Human sciences: whose responses support the claim?
| English | Français |
|---|---|
| sample/ˈsæmpl/ | sample |
| population/ˌpɒpjʊˈleɪʃn/ | population |
| selection bias/sɪˈlekʃn ˈbaɪəs/ | selection bias |
An online vote seems decisive
- A school has 400 students. An optional online poll receives 100 responses, and 80 support later library hours. Does this show that 80% of all students support the change?
- A sample 样本 is the group from which evidence is collected. Here it contains the volunteers who answered, not all 400 students.
Define the group and the claim
- The · Le population 总体 is the group the question concerns. Decide whether the claim is about respondents, all enrolled students or only regular library users.
- The count establishes 80 out of 100 respondents supported the proposal. Extending that finding to everyone requires reasons why the respondents represent the wider group.
Which percentage is directly established by the recorded poll?
The denominator is the group that actually answered.
Check who entered the sample
- People who strongly want a change may be more likely to answer. Students without access to the poll may be absent. This is possible selection bias 选择偏差: entry into the sample changes the evidence.
- Ask how responses were invited, who could respond and whether relevant groups were missed. More responses through the same narrow channel do not automatically remove the problem.
Adding more volunteers from the same interest channel guarantees a representative sample.
The same selection process can continue to miss relevant groups.
Which missing detail most affects the wider claim?
Participation determines which people the evidence describes.
Separate description from explanation
- A poll describes stated preferences under its wording and conditions. It does not establish that later hours improve learning or explain why each student selected an answer.
- For any follow-up involving people, use the school’s approved process and collect only the information required. Clear wording and privacy safeguards are part of a defensible method.
What does this preference poll fail to establish?
A preference response does not measure a learning outcome or isolate a cause.
Revise the conclusion before making a decision
- Report what is known: 80% of these respondents supported the wording they were shown. State the participation gap and consider an approved method reaching different student groups.
- Ask how a method shapes knowledge about a community. The issue is whether this evidence supports this claim, not whether every survey is reliable or unreliable.
Match each term to this case.
Keep the target group distinct from the evidence actually collected.
Which statements responsibly report the result?
Describe the result and the relevant participation limit.
All 100 respondents used an optional library-interest channel. Eighty supported later hours. The result accurately describes those recorded responses. Since students who rarely use the library may not read that channel, it does not establish the proportion among all 400 students. An approved follow-up could invite a defined range of students through several access routes, record participation and ask a neutrally worded question. Even then, stated support would not by itself measure improved learning.
A precise percentage can still support a narrow claim. Check the group and question before treating precision as representativeness.
Define the population, explain how the sample was selected and keep the conclusion within what the responses establish.