What is confirmation bias?
The moment we believe an idea, our minds start taking sides: they magnify the supporting evidence and overlook or forget whatever runs against it. This quiet favouritism is called confirmation bias, and none of us is exempt.
Short answer
Confirmation bias is the tendency to seek out information that supports our existing beliefs, to trust it more and to remember it better, while playing down or forgetting information that contradicts it. It is not a conscious choice; the mind does it without our noticing.
One bias entering through three doors
Confirmation bias is not a single error but a family of tendencies operating at three stages. Searching: when we look for an answer, we turn to sources likely to support our belief; even the search we type is partial. Interpreting: the same ambiguous evidence is read by each side in its own favour; two people with opposing views can read the same report and both conclude they were right. Remembering: events that confirm the belief stay vivid in memory, while those that refute it fade. When the three combine, a person sincerely believes that the evidence has always supported them. What is more, the three doors feed one another: partial searching supplies material for partial interpretation, and partial recall sets the direction of the next search.
A classic observation: the rule-finding game
In a well-known experiment by the psychologist Peter Wason, participants are given the sequence "2-4-6" and asked to find the hidden rule it follows; they may propose their own sequences as tests. Most participants assume a rule such as "even numbers increasing by two" and go on proposing sequences that would confirm that assumption: 8-10-12, 20-22-24, and so on. The hidden rule is in fact much broader: any increasing sequence. The way to find the rule is to make attempts that would break the assumption, by proposing 1-2-3 or 5-4-3, for instance. The lesson of the experiment is clear: confirming is easy and satisfying, while what teaches you something is the attempt to refute.
Examples from the sky to everyday life
Horoscopes are the most familiar stage for this bias: the accurate line in the weekly reading stays in mind while the five that missed are quietly forgotten. The same mechanism works in other areas. Someone who believes a supplement is working credits the good days to the supplement and the bad days to other factors. Someone who trusts a brand reads the positive reviews and treats the negative ones as malicious. News feeds, too, learn from our preferences and keep showing us content we will endorse; this time our bias is amplified by software.
Why is it so persistent?
Confirmation bias does not come from laziness; it is a combination of mental economy and the defence of identity. Questioning every belief constantly is exhausting, and the mind works so as to preserve its established order. Moreover, some beliefs are part of our identity: an objection to them feels like an objection to who we are. This is why being intelligent and educated is no protection on its own; skill in argument can even make it easier to justify a mistaken position more convincingly.
A lesson from history: the treatment that "worked" for centuries
To see the cost of confirmation bias, it is enough to look at the history of medicine. Bloodletting, the draining of blood from a patient through practices akin to cupping, remained the standard treatment in Western medicine for close to two thousand years. Physicians sincerely saw that it worked: a patient who recovered counted as the method's success, while a patient who died was explained away as "having been too far gone anyway". Both outcomes fed the belief, and no outcome could shake it. Questioning the method became possible only when counting began: when the outcomes of patients who were bled and those who were not were compared systematically, it emerged that the practice brought no benefit in many conditions and even did harm. The lesson of the story is that error can persist without any bad faith: intelligent, experienced and well-meaning people are defeated by their own observations as long as they recall instead of counting.
The antidote: acting as your own opposing counsel
We cannot abolish the bias, but we can reduce its effect. The strongest tool is to make a habit of one question: "if this belief of mine were wrong, how would I know?" It helps to look deliberately for contrary evidence when assessing a claim, to keep records and count the hits and the misses together, and to read regularly a good-quality source that disputes the view we favour. Science has institutionalised exactly this: blind and double-blind trials are designed to stop researchers' own expectations from shaping the result. For the sister concept, see our guide to falsifiability.
Checklist
- Write down an answer to "how would I know if I were wrong?" for one important belief
- When researching a claim, deliberately read a good-quality source from the opposing view
- Note your predictions in advance; count the hits and the misses together
- Add a source that disagrees with you to feeds that always take your side
- In a discussion, try summarising the other side's strongest argument in your own words
Frequently asked questions
Does everyone have confirmation bias?
Yes. It operates in everyone regardless of education or intelligence. The difference lies in whether a person acknowledges the tendency and takes countermeasures against it.
If I notice my bias, am I free of it?
Awareness is the first step but not enough on its own; the bias goes on working unconsciously. What is effective is turning the search for contrary evidence and the keeping of records into habits.
What is the difference between confirmation bias and apophenia?
Apophenia is seeing patterns in random data; confirmation bias is favouring evidence that supports an existing belief. The two often work together: first the pattern is seen, then only the examples that confirm it are collected.
Do scientists fall into this bias too?
Yes, scientists are human. The strength of science comes not from the impartiality of individuals but from mechanisms that offset bias systemically, such as peer review, replication studies and blinded trials.
This guide draws on standard textbooks in cognitive psychology and on peer-reviewed research findings in decision-making and reasoning.