Belief Perseverance

At a Glance

1. Quick Summary

Belief perseverance is what happens when the ground is pulled out from under a belief and the belief stays where it was. In the original 1975 experiment, students were given invented scores on a task, were then told plainly that the scores had been invented, and still rated their own ability in line with the fake numbers. The reason is explanation: as soon as we accept a claim we start building an account of why it makes sense, and that account does not vanish when the claim is retracted. It is distinct from confirmation bias, which concerns how we gather evidence while a belief is alive, and from the backfire effect, the claim that corrections strengthen false beliefs, which large recent studies have found to be rare. Corrections do work, partly. They just leave a residue.

Simply About the Bias

Belief perseverance means you keep believing something even after you find out the reason you believed it was wrong. Your brain built a little story around the idea, and the story stays behind when the facts leave.

Say a friend tells you a kid in your class cheated on a test. You think about it and it kind of fits. He finished really fast, and he always sits near the smart kids. The next day your friend says sorry, she made it up because she was annoyed with him. You say okay. But a week later he gets a good grade, and the first thing you think is, hmm. The rumor is gone. The feeling it left behind is not.

Here is one way to catch it. When you learn that something you believed was false, say out loud what you had thought was the reason, and then ask: what else could explain that? He finished fast because he studied. He sits near the smart kids because his friends sit there. Giving the old clues a new explanation helps the old belief let go.

2. The Science Behind It

2.1. Discovery and History

Francis Bacon wrote in 1620 that once the mind has adopted an opinion it "draws all things else to support and agree with it." Leon Festinger watched it happen in 1954, when a group whose leader had predicted the end of the world met the morning after with a fresh explanation and renewed zeal; his When Prophecy Fails framed this as cognitive dissonance, but what it documents is perseverance.

The laboratory work began at Stanford with a practical worry: psychology experiments deceive participants and then "debrief" them, and Lee Ross, Mark Lepper and Michael Hubbard asked whether debriefing undoes the deception. In 1975 they found that it did not, or only partly. Lord, Ross and Lepper (1979) showed that opposed partisans read the same mixed evidence and each came away more convinced (biased assimilation). Anderson, Lepper and Ross (1980) extended perseverance to beliefs about the world, and Craig Anderson's 1982 and 1983 papers found the engine: people who had written an explanation clung hardest, and explaining the opposite was the one intervention that reliably loosened their grip.

The memory literature added a sibling, the continued influence effect, in which a retracted detail still shapes later inferences (Johnson & Seifert, 1994), and Lewandowsky and colleagues' 2012 review joined the two literatures. Then came a political twist. Nyhan and Reifler (2010) reported that correcting false political claims could make some partisans believe them more, the "backfire effect," and the idea spread fast. Wood and Porter (2019), with more than ten thousand participants and 52 issues, found backfire almost nowhere, and Nyhan argued in 2021 that misperceptions endure for other reasons. The honest summary: the debriefing result is solid, explanation is the mechanism, and corrections reduce false beliefs without erasing them.

Key milestones:

• 1956: Festinger, Riecken and Schachter publish When Prophecy Fails
• 1975: Ross, Lepper and Hubbard demonstrate perseverance in the debriefing paradigm and show that a "process debriefing" undoes it
• 1979 to 1982: Lord, Ross and Lepper document biased assimilation; Anderson, Lepper and Ross tie perseverance to explanation; Anderson shows counter-explanation reduces it
• 2010: Nyhan and Reifler report a backfire effect; 2019: Wood and Porter fail to find it across 52 issues

2.2. Key Researchers

2.3. Landmark Studies

Perseverance in the debriefing paradigm (Ross, Lepper & Hubbard, 1975)

Stanford undergraduates read 25 pairs of suicide notes and tried to pick the genuine one in each pair, a task nobody can really do. Feedback was fixed in advance: "success" participants were told they got 24 of 25 right, "failure" participants 10, and an average group 17. The experimenter then explained, at length, that the feedback had been predetermined, and participants rated how good they actually were at the task.

The fake scores should have been inert by then. They were not. Participants told they had succeeded rated themselves as more able than those told they had failed, and observers who had watched everything, debriefing included, showed the same pattern. A second experiment added a "process debriefing" that explained the perseverance phenomenon itself, and that version worked.

Perseverance of social theories (Anderson, Lepper & Ross, 1980)

Participants read two case histories of firefighters, paired so that in one condition the risk-taker succeeded and in the other the cautious one did. From two cases, people formed a confident theory about risk-taking and firefighting. They were then told the cases were fictitious. Asked afterward, they still believed whatever their invented cases had suggested, and those who had written out an explanation beforehand held on hardest.

Biased assimilation and attitude polarization (Lord, Ross & Lepper, 1979)

Supporters and opponents of capital punishment read summaries of two studies, one finding that the death penalty deters murder and one finding it does not. Both groups rated the study that agreed with them as better designed, found flaws in the other, and ended more polarized than they began. This is how an established belief digests new information in a way that protects it.

2.4. Neurological Basis

The neuroscience of belief perseverance specifically is thin. No imaging study I know of has reproduced the debriefing paradigm in a scanner, so this is a plausible mechanism built from better-established findings.

The closest study is Kaplan, Gimbel and Harris (2016), who scanned people while their political beliefs were challenged. Challenges to firmly held beliefs produced more activity in the default mode network and in the amygdala and insula, and participants with stronger activity there changed their minds less. That concerns resistance to challenge rather than perseverance after discrediting, so it shows that belief defense recruits self-related and emotional systems without mapping the debriefing effect.

The memory side is firmer. Retracting information does not delete it: the original trace remains, the retraction is stored alongside as a competitor, and whichever is more accessible wins at retrieval. The original often wins because it was encoded first and fits the model the person built, which is why corrections that supply an alternative explanation work better. The hypercorrection effect (Butterfield & Metcalfe, 2001), in which high-confidence errors are more likely to be fixed once feedback arrives, shows the system is not simply stubborn. A surprising, attended correction can be encoded strongly. Perseverance is what happens when it is not.

3. Evolutionary Origins

Nobody has tested an evolutionary account of belief perseverance directly, so this is reasoning about why a mind might be built this way rather than evidence that it was.

Beliefs have to be stable to be useful. A creature that dropped a belief at every contrary observation could never act. Ancestral evidence was noisy: a watering hole that was safe yesterday is probably safe today even if a predator was seen nearby once. Perseverance is that sensible policy applied to a case it was never built for, where the new datum is a trustworthy report that the old evidence was fake.

Explanation is the brain's main tool for prediction. We store causal models rather than isolated facts, because a model predicts what happens next. The cost is that when the fact is withdrawn, the model, now resting on other things we know, keeps producing the same prediction.

Consistency pays socially. A person whose beliefs change with every rumor is hard to rely on and easy to manipulate. Holding a position builds a reputation for reliability, which is one reason perseverance is stronger when a belief has been voiced or acted on.

Giving up a belief is expensive. Revising one belief means revising the others that depend on it and admitting error. Keeping it is cheaper in the short term even when it is wrong.

4. How This Bias Manifests

4.1. In Everyday Life

• A rumor about a neighbor gets corrected, and you still watch them more closely
• A health scare you read years ago, since debunked, still shapes what you buy
• A first impression formed on someone's bad day persists after you have seen them on good ones
• A teacher told you that you were "not a maths person," was wrong, and you still avoid spreadsheets

4.2. In the Workplace

• A manager's view of an employee formed from one botched presentation survives a year of solid work
• A strategy adopted on a market report later withdrawn keeps running, because everyone has already explained why it makes sense
• Post-mortems that correct the record, after which people keep blaming the original suspect
• A reference check that turns out to have been about someone else with the same name, and the panel's lingering doubt

4.3. In Business and Marketing

• Brand beliefs seeded by an advertising claim later ruled misleading, which customers go on repeating
• Investors who built a thesis around a founder's early story and hold it after the story is retracted
• Product myths ("replace it every year") that survive the manufacturer's own corrections
• Consumers who read a correction and remember only the claim

4.4. In Politics and Media

The political case is where the evidence is richest and most misread. Lewandowsky, Stritzke, Oberauer and Morales (2005) surveyed people in the United States, Germany and Australia during the 2003 Iraq war about claims that had been reported and retracted. American respondents went on using the retracted claims; German and Australian respondents mostly did not, and the difference tracked suspicion about the motives behind the war. Nyhan and Reifler (2010) then reported the backfire effect, which held sway until Wood and Porter (2019) found corrections reducing false beliefs across the board.

Typical patterns:

• A false claim spreads widely, the correction narrowly, and years later much of the public still holds the claim
• Partisans read the same inquiry report and each side finds it vindicates them
• Corrections that restate the original claim prominently and cement it

4.5. In Healthcare

Ignaz Semmelweis showed in 1847 that handwashing with a chlorine solution cut deaths from childbed fever; the profession rejected it for decades because it contradicted the accepted theory of disease, and he died in 1865 without seeing it accepted. Barry Marshall and Robin Warren identified Helicobacter pylori as the cause of most stomach ulcers in the early 1980s, the stress theory persisted for years, and the two shared the Nobel Prize in 2005.

In practice:

• Patients who read a retracted study linking a treatment to harm still refuse the treatment
• Clinicians who keep a diagnosis formed on an early, later-corrected test result
• Parents who accept that a vaccine-autism link was fabricated and still feel uneasy at the appointment

4.6. In Finance and Investing

• An investment thesis formed on a company's early narrative survives the restatement of the figures it rested on
• Traders who keep a position after the analyst report that justified it is withdrawn
• Beliefs about a fund manager's skill formed from a lucky early run and kept through years of average returns
• Clients told a product was "guaranteed" by a salesperson later sanctioned for saying so, who still describe it that way

5. Real-World Case Studies

Case Study 1: The MMR Vaccine and Autism

• Context: In 1998 Andrew Wakefield and co-authors published a paper in The Lancet describing twelve children and suggesting a possible link between the measles, mumps and rubella (MMR) vaccine and autism. Press coverage was extensive.
• What happened: Large epidemiological studies found no link. In 2004 most co-authors withdrew their support after a journalist revealed undisclosed conflicts of interest; in 2010 The Lancet retracted the paper and the UK General Medical Council struck Wakefield off the register. By then MMR uptake in England had fallen from over 90 percent to around 80 percent.
• The bias at work: Parents built an explanation: the timing fits, the vaccine is given around the age autism is noticed, a doctor said so. The retraction removed the evidence and left the explanation. Many parents who accepted that the study was fraudulent still reported unease.
• Consequences: Measles cases rose across Europe and North America, concentrated in under-vaccinated communities. The World Health Organization confirmed in 2019 that the United Kingdom had lost its measles elimination status.
• Lessons learned: A total correction still competes with a story that has had years to settle. Corrections that explain the timing coincidence do better than ones that only say the study was false.

Case Study 2: Weapons of Mass Destruction in Iraq

• Context: The 2003 invasion of Iraq was justified largely by claims that the country held chemical and biological weapons. Searches found none, and the claims were publicly retracted by the governments that had made them.
• What happened: Lewandowsky, Stritzke, Oberauer and Morales (2005) tested whether people in the United States, Germany and Australia still reasoned from war-related claims they knew had been retracted. American participants did; German and Australian participants largely did not, and what separated them was suspicion of the official reasons for the war. Polling years later still found a sizeable minority of Americans believing weapons had been found.
• The bias at work: Those who accepted the case for war had an explanation ready: the regime had used such weapons before, it had obstructed inspectors, officials said the evidence was solid. When the evidence was withdrawn, the explanation kept producing the conclusion.
• Consequences: The belief shaped political debate for a decade and became the test case for the overstated backfire effect when Nyhan and Reifler used a WMD correction in 2010.
• Lessons learned: Whether a correction works depends on whether the person has a reason to want to update. Wood and Porter later showed that even here, clear corrections moved beliefs toward accuracy without restoring them fully.

Historical Example: When Prophecy Fails (1954)

In the autumn of 1954 Leon Festinger, Henry Riecken and Stanley Schachter read that a woman in a Chicago suburb, called "Marian Keech" in their book, was receiving messages from another planet predicting a great flood on December 21. A small group had gathered around her; several had quit jobs or given away possessions. The researchers joined as observers.

The prophecy failed on schedule. The group sat through the night of December 20 waiting for a flying saucer that did not come. In the early hours Keech received a new message: the group's faith had been so strong that the planet had been spared. Members who had avoided reporters now called them. Critics have noted that the observers may have influenced the group, but the pattern has recurred in groups whose predictions fail. Disconfirmation does not remove the belief; it creates a demand for an explanation, and the one reached for keeps the belief alive.

6. The Cost of This Bias

6.1. Personal Costs

• Relationships damaged by impressions that were corrected but never let go
• Years of avoiding a subject or career because of a judgment made on evidence later shown to be wrong
• Health decisions driven by scares since retracted
• Being easy to manipulate: whoever gets a story in front of you first owns part of your belief

6.2. Professional Costs

• Hiring and promotion decisions shaped by corrected but remembered rumors
• Strategies continued past the point where their justification collapsed
• Teams that reach the right conclusion in a post-mortem and act on the old one
• Slow organizational learning, because corrections are filed and the original beliefs are acted on

6.3. Societal Costs

• Public health measures undermined by scares whose evidence was retracted years earlier
• Political debate conducted on claims both sides know were withdrawn
• Wrongful convictions sustained by jurors' first impressions, which instructions to disregard evidence do not fully remove
• A public that cannot be corrected, which makes misinformation cheap to produce and expensive to undo

6.4. Statistical Impact

• In Ross, Lepper and Hubbard (1975), participants told their fake scores were fake still differed in self-ratings of ability; debriefing reduced the gap without closing it, and a process debriefing eliminated it
• Anderson, Lepper and Ross (1980) found stronger perseverance in participants who wrote an explanation before the discrediting
• Chan, Jones, Hall Jamieson and Albarracín's 2017 meta-analysis found that corrections reliably reduce belief in misinformation but rarely return it to the level of people who never saw it
• Wood and Porter (2019), across 52 issues and more than ten thousand participants, found corrections reducing false beliefs in nearly every case and no reliable backfire

7. The Hidden Benefits

Belief perseverance is the cost of something we would not want to give up: a mind whose beliefs have inertia.

• Stability Under Noise: Most contrary observations are noise. A belief system that stays put until contrary evidence accumulates outperforms one that flips on every report.

• Protection Against Manipulation: If a single retraction could fully erase a belief, so could a single false retraction. People who hold on to what they have worked out are harder to spin.

• Persistence Through Early Failure: Alfred Wegener's continental drift was rejected for half a century before the evidence arrived. Barry Marshall drank a culture of H. pylori to prove a point nobody wanted to hear. Holding a belief against the current evidence keeps a correct but unpopular idea alive long enough to be vindicated.

• Psychological Steadiness: A self-concept that any piece of feedback could overturn would be exhausting. Some resistance to revision lets people absorb criticism without collapse.

These benefits hold when the belief was well founded. They become costs when the foundation turns out to have been invented.

8. Self-Assessment: Do You Have This Bias?

8.1. Warning Signs Checklist

• You have learned that a claim you repeated was false, and you still catch yourself believing it
• When a source you trusted is discredited, your opinion of what it criticized does not improve
• You can name a first impression of someone that you never revised despite years of contrary evidence
• You think "the study was flawed, but the basic idea is probably still true" without checking for other support
• You remember the headline of a scare story more clearly than the correction that followed
• When a belief turns out to rest on a single source, you look for reasons it might be true anyway
• You have continued a habit after the advice behind it was withdrawn
• Giving up a belief after a correction feels like it would make you look inconsistent
• You have explained to someone why a thing is true, and found the explanation hard to drop after the thing was shown to be false
• You rate corrections as "technically right" and go on as before

Scoring:

• 0-2 checked: Low susceptibility
• 3-5 checked: Moderate susceptibility
• 6-8 checked: High susceptibility
• 9-10 checked: Very high susceptibility

8.2. Self-Reflection Questions

1. What belief do you hold now that you first picked up from a source you no longer trust? Have you re-examined it?
2. Think of a time you were told that something you had been told was false. What did you do with the explanation you had already built around it?
3. Is there a person about whom you formed a strong early impression that later information should have changed? Did it?
4. When you read a correction, do you remember the correction a month later, or the original claim?
5. Which of your beliefs would you be most reluctant to give up if its original evidence were retracted tomorrow, and why?

8.3. Quick Diagnostic Scenario

Scenario: Six months ago you read a widely shared article reporting that a popular food additive causes memory problems. You cut it from your diet and explained to friends why. Today you learn that the study behind the article has been retracted for fabricated data, and that no other study has found the effect.

How would you respond?

• A) Keep avoiding the additive; the study may have been bad, but where there's smoke there's fire → High susceptibility
• B) Accept that the evidence is gone and resume buying the product, with a lingering unease every time → Moderate susceptibility
• C) Accept the retraction, ask what else you believed on the strength of that article, and tell the friends you warned → Low susceptibility

9. Identifying This Bias in Others

9.1. Behavioral Indicators

• Acknowledges a correction and then argues as though it had not been made
• Treats the retraction of a source as a technicality rather than the removal of the evidence
• Shifts from "the evidence shows" to "everyone knows" when the evidence is withdrawn
• Finds new reasons for the old conclusion within minutes of losing the original one
• Remembers the claim vividly and the correction vaguely

9.2. Conversational Red Flags

Phrases people say when under this bias:

• "Okay, that study was fake, but the underlying idea makes sense."
• "There's no smoke without fire."
• "I know it was retracted, but I still think..."

Types of arguments they make:

• Replacing the withdrawn evidence with a general plausibility argument
• Citing the explanation they built as if it were independent evidence

Questions they avoid asking:

• "If I had never seen the original claim, would I believe this now?"
• "What evidence, apart from the discredited source, do I actually have?"

9.3. Situational Triggers

• The belief was formed from a vivid case rather than from statistics (Anderson 1983 found concrete cases produced stronger perseverance than abstract data)
• The person explained the belief to someone else before the correction arrived
• The correction says only "that was false" and offers no alternative account
• The correction arrives long after the claim, or in fragments
• The person distrusts whoever issued the correction more than whoever made the claim

10. Cognitive Debiasing Strategies

10.1. Immediate Techniques

• Counter-explain: When a belief loses its evidence, spend two minutes explaining how the opposite could be true. Anderson (1982) found this reliably cut perseverance, because it gives the original explanation a competitor.
• Consider the opposite: Lord, Lepper and Preston (1984) showed that asking "would I have judged this evidence the same way if it had pointed the other way?" reduced biased assimilation where instructions to be fair did not.
• Fill the gap: Write down what actually happened. Johnson and Seifert (1994) found that corrections supplying an alternative explanation reduced continued influence far more than bare retractions.
• Name the effect: The process debriefing in Ross, Lepper and Hubbard (1975) worked because participants learned that perseverance exists. Telling yourself "this is the thing where the story outlives the facts" is a small version of the same move.

10.2. Long-Term Strategies

• Keep a retraction ledger: Record beliefs you have had to withdraw, their source, and what you did about the downstream beliefs. Writing them down makes the correction as memorable as the claim.
• Delay explanation: Postpone explaining a striking claim until you have checked it. An unexplained claim is easy to drop; an explained one is not.
• Prefer base rates to stories: Beliefs anchored to a single vivid case are the ones that cling hardest.
• Practice public revision: Say "I used to think X, and I was wrong because Y" out loud and often. Each repetition lowers the social cost of the next one.

10.3. Environmental Design

• Corrections with the same reach as the claim: Route a correction through every channel the original claim used, with a headline of its own.
• Decision records with sources: Log the evidence behind each major decision so that a withdrawn source can be traced to the decisions it supported.
• Prebunking: Warn people about a likely false claim, and the explanation they will build, before it arrives. Anderson's 1982 inoculation condition and the misinformation-inoculation work of the 2010s point the same way.
• Lead with the truth: State the true account first and mention the false claim once, after it, so the truth is what the reader rehearses.

10.4. When to Seek External Input

• When a belief you are acting on came from a source now known to be unreliable, and you cannot say what else supports it
• When your impression of a person is strongly negative and was formed early
• When you notice yourself generating fresh reasons for an old conclusion right after losing the original one

Whom to ask: someone who did not see the original claim. Describe the current evidence without the discredited part and ask what they would conclude. If their answer differs from yours, the difference is the residue of the retracted evidence.

11. Practical Exercises

Exercise 1: The Retraction Ledger

• Objective: Make corrections as memorable as the claims they correct
• Time required: 10 minutes per entry, 20 minutes monthly review
• Materials needed: A notebook or a note on your phone
• Difficulty level: Beginner
• Instructions:
  1. Each time you learn that something you believed was false, write down the belief, where you got it, and the correction
  2. List the other beliefs and decisions that depended on it
  3. For each, note whether it still has independent support
  4. Decide what, if anything, you will do differently
  5. Once a month, reread the ledger and check whether the retracted beliefs have crept back
• Reflection questions:
  • Which retracted beliefs were hardest to let go of, and what did they have in common?
  • Did any retracted belief reappear in your thinking after you wrote it down?
• Frequency: Ongoing, with a monthly review

Exercise 2: The Counter-Explanation Drill

• Objective: Build the habit of giving every explanation a competitor
• Time required: 15 minutes
• Materials needed: Paper and pen
• Difficulty level: Intermediate
• Instructions:
  1. Pick a belief you hold about why something is the case (why a colleague behaves as they do, why a product sells)
  2. Write your explanation in three or four sentences
  3. Write an equally detailed explanation for the opposite conclusion, as if you had to defend it
  4. List what evidence would distinguish the two
  5. Note how confident you feel in the original belief now compared with before
• Reflection questions:
  • Was the counter-explanation harder to produce than you expected?
  • What evidence do you actually have beyond the original story?
• Frequency: Weekly for two months, then whenever a belief feels suspiciously settled

Exercise 3: The Belief Autopsy

• Objective: Trace a long-held belief to its source and test whether the source still stands
• Time required: 45 minutes
• Materials needed: Paper, internet access
• Difficulty level: Intermediate
• Instructions:
  1. Choose a belief you have held for years about health, money, a group of people, or how the world works
  2. Write down when and where you first came to believe it
  3. Look up the original source. Is it still regarded as sound, or has it been retracted or contradicted?
  4. Sort every reason you now hold for the belief into "independent evidence" and "explanations I built after accepting the claim"
  5. If the independent column is empty and the source is discredited, write out what you would believe meeting the question for the first time today
• Reflection questions:
  • How much of your confidence came from the source, and how much from your own explanation?
  • What would it take to actually change the belief, as opposed to acknowledging that you should?
• Frequency: Quarterly, one belief each time

Daily Practice

The "Where Did I Get That?" Question

Once a day, when you notice yourself stating something as fact, pause and ask where it came from. If the answer is a single source, note whether you have ever checked it.

• Suggested duration: 2 minutes
• Best time of day: Whenever you catch yourself passing on a claim
• How to track progress: Tally claims you could trace to a verified source against claims you could not

Weekly Challenge

Correct Yourself in Public

Each week, find one thing you told someone that you now know to be wrong or unsupported, and go back to them with the correction and what actually appears to be true.

• Expected outcomes after 4 weeks: Lower discomfort with revision; a clearer sense of which beliefs rest on single sources
• Journaling prompts for reflection:
  • Which correction was hardest to deliver, and was that about the facts or about looking inconsistent?
  • Did the person you corrected update, or show perseverance of their own?

12. For Specific Audiences

For Leaders and Managers

• Impressions of People: First impressions of staff formed from early incidents persist after the record has changed. Review your view of each report against the last year's evidence, not the first month's.
• Corrections in Organizations: When a claim that drove a decision is withdrawn, issue the correction through every channel the claim used, and name the decisions it affects.
• Explanation Discipline: Ask teams to hold off on explaining a surprising result until it is verified. A team that has explained a result will defend it after it is overturned.

For Parents and Educators

• Age-Appropriate Explanation: "Sometimes we find out that a thing we believed came from a mistake. Our brain likes to keep the idea anyway, because we already made up a reason for it. When that happens, try making up a reason for the opposite."
• Misconceptions in Learning: Students arrive with intuitive theories (heavier objects fall faster, for instance) that survive instruction because it corrects the fact but not the explanation. Having students predict, observe, then explain the gap works better than stating the right answer.
• Modeling Revision: Say "I was wrong about that, and here is what I think now" in front of children. They learn to update by watching it done without drama.

For Healthcare Professionals

• Diagnostic Anchors: A working diagnosis formed on an early test result persists after the result is corrected. When a result is revised, restate the differential from scratch rather than editing the old one.
• Patient Beliefs: Patients who accept that a scare was based on fraud may still feel it. Supply the alternative account of the coincidence they noticed, not only the retraction.
• Communicating Corrections: Lead with the true account, state the false claim once, and explain why people believed it. Corrections that lead with the myth refresh it.

For Financial Professionals

• Thesis Hygiene: Record the specific evidence behind each investment thesis. When a piece of it is withdrawn, re-derive the thesis from what remains rather than asking whether it "still feels right."
• Client Beliefs: Clients hold convictions about products and managers seeded by sources since discredited. Ask where a conviction came from before arguing with it.
• Pre-Commitment: Decide in advance what evidence would cause you to exit a position, and write it down. A written trigger survives the explanations you will generate later.

13. Interactions with Other Biases

Biases That Amplify This One

Biases That Counteract This One

Common Bias Chains

Anchoring → Confirmation Bias → Belief Perseverance → Continued Influence Effect

A first claim sets the anchor. Confirmation bias collects supporting observations and the person builds an explanation. When the claim is retracted, the explanation and the collected observations keep the belief alive, and the retracted claim goes on shaping inferences even when the person can correctly state that it was retracted.

Interrupting the chain: The cheapest intervention is early: withhold explanation until the claim is checked. The most effective one after the fact is counter-explanation together with an alternative account that fills the gap.

14. Cultural Perspectives

Direct cross-cultural research on belief perseverance is sparse, so this section is partly inference. The one well-documented national difference, from Lewandowsky and colleagues' 2005 Iraq study, was not cultural in the usual sense: Americans kept using retracted claims while Germans and Australians did not, and the gap was explained by suspicion of the claims' source. Trust in the source, more than nationality, predicted perseverance.

Work on reasoning styles offers a plausible moderator. Peng and Nisbett (1999) found that Chinese participants were more comfortable than Americans with holding apparently contradictory propositions, a style they called dialectical thinking. A thinker who tolerates contradiction may be less driven to build a single explanation, or may simply hold the belief and its correction side by side. The literature does not yet say which. Cultures also differ in the cost of visible revision: where changing one's mind reads as weakness, perseverance has a social reward on top of its cognitive one.

15. Myths and Misconceptions

16. Expert Insights

"The human understanding when it has once adopted an opinion (either as being the received opinion or as being agreeable to itself) draws all things else to support and agree with it." — Francis Bacon, Novum Organum (1620)

"A man with a conviction is a hard man to change. Tell him you disagree and he turns away. Show him facts or figures and he questions your sources. Appeal to logic and he fails to see your point." — Leon Festinger, Henry Riecken and Stanley Schachter, When Prophecy Fails (1956)

"A new scientific truth does not triumph by convincing its opponents and making them see the light, but rather because its opponents eventually die, and a new generation grows up that is familiar with it." — Max Planck, Scientific Autobiography (1950)

Paraphrasing Lewandowsky, Ecker, Seifert, Schwarz and Cook (2012): a retraction on its own rarely works, because it leaves a gap in the person's model of events; the correction that works supplies an alternative account, avoids repeating the myth more than necessary, and comes from a source the person trusts. — Summary of Misinformation and Its Correction, Psychological Science in the Public Interest

17. Key Takeaways

1. The belief outlives its evidence: Retracting a claim removes the evidence but leaves the belief, which by then rests on an explanation the person built.

2. Explanation is the mechanism: People who explain a relationship before it is discredited cling to it hardest, and explaining a relationship creates belief in it even without evidence.

3. It is not confirmation bias and not backfire: Confirmation bias is about gathering evidence; perseverance is about surviving its loss. Backfire has not held up in large tests.

4. Corrections work, partly: Meta-analyses find that debunking reliably reduces false beliefs without returning them to baseline. The residue is the perseverance.

5. Counter-explanation is the best-tested remedy: Explaining how the opposite could be true, or receiving an alternative account that fills the gap, cuts perseverance in a way bare retractions do not.

6. Trust in the source matters more than nationality: Whether a retraction is accepted depends on how the person regards whoever made the claim and whoever is correcting it.

7. Knowing about the effect helps: The process debriefing that eliminated perseverance in 1975 worked by telling people the mechanism. Naming it when it happens to you is a small version of the same cure.

18. Further Resources

Academic Papers

• Ross, L., Lepper, M. R., & Hubbard, M. (1975). Perseverance in self-perception and social perception: Biased attributional processes in the debriefing paradigm. Journal of Personality and Social Psychology, 32(5), 880-892.
• Anderson, C. A., Lepper, M. R., & Ross, L. (1980). Perseverance of social theories: The role of explanation in the persistence of discredited information. Journal of Personality and Social Psychology, 39(6), 1037-1049.
• Lord, C. G., Ross, L., & Lepper, M. R. (1979). Biased assimilation and attitude polarization: The effects of prior theories on subsequently considered evidence. Journal of Personality and Social Psychology, 37(11), 2098-2109.
• Anderson, C. A. (1982). Inoculation and counterexplanation: Debiasing techniques in the perseverance of social theories. Social Cognition, 1(2), 126-139.
• 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.
• Wood, T., & Porter, E. (2019). The elusive backfire effect: Mass attitudes' steadfast factual adherence. Political Behavior, 41(1), 135-163.

Books

• Nisbett, R. E., & Ross, L. (1980). Human Inference: Strategies and Shortcomings of Social Judgment. Prentice-Hall.
• Festinger, L., Riecken, H. W., & Schachter, S. (1956). When Prophecy Fails. University of Minnesota Press.

Book Chapters

• Ross, L., & Anderson, C. A. (1982). Shortcomings in the attribution process: On the origins and maintenance of erroneous social assessments. In D. Kahneman, P. Slovic, & A. Tversky (Eds.), Judgment Under Uncertainty: Heuristics and Biases (pp. 129-152). Cambridge University Press.
• Anderson, C. A. (2007). Belief perseverance. In R. F. Baumeister & K. D. Vohs (Eds.), Encyclopedia of Social Psychology (pp. 109-110). Sage.

19. Summary Card

20. Glossary of Terms Used

21. Discussion Questions

For book clubs, classrooms, or self-reflection:

1. The 1975 participants had no stake in whether they were good at judging suicide notes, and still kept the belief. How much of your own resistance to correction is about ego, and how much is about the way explanation works?

2. Corrections reduce false beliefs but do not erase them. Should news organizations and public health bodies change how they issue corrections, and how?

3. Think of a scientific or medical idea that was overturned during your lifetime. How long did it take for people around you to stop reasoning from the old idea?

4. The same inertia that keeps a discredited belief alive also kept continental drift and the bacterial theory of ulcers alive through years of rejection. Is there a way to tell the two cases apart from the inside?

5. If delaying explanation until a claim is verified is the cheapest defense, what would it look like to build that delay into a classroom, a newsroom, or a team?

Laas opgedateer 15 Oktober 2026