Illusory truth effect: repeat it until it seems true
In the weeks following Russia’s full-scale invasion of Ukraine in February 2022, a coordinated wave of messaging emerged across Telegram channels, fringe forums, and state-affiliated media outlets: claims that Ukraine was developing biological weapons in U.S.-funded laboratories. The allegation had no credible evidentiary basis. Yet repeated exposure across multiple platforms — each iteration slightly reworded, each amplified by different accounts — produced measurable effects on public belief. Polling conducted by YouGov in March 2022 found meaningful segments of Western audiences reporting uncertainty about the biolab claims, despite unified debunking from the WHO and U.S. Department of Defense. This was not an accident. It was a deliberate exploitation of what cognitive scientists call the illusory truth effect: the empirically documented tendency for people to rate repeated statements as more true than statements they encounter for the first time. Understanding how this effect works — its mechanisms, its limits, and its operational application — is now a foundational requirement for anyone working in cognitive security, information operations analysis, or policy response.
The cognitive mechanism behind the illusory truth effect
Processing fluency and the truth heuristic
The illusory truth effect was first rigorously documented by Lynn Hasher, David Goldstein, and Thomas Toppino in a 1977 paper published in the Journal of Experimental Psychology: General. Their finding was straightforward but consequential: participants who heard statements repeated across sessions rated those statements as more likely to be true, regardless of the actual validity of the claims. The mechanism proposed — and subsequently refined — centers on processing fluency, the subjective ease with which information is encoded and retrieved by the brain.
When a statement is encountered for the second or third time, the cognitive system processes it more quickly. That fluency — that feeling of smooth, effortless recognition — is misattributed to familiarity with a fact rather than mere repetition of an exposure. The brain, operating under cognitive load constraints, uses fluency as a proxy for truth. It is a heuristic, not a deliberate judgment. This is sometimes framed within dual-process theory: System 1 processing (fast, automatic, associative) substitutes familiarity for validity without engaging the slower, effortful System 2 verification mechanisms.
Repetition bypassing prior knowledge
Perhaps the most operationally significant finding in this literature is that the illusory truth effect operates even when participants already know the correct answer. A 2015 study by Lisa Fazio and colleagues at Vanderbilt University — published in the Journal of Experimental Psychology: General — demonstrated that repeated exposure to false statements increased their perceived truth even among participants who initially answered related knowledge questions correctly. Participants who knew that the Earth’s atmosphere is primarily nitrogen still rated the false claim “the Earth’s atmosphere is mostly oxygen” as more plausible after repeated exposure. This is not a story about ignorance. It is a story about cognitive architecture that can be exploited at population scale.
Scientific grounding and the limits of the research
What the replication record actually shows
The illusory truth effect is among the more robustly replicated findings in cognitive psychology. It has been reproduced across varied stimulus materials, time intervals between exposures, and participant populations. A 2019 meta-analysis by Dechêne and colleagues, covering decades of studies, confirmed the effect’s consistency across experimental contexts. That said, intellectual honesty requires acknowledging where the evidence gets thinner.
Most laboratory studies use short repetition intervals and highly controlled stimuli — trivia statements, abstract claims. The ecological validity question — how well these findings generalize to the chaotic, multi-platform information environments in which adversarial influence campaigns operate — remains genuinely open. Real-world information exposure involves competing narratives, emotional valence, source credibility signals, and social reinforcement mechanisms that laboratory paradigms do not fully capture.
Cultural variability and boundary conditions
There is also limited cross-cultural research on whether the magnitude of the effect varies across different epistemic communities, trust-in-institutions baselines, or media literacy environments. Gordon Pennycook and David Rand’s work at MIT’s Computational Story Lab has pointed toward attention and analytical thinking as partial moderators: individuals primed to think analytically show reduced susceptibility. But priming is an intervention, not a stable population characteristic. The strategic implication is that populations under high cognitive load — during crises, during information floods, during periods of institutional distrust — are more vulnerable, not less.
How adversarial actors operationalize the illusory truth effect
Repetition architectures in FIMI campaigns
FIMI — Foreign Information Manipulation and Interference, the term adopted by the European External Action Service — campaigns rarely rely on a single broadcast. They are engineered for repetition across platforms, time, and source personas. The EU DisinfoLab’s 2019 investigation into the Indian Chronicles operation documented networks of hundreds of fake media outlets that reprinted and slightly reformatted the same core narratives, creating the illusion of independent corroboration while mechanically amplifying exposure frequency. Each re-encounter with a claim — even under a different headline — functions as another repetition trial.
What makes this architecturally sophisticated is that it exploits the illusory truth effect without requiring the audience to consciously recognize the repetition. The claim appears in different wrappers, attributed to different notional sources. The cognitive system registers familiarity with the underlying proposition while the meta-cognitive check — “wait, haven’t I seen this before?” — is bypassed by surface variation.
Firehose strategies and epistemic exhaustion
The RAND Corporation’s 2016 analysis of Russian propaganda, authored by Christopher Paul and Miriam Matthews, described what they termed the “firehose of falsehood” model: high-volume, high-velocity messaging that overwhelms fact-checking capacity and saturates the information environment. This strategy has an underappreciated relationship to the illusory truth effect. Volume is not simply about drowning out corrections — it is about increasing repetition exposure for favorable narratives while simultaneously exhausting the cognitive resources needed to engage System 2 verification. In that context, the illusory truth effect and information overload function as a compound vulnerability.
Does correction work — and under what conditions?
The asymmetry between encoding and correction
This is where the operational picture becomes genuinely concerning. Research on corrections and retractions consistently shows that corrections reduce but rarely eliminate the influence of repeated misinformation. The work of Ullrich Ecker and colleagues at the University of Western Australia has documented a “continued influence effect”: corrected misinformation continues to shape reasoning even after the correction is accepted. When the original false claim has been repeated more than the correction, the fluency asymmetry favors the original narrative.
This creates a structurally asymmetric contest between information operations and counter-messaging efforts. The adversary repeats a false claim dozens or hundreds of times before a correction is even issued. The correction, issued once or twice through official channels, faces a fluency deficit from the outset. Prebunking — exposing audiences to weakened forms of manipulation techniques before they encounter the real campaign — has shown more promise in recent research, including work supported by the UK’s RESIST counter-disinformation toolkit framework.
Platform architecture as an amplification layer
Recommendation algorithms on major social platforms are not neutral with respect to the illusory truth effect. Engagement-optimized systems create what might be called automated repetition infrastructure: content that generates interaction — including outrage-driven interaction — is re-served to users across sessions. A false claim encountered Monday may be algorithmically re-surfaced Thursday and again Sunday. The platform is, functionally, running repetition trials on its user base as a byproduct of its engagement optimization. This is not a designed influence operation — but its cognitive effects are structurally indistinguishable from one.
Myth vs. reality: common misconceptions about illusory truth susceptibility
Myth: only low-information or poorly educated people are affected
This is the most operationally dangerous misconception in this space. The evidence does not support it. Fazio’s 2015 research demonstrated the effect among participants with demonstrably accurate prior knowledge. Pennycook and Rand’s work on “bullshit receptivity” shows that analytical thinking moderates susceptibility somewhat, but does not eliminate it. Expert communities are not immune — they are simply exposed to different information environments with different repetition patterns. In my assessment, the belief that expertise confers blanket protection is itself a cognitive vulnerability: it reduces vigilance precisely in the domains where it matters most.
Reality: familiarity is a near-universal truth signal
The processing fluency mechanism is not a bug in a subset of cognitive architectures. It is a feature of human cognition at population scale. The inferential shortcut — familiar equals trustworthy — was adaptive in low-information environments where repeated exposure to a concept generally indicated its validity. In high-volume, algorithmically amplified information environments, that same shortcut becomes a systematic liability. No demographic is fully outside its reach.
A framework for assessing illusory truth exploitation in information operations
For practitioners analyzing influence campaigns, the following indicators suggest deliberate operationalization of the illusory truth effect:
- Cross-platform claim synchronization: The same core proposition appearing across ideologically distinct platforms within a compressed timeframe, often with surface-level variation in wording.
- Source persona diversity without narrative diversity: Multiple apparently independent outlets or accounts amplifying a functionally identical claim — a signature pattern identified in EU DisinfoLab and Stanford Internet Observatory investigations.
- Pre-correction saturation: High repetition volume established before credible counter-messaging enters the information environment, creating a fluency deficit for corrections.
- Emotional packaging variation: The same factual claim embedded in different emotional or partisan frames to reach different audience segments while accumulating repetition exposure across all of them.
- Timing correlation with cognitive load events: Claim saturation campaigns that launch during crises, breaking news cycles, or high-anxiety periods — when System 2 resources are most depleted.
| Indicator | Operational Signal | Assessment Priority |
|---|---|---|
| Cross-platform claim synchrony | Coordinated amplification network | High |
| Source diversity, narrative uniformity | Manufactured credibility through volume | High |
| Pre-correction saturation | Fluency asymmetry exploitation | High |
| Emotional frame variation | Multi-segment targeting | Medium |
| Crisis-timed deployment | Cognitive load amplification | Medium |
OSINT analysts tracking these patterns should prioritize temporal mapping — when does a claim first appear, how quickly does repetition volume build, and what is the lag time before authoritative counter-messaging reaches comparable exposure levels? That temporal gap is where illusory truth exploitation is most consequential.
Conclusion: forward assessment
The illusory truth effect is not an exotic laboratory curiosity. It is a well-documented property of human cognition that adversarial actors — state-sponsored and otherwise — have learned to engineer at scale. The convergence of algorithmic amplification, FIMI campaign architectures, and the structural asymmetry between false claim saturation and correction reach means this vulnerability is likely to be more, not less, exploited in the near term. NATO’s focus on cognitive security within its evolving resilience frameworks acknowledges the problem in principle; translating that acknowledgment into prebunking infrastructure, platform accountability mechanisms, and analytical capacity-building remains the practical challenge. What concerns me here is not the existence of the effect — that is settled science. It is the widening gap between what the research tells us and what institutional responses have actually implemented.
The most dangerous form of the illusory truth effect may be the one operating on the institutions that believe they are already countering it.
Sources
- Hasher, L., Goldstein, D., & Toppino, T. (1977). Frequency and the conference of referential validity. Journal of Experimental Psychology: General, 106(1), 107–112.
- 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.
- Paul, C., & Matthews, M. (2016). The Russian “Firehose of Falsehood” Propaganda Model. RAND Corporation.
- 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.
- EU DisinfoLab. (2019). Indian Chronicles: Deep Dive into a 15-Year Operation Targeting the EU and UN. EU DisinfoLab.
- Ecker, U. K. H., Lewandowsky, S., & Tang, D. T. W. (2010). Explicit warnings reduce but do not eliminate the continued influence of misinformation. Memory & Cognition, 38(8), 1087–1100.
