AI generatedStatistics

Interpreting a sample’s limitations

Interpreting a sample’s limitations.

  • Check who was included, how many responded, when data were collected and who might be missing.
  • Sample results are estimates; another random sample can give different results.
  • Link the limitation to the claim being made.

Example 1

A random sample of 30 students gives 60% in favour of a proposal. Must exactly 60% of the whole school agree?

Different random samples can give different results.

No. The sample percentage is an estimate of the population percentage.

Example 2

A large survey includes only adults from one town. Can it safely describe all adults in the country?

A large sample does not automatically cover the target population well.

Not without further evidence. The town may differ from the rest of the country.

Your turn

Question 1

A random sample is selected, but only half reply. What is one concern?

Check answer
Those who reply may differ from those who do not, causing non-response bias.
Question 2

A travel survey from ten years ago is used to estimate current habits. Give a limitation.

Check answer
Habits or transport options may have changed, so the old results may be out of date.
Question 3

Only 5 people are surveyed to estimate a town’s favourite sport. Why is the estimate uncertain?

Check answer
A few unusual responses can strongly change the sample proportions.
Question 4

A company surveys current customers about why people stopped buying from it. Explain the problem.

Check answer
Former customers are missing, although their reasons are the target of the investigation.
Question 5

A random sample finds that students who exercise more tend to score higher. Can the survey prove exercise causes higher scores?

Check answer
No. It shows an association in the sample, not proof of causation.