Lesson 5 of 7 · 8 min
Numbers and “a study proves”
A number looks like a fact. Four questions show whether there is anything behind it.
Try it yourself: what does the number really say?
Tap the parts where a number or a “study” promises more than it shows. Careful: one of the examples handles its numbers honestly.
If you see nothing suspicious in a message, choose “I see nothing suspicious”.
Round 1 of 3 · Found: 0
An article on a health website
A study proves: people who eat porridge for breakfast have a 50% lower risk of forgetfulness. Nine out of 10 of our customers confirmed it.
Sample no. 245
Four questions for every number
1. Out of how many? Percentages without counts tell you nothing. “A rise of 100%” can mean from two to four. “Nine out of 10” can mean ten people were asked.
2. Compared with what? “Half the risk” sounds great. But if the risk fell from 2 people in a thousand to 1 in a thousand, little has changed for you. Ask for both numbers, before and after.
3. Who measured it, and on whom? A survey of a company’s own customers, a poll on a website or 24 volunteers don’t speak for everyone. And whoever paid for the study has an interest in the result.
4. Does that conclusion really follow? When two things happen at the same time, one need not cause the other. In summer, ice cream sales go up and so does the number of drownings. Ice cream is not to blame; the warm weather is.
“A study proves”
One study is one piece of the jigsaw, not the whole picture. Science is only sure when different teams repeatedly arrive at the same result. An honest article therefore writes “a study suggests” and says who carried it out, on how many people and where it was published.
A chart can distort without lying
A cut-off axis turns a small difference into a chasm. Before you believe a chart, look at where the axis starts and what values are on it.
A real example: it depends on the question
How many people come across disinformation? According to a Eurostat survey from 2025, across the EU 56% of people aged 16–74 saw, within three months, online content they considered untrue or doubtful. In Czechia it was 64%. Yet in the Czech survey CEDMO Trends from autumn 2025, only 25% of respondents said they had been exposed to disinformation.
Both Czech numbers are honest. The two surveys simply asked differently: one about “untrue or doubtful content”, the other about “disinformation”. So with every survey it pays to know how the question was worded.
And one more Eurostat figure is worth remembering: across the EU, fewer than a third of people (29%) checked whether such content was true.
Check your understanding
Quiz · 0/31. An advert says: “Risk of catching a cold cut by 50%!” What do you need to know to make sense of it?
2. On a chart, one bar is twice as tall as the other. Where do you look first?
3. “People who drink coffee live longer. So coffee prolongs life.” What is wrong with this?
Found a mistake or inaccuracy in this lesson?Report a mistake
What next
Done? Next lesson: Tricks in arguments: us versus them (8 min)
The whole course is free, no sign-up needed.