the phenomenon alone. A ranking algorithm cannot make someone believe a claim that meets no psychological need, and a psychological need alone would not have spread QAnon to millions of people in three years without a distribution system built to amplify it. The research-backed response is not to wait for a fringe idea to become mainstream and then debunk it, but to build resistance to the underlying manipulation techniques before exposure happens, which is the specific advantage prebunking has shown over after-the-fact correction.

---

Read next:

• Cognitive Biases Explained: Why Your Brain Shortcuts Cost You
• How Memory Can Be False: The Science of Reconstructive Memory
• The Asch Conformity Experiment

Sources

• Vosoughi, S., Roy, D., & Aral, S. (2018). "The spread of true and false news online." Science, 359(6380). science.org/doi/10.1126/science.aap9559
• MIT Sloan. "Study: False news spreads faster than the truth." mitsloan.mit.edu
• American Psychological Association. "Why people believe in conspiracy theories, with Karen Douglas, PhD." Speaking of Psychology podcast. apa.org/news/podcasts/speaking-of-psychology/conspiracy-theories
• Douglas, K. M., Sutton, R. M., & Cichocka, A. "The Psychology of Conspiracy Theories." Current Directions in Psychological Science. researchgate.net/publication/317401748
• Whitson, J. A., & Galinsky, A. D. (2008). "Lacking Control Increases Illusory Pattern Perception." Science, 322(5898).
• University of Cambridge. "How to 'inoculate' millions against misinformation on social media." cam.ac.uk/stories/inoculateexperiment
• Science/AAAS. "Can people be 'inoculated' against misinformation?" science.org/content/article/can-people-be-inoculated-against-misinformation
• arXiv. "Online conspiracy communities are more resilient to deplatforming." arxiv.org/pdf/2303.12115
• Imhoff, R., & Lamberty, P. (2017). "Too special to be duped: Need for uniqueness motivates conspiracy beliefs." European Journal of Social Psychology.