Concept

Confirmation Bias

cognitive-bias identity persuasion

People favor information, framing, and products that confirm a belief they already hold about themselves — particularly a flattering self-image (helpful, smart, adventurous, savvy) — over information that's simply true or useful. Marketers can exploit this by asking or telling someone something that activates a flattering self-identity, then attaching the desired action to that identity.

Street-survey compliance studies. A 2009 study found people stopped on the street were far more likely to agree to complete a full survey if first asked "are you a helpful person?" — compliance rose from 29% to 77.3% (see Helpful/Adventurous Compliance-Priming Studies). A follow-up found willingness to try a new soft drink rose from 33% to 75.7% when people were first asked "are you adventurous?" — in each case, agreeing with the flattering self-description created pressure to act in a way consistent with it.

Applied at scale by Sainsbury's. Facing slow adoption of its Smart Shop cost-saving app despite a straightforwardly rational value proposition (it saves time and money), Sainsbury's and Ogilvy built a TV campaign entirely around confirmation bias rather than the app's features — "be a smart cookie," directly naming and needling a shopper ("Psst, John... you could have saved money... you silly sausage") for not already using it. The campaign drove 275,000 downloads, 13% above forecast, and a 20% increase in adoption (see Sainsbury's Smart Shop Confirmation-Bias Campaign).

Confirmation bias here works by the same underlying route as Noun-Based Identity Framing — attaching a desired behavior to an identity the audience already wants to hold — but confirmation bias specifically targets people's belief about their existing traits, rather than inviting them to adopt a new one.

The founding demonstration of confirmation bias distorting judgment of evidence itself (Episodes 183, 187). Researchers Lord, Ross, and Leeper showed 48 Americans — half strongly for, half strongly against the death penalty — two fabricated studies, one supporting and one opposing capital punishment's deterrent effect. Each side rated the study matching their own view as well-conducted and convincing, and the opposing study as poorly done — leaving both sides more entrenched than before rather than converging (see Lord Ross Leeper Confirmation-Bias Study). JFK's 1962 Rice University moon speech is analyzed as deliberately exploiting the same mechanism, repeatedly appealing to Americans' existing self-image as global leaders rather than presenting new facts or figures (see JFK Moon Speech Case). Tali Sharot frames the fix as leading with common ground rather than disagreement: a UCLA study found doctors persuaded more parents to vaccinate their children by emphasizing shared, already-agreed-upon facts (that vaccines protect against real diseases) rather than directly contradicting parents' vaccine-autism fears (see UCLA Vaccine Common-Ground Study), and Republican voters donated more to an environmental charity when the appeal was reframed around patriotism rather than harm and fairness (see Republican Climate-Donation Patriotism Study).

Confirmation bias under maximal commitment (Episode 195). A small 1950s doomsday cult's belief in an apocalyptic flood, prophesied by "guardians" who would teleport members to safety, survived the flood's failure to arrive — members grew more committed, not less, after the leaders reframed the non-event as a delay caused by their own faith. Robert Cialdini cites the case in Influence to argue that once the psychological, social, and financial cost of belief is high enough, disconfirming evidence gets reinterpreted to fit rather than accepted (see Festinger Doomsday-Cult Confirmation Bias Study). The same episode demonstrates the bias directly on its own audience: three plausible-sounding "studies" were invented and presented as real mid-episode, and were widely believed specifically because they were unsurprising and consistent with the audience's existing trust in the show and in priming research generally — not because the claims were individually verified (see Nudge Podcast Fabricated-Studies Demonstration — note: those three claims are fabricated and must never be cited as real elsewhere). This pattern of confident, plausible, hard-to-verify claims going unquestioned is also central to how Diederik Stapel fabricated research for nearly a decade; see Truth Bias (Truth-Default Theory) and Replicability Crisis.

Why the "10,000-hour rule" spread so widely despite weak evidence (Episode 224). Alex Edmans, author of May Contain Lies, argues Malcolm Gladwell's popularized rule persists because people want it to be true — the alternative (that talent, not just effort, gates success) is uncomfortable. The underlying study measured only violin students' self-reported practice hours from 13 years earlier, a design vulnerable to reverse causation (skilled violinists over-recall their practice, mediocre ones under-recall), and covered only violin, not the general claim Gladwell built on it. The original paper's real finding was that practice is necessary but not sufficient — talent still matters — a nuance the popularized version dropped entirely (see Ericsson Berlin Violin 10,000-Hour Study).

A neurological basis for the bias, from an fMRI study. Researchers Jonas Kaplan, Sarah Gimbel, and Sam Harris found that showing people evidence contradicting a politically-charged, identity-linked belief (on gun control or abortion) activated the amygdala — the brain's fight-or-flight threat-response region — while contradicting a non-political factual belief (that Edison invented the light bulb) produced no such reaction, even though both beliefs were held with similar confidence. The brain processes a challenge to an identity-linked belief like a physical attack, but not a challenge to a belief with no identity stake attached (see Kaplan Gimbel Harris Political-Belief Amygdala Study).

Real-world stakes beyond marketing. Wellness blogger Belle Gibson fabricated a cancer diagnosis and a "natural cure" recovery story, reaching millions of followers and contributing to at least one death, because her story confirmed pre-existing beliefs (that willpower conquers illness, that natural remedies beat pharmaceutical treatment) rather than challenging them (see Belle Gibson Cancer-Cure Hoax Case). Criminologists Kim Rossmo and Joycelyn Pollock found confirmation bias present in 74% of 50 major US wrongful-conviction exonerations studied — the single most common contributing factor, ahead of unreliable witnesses, media pressure, or flawed forensics (see Rossmo Pollock Wrongful-Conviction Confirmation-Bias Study). Edmans also describes testifying to a UK House of Commons inquiry where a witness cited an unpublished, half-finished 2016 draft study claiming lower CEO-worker pay gaps improve company performance — when the finished 2013 version of the very same study found the opposite — because the flattering, half-finished draft was the version that told the committee what it wanted to hear; the committee's recommendation, based partly on that citation, became UK law.

A practical defense: "imagine the opposite." Before accepting a study or claim that confirms an existing belief, deliberately imagine it had found the opposite result and list the questions you'd raise to challenge it (sample size, timeframe, measurement validity, confounding variables) — then apply those same questions to the version you actually like. Edmans argues this doesn't require statistical training, only consistently applying scepticism people already reserve for claims they dislike. See also Narrative Fallacy for a closely related bias covered in the same episode — inventing a plausible causal story to explain an outcome, rather than favoring belief-confirming evidence directly.

Discussed in

  • Episode 163 — 163-how-sainsbury-s-and-adobe-nudge-customers
  • Episode 183 — 183-the-speech-that-put-a-man-on-the-moon
  • Episode 187 — 187-how-to-win-any-argument-according-to-a-neuroscientist
  • Episode 195 — 195-the-lying-psychologist-who-fooled-the-world
  • Episode 224 — 224-can-10-000-hours-of-practice-make-you-great (10,000-hour rule debunked; amygdala fMRI study; Belle Gibson; wrongful convictions)
  • Episode 225 — 225-why-most-bestselling-business-books-are-bs (companion episode on Black-and-White Thinking; Simon Sinek's "Start With Why")

Related