Confirmation Bias Explained: Why You Only See What You Expect

About this psychology article

You already believe you're not that susceptible to confirmation bias. That belief is, itself, a textbook example of exactly how confirmation bias works.

Confirmation bias might be the most widely known cognitive bias in psychology — everyone's heard of it, everyone assumes they mostly avoid it, and almost everyone is wrong about that second part. It's one of the most well-documented, most consistently replicated findings in cognitive psychology, and understanding the actual research behind it — rather than the loose, everyday version of the term — makes it considerably harder to fool yourself with it.

Confirmation bias is the tendency to search for, interpret, favor, and recall information in ways that confirm your existing beliefs, while giving disproportionately less attention to information that contradicts them. First formally demonstrated by psychologist Peter Wason in a landmark 1960 experiment, confirmation bias operates across how we search for information, how we interpret ambiguous evidence, and what we remember afterward — and research consistently finds that people believe they're less susceptible to it than they actually are.

The Experiment That Discovered It

Psychologist Peter Wason formally documented confirmation bias in a 1960 study using what became known as the 2-4-6 task . Participants were told that the sequence "2, 4, 6" followed a rule the experimenter had in mind, and were invited to propose their own three-number sequences, receiving a simple yes or no on whether each fit the hidden rule, before eventually guessing the rule itself.

The actual rule was strikingly simple: any three numbers in ascending order. But most participants formed an early, more specific hypothesis — commonly something like "even numbers increasing by two" — and then proceeded to test only sequences that would confirm that guess, such as "8, 10, 12" or "20, 22, 24." Very few participants tried sequences specifically designed to disprove their working hypothesis, such as "1, 2, 3," which fit the actual rule but would have immediately revealed their assumed rule was too narrow. As a direct result, a majority of participants confidently announced an incorrect rule, despite having every opportunity to test their assumption properly and having received accurate yes-or-no feedback throughout.

The Three Ways Confirmation Bias Actually Operates

Confirmation bias isn't a single mental glitch — psychologists generally describe it as operating through three related but distinct processes:

Mechanism

What It Looks Like

Biased search

Actively seeking information likely to confirm an existing belief, while avoiding or simply not looking for sources likely to challenge it

Biased interpretation

Reading genuinely ambiguous evidence as supporting a pre-existing view, even when it could reasonably support the opposite conclusion

Biased memory

Recalling confirming instances more easily and vividly than disconfirming ones, gradually skewing a person's sense of the overall evidence

Where It Shows Up

Political and media consumption. Well-documented research on selective exposure consistently finds that people gravitate toward news sources and commentary aligned with their existing views. It's worth adding real nuance here, though: the popular narrative that algorithmic "echo chambers" and "filter bubbles" trap people in completely one-sided information environments has itself been somewhat overstated in some research. Studies by researchers including Andrew Guess and colleagues examining actual browsing behavior have found more cross-cutting exposure to opposing viewpoints than the strongest versions of the filter-bubble narrative suggest — the deeper issue often isn't a total absence of opposing information, but how that information gets interpreted and dismissed once encountered, which is precisely where confirmation bias itself operates most powerfully.

Scientific research. Confirmation bias affects trained scientists, not just casual reasoning — a genuine, actively discussed problem in research methodology, connected to broader concerns about motivated reasoning and selective reporting of results. It's part of why practices like pre-registration of hypotheses, blind data analysis, and adversarial collaboration between researchers who disagree have gained ground as partial correctives in various fields.

Medical diagnosis. Diagnostic error research has documented cases where physicians anchor on an initial impression and subsequently interpret new symptoms or test results as confirming that first diagnosis, rather than genuinely weighing alternative explanations — a well-studied contributor to diagnostic error in clinical decision-making literature.

Relationships and jealousy. Someone suspicious of a partner may notice and remember ambiguous behaviors that seem to confirm the suspicion, while barely registering the far larger number of ordinary, unremarkable interactions that don't fit the narrative — a pattern that can escalate suspicion well beyond what the actual balance of evidence supports.

Financial decisions. Investors frequently seek out information confirming that a prior investment decision was sound, while discounting or avoiding warning signs — a pattern closely related to the psychological mechanisms explored in why people defend decisions and relationships that have already hurt them , since admitting a past decision was wrong carries its own psychological cost.

Why This Happens: Three Explanations

Cognitive efficiency. Searching for and processing confirming evidence is simply less mentally effortful than systematically testing every possible alternative — a form of cognitive shortcut that trades thoroughness for speed, useful in many everyday situations but poorly suited to genuinely uncertain or high-stakes questions.

Identity and self-esteem protection. Confirmation bias overlaps significantly with the self-protective mechanisms explored in research on cognitive dissonance and self-deception — maintaining existing beliefs protects a person's broader sense of being a rational, consistent person who doesn't need to admit error, which can make genuinely disconfirming evidence feel like a threat to more than just the specific belief in question.

An evolved feature, not just a flaw. A genuinely provocative academic theory, the argumentative theory of reasoning proposed by cognitive scientists Hugo Mercier and Dan Sperber in 2011, suggests human reasoning may have evolved primarily as a tool for social persuasion and winning arguments, rather than as a pure truth-seeking mechanism. Under this framework, confirmation bias isn't simply a bug in otherwise objective reasoning — it's closer to a predictable feature of a reasoning system built more for constructing convincing arguments for a position than for neutrally evaluating all available evidence.

The Bias You Probably Have About This Bias

Here's a genuinely humbling finding worth sitting with directly: research on what psychologist Emily Pronin termed the bias blind spot consistently finds that people readily recognize cognitive biases, including confirmation bias, in other people's reasoning, while rating themselves as considerably less susceptible than average. This isn't limited to less self-aware individuals — the effect has been documented across education levels and even among people specifically trained in recognizing cognitive biases, suggesting that simply knowing about confirmation bias intellectually doesn't automatically make a person immune to it in their own reasoning.

How to Actually Counteract It

Genuine debiasing research is honest about a difficult truth: confirmation bias is hard to eliminate, and no single technique fully removes it. That said, several approaches have real, if modest and inconsistent, research support.

The "consider the opposite" strategy. Deliberately asking, for any belief you hold with some confidence, "what specific evidence would change my mind?" has shown some ability to reduce bias in experimental settings, by forcing genuine engagement with disconfirming possibilities rather than defaulting to confirming ones.

Actively seeking out strong opposing arguments , rather than the weakest or most easily dismissed version of an opposing view, forces more genuine engagement with a perspective's real substance rather than a convenient straw man.

Structured organizational techniques , including formally assigning someone to argue the opposing case (sometimes called devil's advocacy) or the more rigorous "red teaming" approach used in intelligence analysis and business strategy, build institutional counterpressure against confirmation bias into a decision process, rather than relying purely on individual willpower or self-awareness. These structural approaches tend to outperform purely individual debiasing efforts specifically because they don't depend on any one person successfully noticing their own bias in the moment.

Blind or masked analysis in research and decision-making contexts — evaluating evidence without knowing which conclusion it supports until after the evaluation is complete — removes some of the opportunity for biased interpretation to operate in the first place.

Common Myths About Confirmation Bias

"Only closed-minded or less intelligent people fall for it." Research consistently finds the opposite in some respects — higher cognitive ability doesn't reliably protect against confirmation bias, and some studies suggest more cognitively skilled individuals can actually construct more sophisticated-sounding justifications for pre-existing beliefs, making the bias harder to detect rather than less present.

"If I actively try to be objective, I won't be affected by it." Simply intending to be objective has limited demonstrated effect on its own. The bias blind spot research specifically shows that awareness and good intentions don't reliably translate into reduced bias — what tends to help more are concrete structural techniques, like deliberately seeking disconfirming evidence, rather than good intentions alone.

"Confirmation bias means people are being deliberately dishonest." The bias generally operates below conscious awareness — biased search, interpretation, and memory typically don't feel like active dishonesty from the inside. Most people experiencing confirmation bias genuinely believe they're weighing evidence fairly, which is part of what makes the bias so persistent and difficult to self-correct.

"Confirmation bias is a modern problem caused by social media and algorithms." While digital platforms can amplify certain aspects of the pattern, confirmation bias was documented experimentally decades before the internet existed, and has been observed across cultures and historical periods well predating any algorithmic influence — it's a basic feature of human cognition, not a byproduct of any particular technology.

Confirmation Bias vs. Motivated Reasoning

Dimension

Confirmation Bias

Motivated Reasoning

Scope

The specific cognitive mechanics of how information is searched for, interpreted, and recalled

The broader phenomenon of a desired conclusion shaping the entire reasoning process

Role

One of the primary mechanisms

The wider "why" behind biased reasoning

Relationship

The "how"

The "why" that confirmation bias helps carry out

These two terms are closely related and often used interchangeably, but they're worth distinguishing. Confirmation bias more specifically describes the cognitive mechanics of how people search for, interpret, and recall information. Motivated reasoning is the broader phenomenon in which a person's desire for a particular conclusion to be true shapes their entire reasoning process toward that outcome, often unconsciously. Confirmation bias is generally understood as one of the primary mechanisms through which motivated reasoning actually operates.

Understand Your Own Thinking Patterns

Given how consistently people underestimate their own susceptibility to this bias, an outside, structured look tends to be more revealing than self-reflection alone — that's precisely what makes biases biases. The Cognitive Biases Assessment on PsychOS gives you a personalized breakdown of your reasoning and decision-making patterns, along with practical insight into where confirmation bias is most likely showing up in your everyday thinking.

→ Take the Cognitive Biases Assessment

Takes about 5 minutes. No account required.

Frequently Asked Questions

What is confirmation bias?

Confirmation bias is the tendency to search for, interpret, favor, and recall information in ways that confirm existing beliefs, while giving disproportionately less consideration to contradicting evidence. It was first formally demonstrated by psychologist Peter Wason in a 1960 experiment.

What was Wason's original confirmation bias experiment?

In Wason's 1960 "2-4-6 task," participants tried to guess a hidden rule behind the number sequence "2, 4, 6." Most formed an overly specific hypothesis and then tested only sequences that would confirm it, rather than sequences designed to disprove it, causing most participants to confidently guess the wrong rule despite the actual rule being much simpler.

Are echo chambers on social media as extreme as commonly believed?

Research findings are more mixed than the strongest popular narrative suggests. Some studies examining actual browsing behavior have found more exposure to opposing viewpoints than a total "filter bubble" would predict, suggesting the deeper issue is often how people interpret and dismiss opposing information they do encounter, rather than a complete absence of it.

What is the "bias blind spot"?

The bias blind spot, a term coined by psychologist Emily Pronin, describes the well-documented tendency for people to readily recognize cognitive biases in others while rating themselves as considerably less susceptible than average — an effect found even among people specifically trained to recognize these biases.

How can you actually reduce confirmation bias in your own thinking?

Approaches with some research support include deliberately asking what evidence would change your mind, actively seeking out the strongest version of opposing arguments, and using structured techniques like devil's advocacy or red teaming in group decision-making. No single technique fully eliminates the bias, but these approaches can meaningfully reduce its influence.