What Is Confirmation Bias? Definition, Examples and How to Overcome It

You are sure your new colleague dislikes you. Over the next week you notice every short reply and every skipped greeting. You don’t notice the friendly comment on Tuesday or the coffee she offered you. By Friday, you have a folder of “proof”.

That is confirmation bias at work. It is one of the best-documented patterns in psychology, and once you know what it looks like, you will see it everywhere: in arguments, in the news, in your own most confident opinions.

What is confirmation bias?

Confirmation bias is the tendency to search for, interpret, favour and remember information in a way that confirms what we already believe, while giving less weight to information that challenges it.

A widely cited review by psychologist Raymond Nickerson described it as a “ubiquitous phenomenon in many guises”, meaning it appears in science, law, medicine, politics and everyday life. It is not lying or stupidity. It is how a mind with limited time and attention takes shortcuts.

It belongs to a larger family of thinking shortcuts. For the wider picture, see what is cognitive bias.

Where confirmation bias operates

It helps to separate three stages:

1. Searching. We look for evidence that agrees with us. If you believe a diet works, you Google “benefits of the diet”, not “problems with the diet”.
2. Interpreting. We read ambiguous evidence in our favour. The same facts seem convincing when they support us and flawed when they don’t.
3. Remembering. We recall supporting examples more easily than contradicting ones. Memory is reconstructive, so it can quietly favour our story (see how memory rebuilds the past).

What the research shows

The 2-4-6 task

In a classic 1960 experiment, Peter Wason gave people the number sequence 2, 4, 6 and asked them to discover the rule behind it by proposing their own sequences. Most people guessed “numbers going up by two” and then tested only sequences that fit it (8, 10, 12), rather than trying ones that could prove it wrong. The actual rule was simply “any ascending numbers”, so many people stayed confidently wrong. The experiment is often used to show how easily we test ideas by looking for confirmation.

Biased assimilation

In a well-known 1979 study, Lord, Ross and Lepper showed people with opposing views on the death penalty the same mixed evidence. Instead of moving closer together, participants rated evidence that supported their side as more convincing and dismissed the rest, and many ended up more convinced of their original view. The same information can polarise rather than unite.

Self-fulfilling questions

Research by Snyder and Swann found that when people were asked to find out whether someone was, for example, extraverted, they tended to ask questions that would confirm it, and this influenced how the other person behaved. Our expectations can shape the very evidence we collect.

Related biases and effects

Confirmation bias rarely works alone. Several neighbouring effects are often confused with it:

• Belief perseverance. Beliefs tend to survive even after the evidence for them is removed. In a classic set of experiments, people kept believing a made-up feedback story about themselves even after being told it was fabricated.
• Myside bias. This is the tendency to evaluate evidence and generate arguments in favour of your own side. Research by Stanovich and colleagues suggests it is largely independent of intelligence, so cleverness is no guarantee against it.
• Halo effect. One good (or bad) trait colours how we judge everything else about a person, and we then notice evidence that fits the impression.
• Anchoring. The first number or impression we meet pulls later judgements toward it, and we then look for reasons it was right.
• The “backfire effect”. The idea that correcting a false belief makes people believe it more became popular, but later, larger studies found it to be rare. In most cases people update at least a little when given a clear correction, so do not assume that challenging someone is pointless.

Is testing for “yes” always irrational?

Not quite. Researchers such as Klayman and Ha pointed out that looking for cases that fit a hypothesis (a “positive test strategy”) is often a reasonable, efficient way to learn, especially when the thing you are looking for is rare. The bias becomes a problem when we only do that, when we never ask what would prove us wrong, or when the stakes are high.

Everyday examples

In relationships

If you think your partner doesn’t care, you notice forgotten chores and miss the small kindnesses. If you think you’re falling in love, you overlook red flags. This is one reason jealousy can feel so convincing: every ambiguous text becomes “evidence”.

At work

A manager who decides early that someone is a “star” notices their wins and excuses their mistakes. Another who decides someone is “difficult” does the opposite. Hiring decisions often suffer the same way.

In health

People with health worries can read symptoms into every twinge, while others dismiss real warning signs because they are “sure it’s nothing”. See health anxiety for how this feeds worry.

In the news and online

Algorithms show us more of what we already click. Feeds full of agreement feel like consensus, which makes our views seem more common and more correct than they are.

About ourselves

If you believe “I’m bad at this”, you remember failures and discount successes. That is a core mechanism of negative self-talk and of imposter syndrome.

Why does it happen?

Researchers point to several overlapping reasons:

• Cognitive efficiency. Checking every belief against all evidence is exhausting. Confirming is faster.
• Emotional comfort. Being wrong is unpleasant. Contradictory evidence creates tension that we tend to resolve in our favour (see cognitive dissonance).
• Identity. Beliefs tied to who we are, our group or our values are defended more fiercely.
• Social reward. Agreement feels good, and we surround ourselves with people who share our views.
• Fast versus slow thinking. Kahneman’s account of fast, intuitive thinking (System 1) explains how first impressions quietly steer slower reasoning (System 2).

Is confirmation bias always bad?

Not always. It saves time, helps us keep a stable sense of the world and sometimes protects motivation. The problem arises when stakes are high or beliefs are wrong: in medical diagnosis, legal decisions, investing, relationships and public debate.

7 ways to reduce confirmation bias

1. Ask “What would change my mind?”

Before you gather evidence, decide what kind of evidence would prove you wrong. If the answer is “nothing”, you are dealing with a belief, not a conclusion.

2. Actively seek disconfirming evidence

Search for the strongest case against your view. Try “why [my belief] might be wrong”. Aim to be able to state the opposing view so well that its supporters would agree with your summary.

3. Consider the opposite

Studies on debiasing suggest that explicitly asking “What are reasons my judgement might be wrong?” can reduce one-sided thinking. Write down two or three.

4. Slow down on important decisions

Bias thrives under speed and stress. Give big decisions a night. Write pros and cons before checking with people who agree with you.

5. Diversify your sources and people

Read outlets you disagree with. Ask a trusted friend to challenge you. The goal isn’t to adopt their view but to test yours.

6. Separate observation from interpretation

Write what actually happened, then what you concluded. “She didn’t say hello” is an observation; “she dislikes me” is an interpretation. Then list other explanations: busy, tired, distracted, shy.

7. Keep score honestly

When you make a prediction, write it down and check later. A record protects you from remembering only your hits.

A worked example

You believe your business idea is brilliant.

• Bias in action: you talk mainly to supportive friends, read success stories, and treat negative feedback as “they just don’t get it”.
• Debiasing: you ask three strangers in your target market for blunt feedback, list the top three reasons the idea could fail, and set a clear test: “If fewer than 10 of 50 people sign up, I’ll rethink.”
• Result: you may still proceed, but with better information and a plan for being wrong.

Common mistakes when fighting bias

• Thinking you are immune. Almost everyone believes they are more objective than others.
• Only spotting it in other people. The useful skill is noticing it in yourself.
• Overcorrecting into “everything is uncertain”. The aim is calibrated confidence, not endless doubt.
• Confusing more information with better information. Extra evidence can simply feed the bias if you filter it.
• Treating disagreement as attack. Curiosity works better than winning.

Frequently asked questions

What is confirmation bias in simple terms?

It is the habit of noticing and believing information that agrees with what you already think, and overlooking or dismissing information that doesn’t.

What is an example of confirmation bias?

Believing a colleague dislikes you and then noticing only the moments that seem to prove it, while overlooking friendly moments.

Is confirmation bias the same as cognitive dissonance?

They are related but different. Cognitive dissonance is the discomfort we feel when our beliefs and actions or evidence clash; confirmation bias is one way we avoid or reduce that discomfort.

Can you get rid of confirmation bias?

Not entirely. It is built into normal thinking. You can reduce its effects with deliberate habits like seeking disconfirming evidence and slowing down.

Does intelligence protect against it?

Not reliably. Intelligent people can be skilled at building arguments for what they already believe.

How does confirmation bias affect relationships?

It can lock in first impressions and assumptions, making misunderstandings and resentment harder to correct. Checking your assumptions out loud helps.

The bottom line

Confirmation bias is not a defect you can remove; it is a default you can manage. The next time you feel very sure, pause and ask what evidence you may have been filtering out. Choose one belief you hold strongly this week and spend ten minutes looking for the best argument against it.