In a selection bias question, compare the group a survey wants to describe with the group its method could reach. Random selection helps within the reachable group, but it cannot include people who had no chance of being contacted. A large sample does not fix a gap between those two groups.
Hints
- Hint 1
The target population is the whole group the firm wants to describe. Identify that group first, then ask whether every part of it had a chance to enter the survey.
- Hint 2
A sampling frame is the group a method can reach. Random selection operates only within that frame. Compare the kinds of telephone numbers the firm could dial with the households in the target population.
Step-by-step
Check who could be reached
Step 1Identify the group the estimate is for
No Desmos needed. The decision is about who could be surveyed, not arithmetic.
The firm wants an estimate for teenagers in the city. That whole group is the target population, meaning the group the estimate is meant to describe.
- Step 2
Find who the calls could miss
The firm randomly dialed landline telephone numbers, so teenagers in households without landlines had no chance of being reached. Random dialing helps choose among landline numbers; it cannot bring those missing households into the survey. Ask who could have been picked, not only whether the picking was random.
- Step 3
Judge the citywide estimate
The -hour mean describes the teenagers surveyed. Even with 600 households, leaving out households without landlines could tilt an estimate for all city teenagers. That is possible selection bias: the method may systematically miss part of the target population. So the landline-only sampling method may be biased. Choice C.