Why is it so hard to change our own minds, even when we know we're wrong?
Sometimes the new fact is clear, and yet an old belief keeps its place. We may have repeated it for years, built plans around it, or found our people through it. Changing a sentence in the mind can feel like changing the story of who said it.
Why some beliefs resist
In a brain-imaging study of 40 politically liberal participants, challenges to strongly held political views engaged networks associated with self-representation more than challenges to nonpolitical views. Resistance also related to activity in regions involved in emotion. That is a finding about this task and sample, not proof that every disagreement is processed as physical danger.
Prior commitments can make us scrutinize unwelcome information harder than welcome information. Yet resistance is not destiny. In two experiments on misinformation corrections, no tested item backfired more under correction than under a comparison condition. Other experiments found that asking people to consider alternatives could help them revise beliefs when presented with contrary evidence.
The move that remains ours
It may help to separate a belief from the person who holds it: “I believed this because...” can be true even when “I still believe this” no longer is. That is an invitation, not a proven universal technique. Social stakes, the quality of evidence, and trust in its source still matter.
A changed mind is not a mind that failed. It is one that kept the question open long enough to learn.
Follow the evidence
Observed: Identity-linked beliefs can show resistance, while factual corrections often improve accuracy in controlled tests.
Our reading: Belonging, self-concept, and trust can shape how evidence is received.
Still open: Which approaches help with deeply held beliefs outside short experiments and across different communities.
Sources
- Kaplan et al., neural correlates of belief maintenance (2016)
- Swire-Thompson et al., corrections and backfire tests (2022)
- Rose et al., counterfactual thinking and belief revision (2024)
Research checked 15 November 2026. See something we should correct? Tell us.
Correction, 15 November 2026: An earlier version equated belief challenges with physical threat and implied corrections entrench beliefs. The cited studies do not support those claims as general rules.
Have a question you cannot stop asking?
Send it. We read every question, and answer the ones we take on in full.
Ask your own Why