| English | 释义 | 拼音 |
|---|---|---|
| population/ˌpɒpjʊˈleɪʃn/ | 总体 | zǒng tǐ |
| continuous data | 连续数据 | lián xù shù jù |
"Dewey Defeats Truman", 1948
- On the night of the 1948 US election, the Chicago Tribune printed the headline "Dewey Defeats Truman". Polls had used quota samples, letting interviewers pick whom to ask, and stopped polling weeks early. Truman won. Pollsters switched to random methods afterwards.
- Higher questions ask you to judge samples and to justify comparisons with evidence.
Why did the 1948 polls fail?
难度 1/5 · 今日首次完成可获 100 金币及少量额外奖励
Quota sampling and old data.
From sample to population
- A random sample of 80 with 24 walkers estimates 30% of the population 总体: about 180 of 600.
- The estimate has uncertainty: a new random sample will differ a little.
In a random sample of 200 trees, 34 have a disease. Estimate how many of 5000 trees in the forest have it.
难度 2/5 · 今日首次完成可获 140 金币及少量额外奖励
0.17 × 5000.
Sources of bias
- Self-selection (voluntary polls), convenience (whoever is nearby), and non-response (people who refuse) all bias a sample.
- Quota sampling let interviewers choose, and in 1948 they chose people who were easier to reach.
Match each sample to its main problem.
难度 2/5 · 今日首次完成可获 140 金币及少量额外奖励
Types of bias.
Worked example: comparing two samples
Sample A: 24 of 80 students walk. Sample B, from another school: 39 of 150 walk. Which school has the higher proportion of walkers?
- Proportions: A $= \dfrac{24}{80} = 0.30$; B $= \dfrac{39}{150} = 0.26$.
- Compare: A's proportion is higher.
- Caution: the difference of 0.04 is small; with samples this size, another pair of samples could reverse it. Both samples must be random.
Samples, populations and justified comparisons
Define the target population and variables before sampling
Classify the worked-case statements, then explain the units or invariant that justifies each decision.
Sample A: 45 of 120 like the new menu. Sample B: 33 of 75. Which has the higher proportion?
难度 1/5 · 今日首次完成可获 100 金币及少量额外奖励
Compare proportions.
Discrete and continuous data
- Discrete data is counted (number of siblings); continuous data is measured (journey time).
- Rounding a journey time to the nearest minute does not make it discrete.
Which are continuous data?
难度 2/5 · 今日首次完成可获 140 金币及少量额外奖励
Measured, not counted.
A good comparison
- Use the same measure for both groups, state the numbers, and write a conclusion in context.
- Mention a limitation: sample size, method or bias.
A difference in two samples is not proof of a difference in populations. Random samples vary. A small difference from small samples may just be chance — say how confident the evidence lets you be.
If sample A's proportion is 0.30 and sample B's is 0.26, population A must have a higher proportion.
难度 1/5 · 今日首次完成可获 100 金币及少量额外奖励
Samples vary.
Estimate population values from random-sample proportions and treat them as estimates. Identify self-selection, convenience, quota and non-response bias. Compare groups with matching statistics, in context, and state a limitation.
A researcher says: "My random sample of 40 students had a mean sleep time of 7.2 hours, so every student sleeps about 7.2 hours." Explain two problems with this claim.
Justified claims.