A number arrives in your feed. It is precise, it is confident, and it supports something you already half-believed. That is exactly the moment to slow down. Statistics are not lies, but they are almost always an argument in disguise: someone chose what to count, how to count it and which part of the result to show you.
In the Curious Minds group we keep a short checklist for this. None of it needs a maths degree. It needs about ninety seconds of honest curiosity.
1. Compared with what?
"Crime rose 20%" means nothing on its own. From what base? Over what period? A rise from five incidents to six is 20%. So is a rise from five thousand to six thousand. Ask for the absolute numbers next to the percentage and half of all alarming headlines quietly deflate.
2. Who counted, and how?
Every figure has a method behind it. Was it a survey, a model, a census, a guess dressed as an estimate? A survey of people who chose to answer tells you about people who like answering surveys. A model tells you about the assumptions of the people who built it. Neither is useless. Both need labelling.
3. What is the denominator?
"Most accidents happen within five miles of home" sounds like a warning about your neighbourhood. It is mostly a statement that you drive within five miles of home more than anywhere else. When a number surprises you, find the thing it is being divided by. Surprises usually live there.
4. Would the opposite result have been reported?
A study that finds nothing rarely makes the news. If you only ever see the striking results, your picture of the world is built from the loudest tenth of the evidence. Ask yourself whether a boring answer to the same question would have reached you. If not, discount accordingly.
5. What would change my mind?
This is the question that separates thinking from cheering. If no possible number could move you, you are not evaluating the statistic; you are using it. Decide, before you look closely, what result would make you update, then look.
The goal is not to distrust every number. It is to know which ones have earned your trust, and why.
A small habit that helps
- Write the claim in one plain sentence before you react to it.
- Find the original source, not the summary of the summary.
- Note what you would need to see to believe the opposite.
- Share the figure with its caveats attached, not stripped off.
Members who follow this routine tell us the same thing: they read less, argue less and remember more. That is a good trade. If you have a favourite example of a statistic that fooled you, post it in Ask the Community. The collection of near-misses is one of the most useful threads we have.
0 comments
No comments yet — be the first to add your view.