Confirmation bias

How it works

Your brain is not a judge weighing evidence. It is a lawyer with a client, and the client is whatever you already believe. New information gets sorted before you are even aware of it: fits goes straight in, doesn’t fit gets a cross-examination. Psychologists call this confirmation bias, and it is probably the most studied flaw in human thinking. Raymond Nickerson’s 1998 review called it “ubiquitous” for a reason: it shows up in juries, doctors, scientists and people arguing about football.

Manipulators do not need to create the bias. They only need to feed it. A propaganda page does not try to convince its opponents; it gives its own audience a steady diet of stories that confirm what they already suspect. Each story feels like proof, so nobody checks. That is why so much disinformation is not even clever. It is simply agreeable. A fake story about the other side behaving exactly as you expected is the easiest thing in the world to share.

The same mechanism explains why fact-checks so often bounce off. A correction that contradicts your side is processed as an attack, so you look for reasons to dismiss it: the source is biased, the fact-checker is funded by someone, the detail does not matter anyway. (Good news from the research: the famous “backfire effect”, where corrections make people believe the myth more, turns out to be rare. Corrections mostly work, just more slowly than we would like.)

Example

@free_voice_now, 4h Here is the minister’s “modest” lake house. 12 bedrooms. Paid from our taxes. Share before they delete it.

@tom.k.1987, 3h Knew it. They are all the same. Shared.

@quiet_checker, 1h Reverse image search: that is a hotel in Austria, photo from 2014. Not saying the minister is a saint, but this one is fake.

Notice that the first two users wanted it to be true, so nobody looked. Now flip the sides: a pro-government page posts “the opposition leader’s secret Brussels salary”, and the other half of the feed does exactly the same thing. The bias has no party.

Why it works on you

• Agreement feels like truth. When a claim matches what you expect, it arrives with a little glow of yes, obviously. That glow is familiarity, not evidence. It is the same feeling that powers the illusory truth effect.
• Disagreement feels like danger. Information that threatens your side is processed as a threat to you. You get defensive before you get curious. See us vs. them.
• Your feed does the filtering for you. You follow people who think like you, the algorithm learns what you click, and after a while the agreement looks like consensus. That is an echo chamber, and it hides the bias from you.

What to do

When a story pleases you: that is your cue, not your reward. Run the swap test: would I believe this if the other side had posted it? If the honest answer is “no, I’d want a source”, go and get the source. The stories you most want to be true are the ones most worth checking.

When you catch yourself dismissing a correction: notice the reasons you are generating. If they appeared in under two seconds, you did not think them. You felt them. Give the correction the same thirty seconds you gave the original claim. Pausing before sharing is the cheapest habit here.

In a comment section: do not tell people they are biased. Everyone is, and nobody has ever been argued out of a belief by being called stupid. Share what you checked and how. “I wanted this to be true too, so I looked it up” lands far better than “fake news”.

Don’t confuse it with

• Trusting sources with a track record is not bias; it is how you survive the internet. Bias is trusting the source because it agrees with you, regardless of its track record.
• Having a strong opinion is fine. The bias is in how you handle evidence against it, not in having the opinion.