16Hit1600
Problem-Solving and Data AnalysisEasy–MediumNon-response as a source of bias

Problem-Solving and Data Analysis practice question

A council mailed a survey about a proposed car park to all 2,000 households in a district. Only 300 replied, and most of the replies came from households strongly opposed to the proposal. Which of the following best describes the main problem with using the replies to judge the district's opinion?

  1. A. Households that chose to reply may differ from those that did not, so the replies may not represent the whole district.Correct
  2. B. 300 replies are too few to say anything about 2,000 households.
  3. C. Surveys sent by post are always unreliable.
  4. D. The survey should have contained more questions.

Answer: A. Households that chose to reply may differ from those that did not, so the replies may not represent the whole district.

The 300 selected themselves by replying, and strong feeling drove many of them to do so. That self-selection, not the count, is what makes the replies unrepresentative.

Why the other answers are wrong

B. 300 replies are too few to say anything about 2,000 households.
300 would be plenty if they were a random sample; the problem is WHO replied.
C. Surveys sent by post are always unreliable.
Postal surveys can work when the response is representative; the issue here is the low, one-sided response.
D. The survey should have contained more questions.
More questions wouldn't change who answered them.

More Problem-Solving and Data Analysis questions

Not affiliated with the College Board. SAT is a trademark of the College Board.