Core framework: Misperceptions vs. information gaps
Nyhan distinguishes between being uninformed (lacking knowledge, often recognized) and being misinformed (holding false beliefs with confidence). Misperceptions are often tied to identity, political ideology, or personal values, making them psychologically resistant. A Mark Twain-era quote captures this: "It is better to know less than to know so much that ain't so"—sometimes being uninformed is preferable to being confidently wrong.
Why the deficit model fails
The traditional approach—give people facts, they update their beliefs—assumes rational Bayesian updating. In practice, when information conflicts with identity or deep-seated attitudes, people activate disconfirmation bias: they scrutinize counter-attitudinal information skeptically, generate counter-arguments, and may end up more entrenched.
Political misperceptions: Evidence from major myths
Iraq and weapons of mass destruction
Despite extensive news coverage of the failure to find WMDs after 2003, belief in the myth persisted. Polling data from Nyhan and colleagues showed that approximately 40% of Americans continued to believe Iraq had active WMD programs as of 2012—a decade after the invasion. This wasn't simple ignorance; belief plateaued and barely declined despite overwhelming, unavoidable evidence.
9/11 conspiracy theories
Partisan gradients appear in conspiracy beliefs: those predisposed by party affiliation to believe the Bush administration aided or allowed the attacks were far more likely than independents or the opposing party to agree. In one 2006 poll, over 40% of the most vulnerable partisan group endorsed this claim.
Obama birth certificate and presidential identity myths
Widespread false claims about Obama's birthplace or religion persisted for years. When educated respondents were exposed to factual information (actual birth certificate), the relationship between political knowledge and belief in the myth actually strengthened. More politically knowledgeable people—who were better able to defend their existing political beliefs—ended up more likely to believe the myth.
Death panels and Affordable Care Act (2009–2011)
Sarah Palin's "death panel" claim about end-of-life counseling spread rapidly. In experimental studies, Nyhan and colleagues showed that corrective information (from nonpartisan healthcare experts) was effective for low-knowledge respondents but had a small or reversed effect for high-knowledge respondents who held warm feelings toward Palin. This suggests sophisticated reasoners use their rhetorical skills to defend attitudinally-congenial misinformation.
Tax cuts and government revenue
In a mock-news study during the Bush administration, people were shown corrective information stating that tax cuts had decreased government revenue. Liberals updated their beliefs downward. Conservatives, however, showed a backfire effect: belief that tax cuts increased revenue doubled when exposed to the correction. The interpretation: they counter-argued and convinced themselves more of the original claim.
Health misinformation and the vaccine-autism debate
Nyhan shifted focus to vaccine hesitancy, testing whether the deficit model works better in health contexts. Data showed ~25% of American parents believe vaccines can cause autism, despite no credible evidence.
The CDC corrective-information study
In a representative experiment with American parents, Nyhan exposed some to CDC information stating vaccines do not cause autism. Results were paradoxical:
- Belief in the autism-vaccine myth went down (good news).
- But intention to vaccinate a future child plummeted among parents with low-to-moderate confidence in vaccines (from ~70% to ~50% likelihood of vaccinating). The worst-case group—vaccine hesitant parents—did not so much double down on the myth itself as they brought other vaccine concerns to mind (side effects, ingredient worries), ending up feeling more negatively about vaccines overall.
This pattern suggests the corrective information, while technically accurate, triggered defensive reasoning about vaccine safety more broadly.
The flu vaccine myth study
Similar results with the false claim "the flu shot can give you the flu": belief in the myth decreased, but among adults with high concern about vaccine side effects, seeing the correction made them less likely to say they'd get a flu shot. Narrative and imagery approaches (showing images of diseased children, telling stories of measles victims) produced no benefits and sometimes increased vaccine-safety fears.
Why misperceptions persist: Psychological mechanisms
Confidence despite ignorance
When Nyhan measured perceived knowledge ("How much do you know about this bill?") versus objective knowledge (quiz questions), a striking pattern emerged: for the partisan group predisposed to misperceptions, higher perceived knowledge correlated with higher belief in myths. They thought they knew more and were actually less informed—a kind of illusory expertise.
Disconfirmation bias and motivated reasoning
When people encounter unwelcome information, they don't passively receive it; they counter-argue. This is especially powerful for identity-linked issues. The more education and sophistication someone has, the better they become at constructing counter-arguments and aligning factual beliefs with political or social identity.
Familiarity breeds acceptance
Repeated exposure to a false claim makes it more familiar, and familiarity itself can feel like a truth signal. This implies that myth-busting journalism that repeatedly names and describes false claims may inadvertently boost perceived accuracy over time.
What works: Evidence-based recommendations
1. Use credible messengers
Government health agencies (23% trust) are far less credible than a child's pediatrician (76% trust). Messages from sources with credibility and professional stake (a doctor saying "I vaccinate my own children") are more persuasive than abstract agency statements.
Evidence from the elite-accountability field experiment
Nyhan and colleagues sent letters to state legislators warning they might be fact-checked. Those legislators were subsequently less likely to make false factual claims in public. This suggests holding elites accountable can reduce misinformation supply, even if it doesn't directly fix citizen beliefs.
2. Rely on cross-cutting messengers
During the 2009 healthcare debate, an ABC news story noting that even healthcare experts who oppose the ACA agree death panels aren't real was more persuasive than agency corrections. When someone argues against their own political interest, partisan audiences find them more credible.
3. Use graphical evidence
Nyhan presented evidence that graphical information (e.g., climate graphs from multiple independent data sources showing consensus) was more effective at correcting climate beliefs than equivalent text. Visual proof of agreement across sources can overcome skepticism.
4. Avoid repeating the myth
Journalist-style myth-busting ("Myth: X. Fact: Y. Myth: X again...") may reinforce familiarity with false claims, making them feel truer over time. Leading with the correct information or using visual evidence appears more effective.
5. Help providers engage hesitant patients
Healthcare providers need time and training to identify and address specific concerns. Pediatricians can probe hesitancy, align messaging with patient values, and use their trust relationship. Alternative vaccine schedules and over-accommodation should be avoided; they put children at risk and start patients "down the road to customization."
The smoking comparison: A success case
Smoking causes cancer represents one of the few factual claims that achieves overwhelming cross-partisan consensus (higher agreement than on evolution, climate change, or Earth's age). This happened because: (1) expert consensus became embarrassing to deny; (2) thousands of personal experiences with loved ones dying of smoking-related illness provided lived evidence; (3) court documents from tobacco companies themselves admitted they knew smoking was dangerous. The lesson: when elites on both sides endorse a claim and personal experience aligns with expert consensus, factual agreement can transcend polarization.
Broader implications
Polarization is historically normal
Nyhan notes that the mid-20th-century American era of bipartisan coalition-building was a historical anomaly, driven by racial segregation in the South. Current polarization is closer to the historical norm. The problem isn't polarization per se (opinion differences are healthy), but factual polarization—disagreement about basic facts.
The challenge for democracy
Our institutions were not designed for high factual polarization. Blocking is easy in the U.S. legislative system, and when both sides dispute basic reality, passing policy becomes nearly impossible. Resolving factual disputes requires media accountability, expert consensus, and elite incentives to be responsible.
Individual vs. systemic solutions
Nyhan is cautious about hopes for citizen-level fixes (more media literacy, better education). While education can help, more sophisticated people sometimes become better at defending preferred beliefs. Solutions must work at the system level: media norms, fact-checking that holds elites accountable, and expert communities that enforce shared epistemic standards.
Notes
This colloquium talk stands out for its clarity and empirical grounding across two domains (politics and health), showing that misperception dynamics transcend topic. Nyhan demonstrates that the problem is not simply informational deficit but psychological defense of identity-linked beliefs, especially among sophisticated reasoners.
The vaccine results are particularly striking because they show that corrections can simultaneously reduce myth-belief and increase hesitancy through broader defensive processing. Parents didn't deny the specific claim so much as activate other concerns, suggesting that corrective messaging in health contexts must consider the full attitude structure, not just isolated myths.
The elite accountability angle (field experiment with legislators) hints at a systems-level lever: if political and institutional incentives reward truthfulness, misinformation supply can be reduced even if it doesn't directly fix citizen psychology. This aligns with literature on how social norms and institutional design shape what elites claim.
A limitation is that short-term experimental effects may not reflect long-term real-world dynamics (reinforcement, social influence, habituation). Nyhan acknowledges this in Q&A, noting that experimental effects decay and longer-term longitudinal work is needed. Nevertheless, the evidence for backfire and defensive reasoning is robust enough to warrant caution in how corrections are framed and delivered.