The Student Council's Survey on Later Start Times
A student council wants to argue that school should start one hour later. They survey 150 students during lunch. They ask: 'Do you agree that a later start would help you feel less tired and improve your learning?' 118 students say yes. The council writes: '79% of students support a later start, so the school should change its schedule.' A teacher points out that the survey was conducted only at lunch, only 150 of 640 students were asked, and the question combined two ideas (feeling less tired AND improving learning). She also notes that students who were absent or who eat lunch elsewhere were not included. The council wants to revise their report before presenting it to the principal.
1. Identify the weaknesses in the council's survey and explain how each weakens the claim.
Read the worked answer
Several problems undermine the 79% figure:
- Leading / double-barrelled question. 'Help you feel less tired and improve your learning' asks about two things at once. A student who agrees with one but not the other cannot answer cleanly. Agreement may reflect only one part.
- Convenience sampling. Only students present at lunch were surveyed. Students who skip lunch, are absent, or eat elsewhere are excluded - they may differ systematically (e.g., more tired, more disengaged).
- Small, non-random fraction. 150 of 640 is about 23%. Without random selection, the sample may not represent the whole school.
- Social pressure / context. Surveying in a group at lunch can encourage agreement with peers.
- Wording bias. 'Do you agree that…' invites yes; a neutral question ('What time should school start?') would reduce bias.
Each of these means the 79% is not a valid estimate of whole-school support.
2. Rewrite the council's conclusion so it is supported by the evidence, and suggest one improvement to the survey design.
Read the worked answer
Revised conclusion:
'In a lunchtime survey of 150 students, 118 (79%) agreed that a later start might reduce tiredness and improve learning. Because the survey used a single combined question and did not sample students randomly, these results suggest interest in the idea but cannot be generalised to all 640 students. A broader, neutral survey is needed before recommending a schedule change.'
Design improvement (choose one, justify):
- Random sampling: select students randomly from the full enrolment list, including absentees via follow-up.
- Split the question: ask separately about (a) tiredness and (b) learning, using a neutral scale (strongly agree → strongly disagree).
- Pilot and reword to remove leading language.
Any of these directly addresses the identified weakness rather than just adding more respondents to a biased sample.