Belief Perseverance.

01Overview

Belief perseverance is the discovery that conclusions outlive their reasons. Ross, Lepper and Hubbard's experiments showed that participants given entirely fabricated feedback about their performance continued to trust the implied self-assessment even after being told the feedback was invented for the study. Anderson, Lepper and Ross then demonstrated the social version: a belief about what makes a successful firefighter, constructed from two cooked case histories, survived the confession that both cases were fabricated. The findings undermine the intuitive model in which beliefs are supported by evidence like a building by foundations; remove the foundations and the building should fall. Instead, the belief persists as why — a causal narrative that once made sense of data, and continues to operate autonomously after the data is withdrawn. This is the mechanism behind dismissed evidence that nonetheless shapes policy, discredited accusations that still stain, and debunked claims whose 'something in it' feeling refuses to leave.

02Key theorists

The 1975 Stanford experiments — 'perseverance in self-perception and social perception' — created false success and failure feedback, manipulated debriefing almost to the level of confession, and found that subjects' self-evaluations and predictions had shifted permanently. The 1980 follow-up showed that even after the evidence was explicitly retracted, the constructed belief drove later judgements, and it survived even when participants were given a general warning about the fallibility of evidence. Kruglanski's lay epistemics supplied the cognitive architecture: a conclusion, once formed, becomes its own epistemic support. Education research later confirmed the phenomenon in classrooms (misconceptions persisted after correction) and it became central to understanding misinformation, where corrections famously fail to un-believe.

03How it works

1. 01Conclusion-memory split: the mind stores the belief; the source gets lost, discredited, or forgotten
2. 02Causal-narrative persistence: a generated explanation for a finding keeps the finding alive
3. 03Post-retraction reinforcement: the effort of understanding a claim can implant it even when the claim is marked false
4. 04Selective evidence recycling: the belief survives by recruiting new support from memory
5. 05Inference leakage: the discredited claim's implications — about the self, the accused, the policy — remain live
6. 06Counterfactual effort: imagining a world where the false thing occurred strengthens familiarity with it

04Where it shows up

Juries influenced by evidence stricken from the record; careers damaged by accusations later found false; policies built on studies later retracted; patients who recall health advice that was already corrected; historical 'truths' preserved in national memory long after the archive was disproved.

05The propaganda link

The Ministry of Truth's revision worked because it needed no agreement in the moment — beliefs already in the citizen's head kept operating through habit and memory, while the corrected record merely finished the job. Belief perseverance is the passive collaborator in every attempt to rewrite events: conclusions the citizen no longer defends keep functioning as truths and the state's retractions are never quite believed, even when accepted.

06How to resist

• Attack the inference, not just the information: correct the implications explicitly, not only the claim
• Replace the narrative: supply a causal account the mind can substitute, because a bare retraction leaves the old account running
• Track sources on the conclusion itself, so it cannot outlive its base unobserved
• Assume a correction is not finished until the belief's downstream uses have been addressed
• For institutions: evaluate decisions later, with the claim's status reintroduced — memory of a policy rarely carries its retraction