Confirmation bias is strongest when you’re actively searching for new information, rather than weighing evidence you’ve already collected. In experimental settings, the bias effect during information search is roughly four times larger than when people evaluate evidence placed in front of them. But the search phase is just one amplifier. Several other conditions, from emotional arousal to identity threat, can push confirmation bias from a mild tilt into a powerful distortion.
The Search Phase Is Where Bias Peaks
Confirmation bias isn’t a single behavior. It shows up in three distinct ways: how you search for information, how you weigh evidence once you have it, and how you remember facts later. These three components aren’t equally strong. When researchers measured the bias across all three, the effect size during information search was 1.18, compared to 0.59 for memory recall and just 0.31 for weighing evidence. In practical terms, this means you’re far more likely to seek out information that supports what you already believe than you are to misremember or misinterpret facts that contradict you.
This matters because the internet has turned most of us into constant information searchers. Every time you type a question into a search engine, choose which article to click, or decide which social media posts to read, you’re in the phase where confirmation bias exerts its greatest pull. The bias isn’t just about how you process what you find. It’s primarily about what you choose to look at in the first place.
When Emotions Run High
Strong negative emotions, particularly anger, amplify confirmation bias significantly. Research on investigative interviewers found that angry participants drew conclusions without thoroughly analyzing the evidence, while sad participants examined case materials more carefully and adjusted their conclusions based on inconsistencies. Anger essentially narrows your cognitive focus and pushes you toward the conclusion you already hold.
Disgust has a similar effect. People experiencing disgust tend to seek information that confirms their initial suspicion rather than testing it. This pattern intensifies among people who have difficulty regulating their emotions. If you’re someone who feels things strongly and struggles to step back from those feelings, your confirmation bias tends to run hotter across the board. The emotional charge doesn’t even have to come from the topic itself. General emotional arousal about a subject area can spill over into biased questioning and interpretation.
When Beliefs Are Tied to Identity
Confirmation bias shifts into a qualitatively different gear when the belief in question is connected to who you are. Beliefs tied to long-held identities resist change and bias the processing of new information in ways that go beyond simple preference for agreeable facts. When a belief is wrapped up in a political identity like “liberal” or “conservative,” or a group membership like a profession, religion, or social community, people abandon it only with great reluctance.
This isn’t stubbornness in the ordinary sense. Accepting contradictory evidence would carry a real psychological cost: it would threaten your sense of self and potentially your standing in a group you value. So you resist persuasion, reject compromises, and filter incoming information through a lens that protects the identity rather than updating the belief. Some researchers argue this identity-protective pattern is so distinct from ordinary confirmation bias that it deserves its own label, “myside bias,” the tendency to find arguments specifically supporting your own views rather than simply confirming any hypothesis you happen to be testing.
When You’ve Already Committed to a Conclusion
One of the most striking findings about confirmation bias comes from brain imaging research. Information that contradicts a previous choice is encoded precisely in the brain, meaning your neurons register it just as accurately as confirming information. But it has little impact on subsequent behavior. The brain essentially reads out confirming evidence more readily than disconfirming evidence from the same neural regions. You’re not failing to perceive contradictory facts. You’re filtering them out downstream, after they’ve already been recorded.
This helps explain why first impressions and preliminary conclusions are so sticky. Once you’ve committed to a position, even tentatively, your brain begins treating supporting and opposing evidence asymmetrically. In a study of psychiatric diagnosis, clinicians who searched for information in a confirmatory way after forming a preliminary diagnosis made the wrong diagnosis 70% of the time. Those who deliberately searched for disconfirming evidence were wrong only 27% of the time. The preliminary conclusion itself became the trap.
When the Task Is Abstract or Rule-Based
Not all types of thinking trigger confirmation bias equally. The bias is stronger when you’re testing logical or numerical rules (effect sizes of 0.72 and 0.91) than when you’re forming impressions of other people’s personality traits (effect size of 0.36). Abstract reasoning tasks, where you’re evaluating whether a pattern or rule is correct, seem to invite more confirmatory thinking than social judgments, where you may naturally consider multiple explanations for someone’s behavior.
This has implications for fields that rely heavily on rule-based analysis: programming, data science, financial forecasting, legal reasoning. If your work involves testing whether a hypothesis or model is correct, you’re operating in exactly the cognitive territory where confirmation bias runs strongest.
A Consistent Trait Across People
Confirmation bias isn’t random. People who show strong bias in one task tend to show it in others, suggesting it operates as a stable individual trait. Across experiments, participants chose confirming information 69% of the time and disconfirming information only 48% of the time, even when the hypotheses being tested were assigned by researchers and had nothing to do with participants’ personal beliefs. This general tendency to confirm shows up regardless of whether the topic matters to you personally.
One reliable predictor of stronger confirmation bias is belief in pseudoscience. People who endorse pseudoscientific claims score higher on all three components of confirmation bias: biased searching, biased interpretation, and biased memory. This correlation doesn’t tell us which comes first, but it suggests that the same cognitive style that makes someone prone to confirmation bias also makes them more receptive to claims that lack scientific support.
Reducing the Bias When It’s Strongest
The conditions that maximize confirmation bias, active information search, strong emotions, identity-linked beliefs, and early commitment to a conclusion, are common in everyday life. But the bias isn’t fixed. A single 40-minute training session, in which participants learned about decision-making biases and worked through a case study illustrating confirmation bias in action, produced significant reductions in biased reasoning. This held true for both trained professionals (national risk analysts) and university students, and the improvement showed up across different subject areas, not just the domain covered in training.
The most effective practical strategy maps directly onto the diagnostic research: deliberately search for evidence that would prove you wrong. Clinicians who used a disconfirmatory search strategy were roughly seven times more likely to reach the correct diagnosis than those who searched in a confirmatory way. You don’t need to distrust your own thinking entirely. You just need to build the habit of asking one extra question: what would I expect to see if I were wrong?