MyLeoNes™

Samples vary by chance — Data, 11–13 years

Two fair samples from the same group can give different results, even when nobody made a mistake.

A sample is a snapshot

A sample is only part of a larger group, so it cannot show every detail perfectly. Choose ten pupils twice from the same year and the groups may have different reading habits by chance. Both results can be honest observations, but neither should be treated as the exact answer for everyone.

Why results wobble

People in a group are not identical. A small sample may happen to include more keen readers, taller plants or rainy days than another sample. Understanding this natural variation stops us treating a small difference as proof that the whole group has changed.

Two draws of sweets

A bag has 50 sweets, 20 of them lemon flavour, so the true share is 40%. Draw 5 sweets without looking: you might get 3 lemon sweets, or only 1. Those samples give 60% and 20%, although the bag did not change. Drawing 20 sweets usually gives a steadier estimate because chance has less influence.

One sample must be right

It is easy to think that a carefully chosen sample must give the exact group result. That is reasonable because careful selection reduces unfairness. Yet chance still decides who is included, especially in a small sample. A result can be useful without being exact, so it should be reported as an estimate.

Where it helps

Pollsters, ecologists and quality inspectors often study a sample instead of every person, animal or product. They repeat or enlarge samples when a decision matters. When you see a survey result, asking how many people were asked helps you judge how much chance may have affected it.

Keep exploring

Other languages

Loading MyLeoNes™…