Definition · Plain-language
Confirmation bias
Confirmation bias is the tendency to search for, interpret, favour and recall information in ways that confirm one’s existing beliefs or hypotheses.
The step most authors miss
Doing CRediT right? Don’t stop at the statement.
A CRediT statement credits you inside one paper. The recognition CRediT was built for happens when those roles are tied to you, persistently. Sign in with your ORCID — free — and claim your CRediT contributions on casrai.org, the home of the standard. They become a verified, portable part of your identity, not a line that disappears into one PDF.
Free: claim your contributions, then export a journal-ready CRediT statement, schema.org structured data, JATS XML, CSV or BibTeX — and preview your public profile. A membership publishes that profile publicly and verifies the journals you serve.
Wason’s 2-4-6 task and three forms of the bias
Peter Wason’s 1960 experiment asked participants to discover a rule that generated number triples (the rule was "any ascending sequence"). Participants generated triples to test hypotheses and consistently proposed only confirming examples — few tried the falsifying approach of testing outside their guessed pattern. This demonstrated confirmation bias in controlled conditions. The bias operates through three distinct mechanisms: selective exposure (seeking out information that confirms the belief), biased interpretation (interpreting ambiguous evidence as confirmatory), and selective recall (more easily remembering instances that confirm than those that disconfirm).
Myside bias, motivated reasoning and echo chambers
Jonathan Haidt and colleagues have documented myside bias — the tendency to generate and evaluate arguments in ways that favour one’s own side. Ziva Kunda (1990) formalised motivated reasoning: when people want to reach a particular conclusion, they selectively access memories and apply inferential rules in ways that support it, while maintaining the subjective feeling of being objective. Echo chambers in social media algorithmically amplify confirmation bias by preferentially showing users content that matches their existing views, reducing exposure to disconfirming perspectives and reinforcing in-group beliefs.
Countering confirmation bias: pre-mortems, falsificationism and devil’s advocates
Several evidence-based techniques reduce confirmation bias. Pre-mortems (Klein 1993) require decision-makers to imagine that a project has failed and then work backwards to identify what went wrong, surfacing disconfirming considerations before a decision is locked in. Karl Popper’s falsificationist science institutionalises the antidote: scientists are trained to seek evidence that could disprove a hypothesis rather than merely confirm it. Organisations use devil’s advocate roles — appointing someone to argue against the prevailing view — to prevent groupthink. Structured analytic techniques used in intelligence analysis formalise these approaches to reduce individual and collective confirmation bias.
Key facts
At a glance
- Definition: favouring information that confirms existing beliefs
- Early identification: Francis Bacon (Novum Organum, 1620)
- Wason 1960: 2-4-6 task demonstrated confirmation bias experimentally
- Three forms: selective exposure, biased interpretation, selective recall
- Myside bias: Haidt — generating arguments that favour one’s own position
- Motivated reasoning: Kunda (1990) — goal-directed information processing
- Echo chambers: algorithmic amplification of confirmation bias in social media
- Counters: pre-mortems, Popperian falsificationism, devil’s advocates
Common misconceptions
What people often get wrong
Often heard: Confirmation bias only affects people who are unintelligent or uneducated.
Actually: Research consistently shows that confirmation bias affects people across all intelligence and education levels. Smarter people are sometimes better at rationalising their existing beliefs, which can make them more susceptible rather than less.
Often heard: Looking for supporting evidence is always confirmation bias.
Actually: Seeking evidence is not the fallacy — the bias lies in exclusively or disproportionately seeking confirming evidence while ignoring or dismissing disconfirming evidence. A balanced search strategy that actively looks for both is the appropriate counter.
Often heard: Confirmation bias can be fully eliminated with training.
Actually: Research suggests confirmation bias can be reduced but not eliminated by training, awareness and structured techniques. It is deeply rooted in normal cognitive processes, and even people aware of the bias show it under time pressure or when emotionally invested in an outcome.
Common questions
FAQ
What is the clearest experimental demonstration of confirmation bias?+
Peter Wason’s 2-4-6 task (1960) is the classic demonstration. Participants had to discover the rule governing a sequence by generating their own triples. Most tested only confirming examples of their guessed rule, rarely attempting disconfirming tests, even though disconfirming tests are more informative.
How does confirmation bias affect scientific research?+
It can lead researchers to design studies that are more likely to confirm their hypothesis, interpret ambiguous data favourably, and publish positive results while leaving null results unpublished (contributing to publication bias). Peer review, pre-registration of hypotheses, and replication requirements are institutional defences against it.
What is the difference between confirmation bias and motivated reasoning?+
Motivated reasoning (Kunda 1990) is the broader process of goal-directed thinking, where people reason toward a desired conclusion. Confirmation bias is one specific mechanism through which motivated reasoning operates — the selective search for and interpretation of confirming information. Motivated reasoning can also use other tactics, such as selective application of logical standards.
Going deeper
