Changing your mind
Hold beliefs with a confidence level, move them by the right amount when evidence arrives, and correct yourself in public without losing face.
A DIAL, NOT A SWITCH
Beliefs come in degrees
We often talk as if beliefs were on/off switches: true or false, believe it or don't. Real knowledge is more like a dial. Try giving your beliefs a rough number: "I'm about 90% sure the pharmacy closes at 9", "maybe 60% that the new road will open this year". A dial makes it easy to change your mind a little, instead of all or nothing.
"I'm 100% sure" about something you've never checked is a warning sign, not confidence. Save the top of the dial for things you have strong reasons for.
Check yourself
Saying "I'm about 70% sure" instead of "I know" makes it easier to adjust your view when new evidence arrives.
Show the answer
True
True. With a number, new evidence can move you from 70% to 60% without a dramatic "I was wrong about everything". Small, honest adjustments are much easier than full reversals, for you and for the people you're talking to.
HOW FAR TO MOVE
Update by the right amount
Strong evidence should move you a lot; weak evidence should move you a little. Several good studies agreeing, or an official record, can turn the dial far. One anecdote, one viral post or one angry comment should nudge it at most. There are two ways to get this wrong: never moving at all (stubbornness) and jumping to whatever you read last (gullibility).
You're fairly sure a medicine is safe. A friend says it gave her a headache: a small nudge. A large review finds a real risk: a big move.
Check yourself
How far should each piece of evidence move you?
- A neighbour's story about what happened to her cousin
- Several large studies that all find the same result
- A viral post with lots of likes and no source
- The official announcement on the organisation's own website
- One angry comment under a news article
- A correction published by the outlet that ran the original story
Show the answer
A small nudge: A neighbour's story about what happened to her cousin, A viral post with lots of likes and no source, One angry comment under a news article
A big move: Several large studies that all find the same result, The official announcement on the organisation's own website, A correction published by the outlet that ran the original story
Check yourself
Arash is about 80% sure a common food additive is safe, based on several reviews. Then he reads one person's story of feeling ill after eating it. What's the best response?
- Ignore it completely; stories never count
- Lower his confidence a little, and maybe look for better evidence
- Flip to believing the additive is dangerous
- Stay at exactly 80%, because changing looks weak
Show the answer
Lower his confidence a little, and maybe look for better evidence
Right. One story is weak evidence: worth a small nudge, and maybe a reason to look further, but not enough to outweigh several reviews.
Check yourself
According to Philip Tetlock's forecasting research, what did the most accurate forecasters tend to do?
- Make one careful prediction and stick to it
- Update their predictions often, in small steps, as news came in
- Follow the opinion of the most famous expert
- Change their prediction completely after every headline
Show the answer
Update their predictions often, in small steps, as news came in
Yes. Frequent, small updates beat both stubbornness and lurching. Each new piece of evidence moved them by about as much as it deserved.
Two mindsets (Julia Galef, The Scout Mindset, 2021)
Soldier
Defends a position. Evidence is ammunition: welcome if it helps, attacked if it doesn't. Changing your mind feels like losing a battle.
Scout
Wants an accurate map, even when the news is unwelcome. Evidence is information about the terrain. Changing your mind means the map got better.
Check yourself
What does the "scout mindset" mean?
- Always taking the opposite side to test people
- Wanting an accurate picture more than winning the argument
- Never holding a firm opinion about anything
- Collecting as much evidence as possible for your own side
Show the answer
Wanting an accurate picture more than winning the argument
Right. A scout's goal is a map that matches the territory. If the evidence says the bridge is out, the scout wants to know, even if it spoils the plan.
A SKILL, NOT A DEFEAT
"I was wrong about that"
Saying it out loud feels costly, but people generally trust someone more when they see them correct themselves; it shows their other claims were checked too. It helps to keep your identity separate from any single belief: you are someone who cares about getting things right, not someone who is never wrong.
"I said the new bus route starts Monday. I checked, it's actually the 1st of next month. My mistake." Nobody in the group thinks less of the person who wrote that.
When you've shared something false
- Correct it in the same place
Same group, same thread, same audience. A correction somewhere else doesn't reach the people who saw the original.
- Lead with what's true
Use the truth sandwich from lesson 4: the fact first, then a brief flag on the false claim ("the post I shared said X; that's not right"), then the fact again.
- Delete or edit the original if you can
So it can't be forwarded further. If you can't, reply to it with the correction so the two travel together.
- Do it quickly
Corrections usually reach fewer people than the original. The sooner you post one, the fewer people see the false version first.
Check yourself
Yesterday you shared a health tip in the family group. Today you find out it's false. What's the best fix?
- Say nothing; people will have forgotten it by now
- Post a clear correction in the same group, say what's true, and delete the original if you can
- Post the correction on your personal page instead
- Share a different, true health tip so the false one gets buried
Show the answer
Post a clear correction in the same group, say what's true, and delete the original if you can
Yes. The correction has to reach the same people, say clearly what's true, and stop the original from travelling further. Doing it quickly is part of doing it well.
Helping someone else change their mind
- Start from common ground: "We both want the kids to stay healthy."
- Ask what would convince them, and listen to the answer.
- Offer a face-saving path: "I believed it too at first, until I saw where it came from."
- Don't expect a full turn in one chat. A small move from 90% to 70% is a real win.
Lesson recap
- Hold beliefs on a dial, not a switch: "about 70% sure" is easier to adjust than "I know".
- Strong evidence should move you a lot, weak evidence a little. Avoid both never moving and jumping to the latest post.
- In Tetlock's research, the most accurate forecasters updated often, in small steps.
- A scout wants an accurate map; a soldier wants to win. Changing your mind means the map got better.
- If you shared something false: correct it in the same place, lead with what's true, remove the original, and be quick.
- To help others change their minds: common ground, a face-saving path, and patience.