📚 The Conservative Party: A-Level Statistics in Election Polls | 保守党：选举民调中的 A-Level 统计
In Edexcel A-Level Mathematics, statistical modelling is often tested through real-world contexts such as opinion polls. This article uses the Conservative Party’s reported voting intention as a case study to revise key topics: sampling methods, the binomial distribution, normal approximation, confidence intervals, hypothesis testing, and margin of error. The aim is to show how a political dataset can be handled rigorously using the statistical tools required by the Edexcel specification.
在 Edexcel A-Level 数学中，统计建模常常通过民意调查等真实情境来考查。本文以保守党报告中的投票意向为案例，复习关键知识点：抽样方法、二项分布、正态近似、置信区间、假设检验以及误差幅度。目的是展示如何利用 Edexcel 考试大纲所要求的统计工具，严谨地处理一组政治数据。
1. Statistical Context: Why Polls Are Useful | 统计背景：民调为何有用
An opinion poll reports the proportion of respondents who say they would vote for the Conservative Party. This proportion, denoted p̂, is a statistic calculated from a sample. Since the whole electorate is too large to survey, pollsters use a random sample and then apply inference to estimate the true population proportion p.
民意调查报告的是受访者中表示会投票给保守党的比例。这个比例记作 p̂，是根据样本计算出的统计量。由于全体选民数量太大，无法逐一调查，民调机构使用随机样本，然后应用推断方法来估计真实的总体比例 p。
- p̂ is the sample proportion, an estimator of p.
- p̂ 是样本比例，是 p 的估计量。
- Different samples will give different values of p̂; this is sampling variability.
- 不同样本会得到不同的 p̂ 值；这就是抽样变异性。
2. Sampling Methods: Random and Stratified | 抽样方法：随机与分层
For a poll to be valid, the sample must be representative. A simple random sample of size n gives every member of the population an equal chance of being selected. In practice, polling companies often use stratified sampling by region, age, and past voting behaviour to reduce bias.
要使民调有效，样本必须具有代表性。容量为 n 的简单随机样本使总体中的每个成员都有相等的被选中机会。实际上，民调公司常常按地区、年龄和过去的投票行为进行分层抽样，以减少偏差。
In Edexcel questions you may be asked to identify the sampling method and explain one advantage. For a stratified sample, an advantage is that it guarantees proportional representation of key groups, improving the accuracy of the estimated Conservative support.
在 Edexcel 考题中，你可能需要识别抽样方法并说明一个优点。对于分层抽样，一个优点是它能保证关键群体的比例代表性，从而提高保守党支持率估计值的准确性。
3. Binomial Model for Voting Intention | 投票意向的二项模型
If we randomly select n voters and let X be the number who support the Conservative Party, then X can be modelled by a binomial distribution provided each selection is independent and the probability p of supporting the party is constant.
如果我们随机抽取 n 位选民，并令 X 为支持保守党的人数，那么只要每次抽取相互独立且支持该党的概率 p 保持不变，X 就可以用二项分布来建模。
X ~ B(n, p)
The conditions for a binomial model are: a fixed number of trials n, two possible outcomes (Conservative or not), a constant probability p, and independent trials.
二项模型的条件是：试验次数 n 固定、只有两种可能结果（支持保守党或不支持）、概率 p 恒定，以及各次试验相互独立。
4. Normal Approximation to the Binomial | 二项分布的正态近似
When n is large and p is not too close to 0 or 1, the binomial distribution can be approximated by a normal distribution. This is useful for calculating probabilities in large election polls without using lengthy binomial sums.
当 n 较大且 p 不太接近 0 或 1 时，二项分布可以用正态分布来近似。这在计算大型选举民调的概率时非常有用，无需进行冗长的二项求和。
X ~ N(np, np(1 − p))
The mean is μ = np and the variance is σ² = np(1 − p). For the sample proportion p̂ = X/n, the approximate distribution is:
均值为 μ = np，方差为 σ² = np(1 − p)。对于样本比例 p̂ = X/n，其近似分布为：
p̂ ~ N(p, p(1 − p)/n)
The approximation is acceptable when np > 5 and n(1 − p) > 5. A continuity correction is required when approximating a discrete binomial probability by a continuous normal probability.
当 np > 5 且 n(1 − p) > 5 时，近似是可接受的。用连续正态概率近似离散二项概率时，需要进行连续性校正。
5. Confidence Intervals for Polling Proportions | 民调比例的置信区间
Suppose a poll of n = 1000 voters finds that 420 support the Conservative Party. The sample proportion is p̂ = 0.42. A 95% confidence interval for the true proportion p is given by:
假设一项 n = 1000 位选民的民调发现，有 420 人支持保守党。样本比例为 p̂ = 0.42。真实比例 p 的 95% 置信区间由下式给出：
p̂ ± 1.96 × √(p̂(1 − p̂)/n)
Substituting the values gives 0.42 ± 1.96 × √(0.42 × 0.58 / 1000) = 0.42 ± 0.0306. Hence the interval is (0.3894, 0.4506), meaning we are 95% confident that the true Conservative support lies between 38.9% and 45.1%.
代入数值得到 0.42 ± 1.96 × √(0.42 × 0.58 / 1000) = 0.42 ± 0.0306。因此区间为 (0.3894, 0.4506)，这意味着我们有 95% 的把握认为保守党的真实支持率介于 38.9% 和 45.1% 之间。
6. Hypothesis Testing: Has Support Changed? | 假设检验：支持率是否改变？
Edexcel questions often ask whether a poll provides evidence that the true proportion has changed from a stated value. For example, test whether Conservative support is greater than 0.40 at the 5% significance level.
Edexcel 考题常常问某次民调是否提供了真实比例相对给定值发生变化的证据。例如，在 5% 显著性水平下检验保守党支持率是否大于 0.40。
H₀: p = 0.40
H₁: p > 0.40
Using the poll data n = 1000 and p̂ = 0.42, the test statistic is:
使用民调数据 n = 1000 和 p̂ = 0.42，检验统计量为：
z = (p̂ − p₀) / √(p₀(1 − p₀)/n) = (0.42 − 0.40) / √(0.40 × 0.60 / 1000) = 1.291
The critical value for a one-tailed 5% test is 1.645. Since 1.291 < 1.645, we do not reject H₀. There is insufficient evidence at the 5% level to claim that Conservative support is greater than 40%.
单侧 5% 检验的临界值为 1.645。由于 1.291 < 1.645，我们不拒绝 H₀。在 5% 显著性水平下，没有足够证据表明保守党支持率大于 40%。
7. Margin of Error and Sample Size | 误差幅度与样本量
The margin of error is the half-width of a confidence interval. For a 95% interval it is 1.96 × √(p̂(1 − p̂)/n). Poll reports often quote a margin of error of ±3%, which reflects the sample size used.
误差幅度是置信区间宽度的一半。对于 95% 的区间，它是 1.96 × √(p̂(1 − p̂)/n)。民调报告通常引用 ±3% 的误差幅度，这反映了所用样本量的大小。
To achieve a required margin of error E at 95% confidence, the minimum sample size is:
要在 95% 置信度下达到所需的误差幅度 E，最小样本量为：
n = (1.96² × p(1 − p)) / E²
If no prior estimate of p is known, use p = 0.5 because it maximises p(1 − p) and gives the most conservative sample size. For E = 0.03, n = (1.96² × 0.25) / 0.03² ≈ 1067.1, so at least 1068 voters are needed.
如果没有 p 的先验估计，使用 p = 0.5，因为它使 p(1 − p) 最大化，并给出最保守的样本量。对于 E = 0.03，n = (1.96² × 0.25) / 0.03² ≈ 1067.1，因此至少需要 1068 位选民。
8. Comparing Two Polls: A Two-Sample Test | 比较两次民调：双样本检验
Sometimes you need to test whether the difference between two polls is statistically significant. Let Poll 1 have n₁ = 1000 and p̂₁ = 0.42, and Poll 2 have n₂ = 1200 and p̂₂ = 0.38.
有时你需要检验两次民调之间的差异是否具有统计显著性。设民调 1 的样本量 n₁ = 1000，样本比例 p̂₁ = 0.42；民调 2 的样本量 n₂ = 1200，样本比例 p̂₂ = 0.38。
The pooled proportion is:
合并比例为：
p̂ = (n₁p̂₁ + n₂p̂₂) / (n₁ + n₂) = (1000×0.42 + 1200×0.38) / 2200 = 0.39818
The test statistic for the difference between two proportions is:
两个比例之差的检验统计量为：
z = (p̂₁ − p̂₂) / √(p̂(1 − p̂)(1/n₁ + 1/n₂))
Substituting gives z = 0.04 / √(0.39818 × 0.60182 × (1/1000 + 1/1200)) = 0.04 / 0.02094 = 1.911. For a two-tailed 5% test, the critical value is 1.96, so the difference is not significant at the 5% level.
代入得到 z = 0.04 / √(0.39818 × 0.60182 × (1/1000 + 1/1200)) = 0.04 / 0.02094 = 1.911。对于双尾 5% 检验，临界值为 1.96，因此该差异在 5% 水平上不显著。
9. Limitations and Bias in Polling Data | 民调数据的局限与偏差
Real election polls often suffer from non-response bias, where certain groups are less likely to answer. Also, respondents may not reveal their true voting intention, especially in politically sensitive contexts. These issues mean that the binomial model’s assumption of independent and truthful responses is rarely perfectly met.
真实的选举民调往往存在无应答偏差，即某些群体不太可能参与回答。此外，受访者可能不会透露真实的投票意向，尤其是在政治敏感的情况下。这些问题意味着二项模型关于独立且真实回答的假设很少被完美满足。
Weighting adjustments are used by polling companies, but Edexcel exam questions usually focus on the mathematical methods rather than advanced survey methodology. You should still be able to state one limitation of using a binomial model for voting intention.
民调公司会使用加权调整，但 Edexcel 考题通常侧重于数学方法，而不是高级调查方法。你仍然应该能够说出使用二项模型模拟投票意向的一个局限性。
10. Exam-Style Worked Example | 考试风格例题
A national poll randomly selects 1600 adults and finds that 672 say they would vote Conservative. Conduct a two-tailed hypothesis test at the 1% significance level to determine whether the true proportion is different from 0.45.
一项全国民调随机抽取 1600 位成年人，其中 672 人表示会投票给保守党。在 1% 显著性水平下进行双尾假设检验，判断真实比例是否不同于 0.45。
H₀: p = 0.45
H₁: p ≠ 0.45
The sample proportion is p̂ = 672/1600 = 0.42. The test statistic is z = (0.42 − 0.45) / √(0.45 × 0.55 / 1600) = −0.03 / 0.01244 = −2.411. The critical values for a two-tailed 1% test are ±2.576. Since |−2.411| < 2.576, we do not reject H₀ at the 1% level.
样本比例为 p̂ = 672/1600 = 0.42。检验统计量为 z = (0.42 − 0.45) / √(0.45 × 0.55 / 1600) = −0.03 / 0.01244 = −2.411。双尾 1% 检验的临界值为 ±2.576。由于 |−2.411| < 2.576，我们在 1% 水平上不拒绝 H₀。
Conclusion: There is insufficient evidence at the 1% significance level to suggest that the true proportion of Conservative voters is different from 45%.
结论：在 1% 显著性水平下，没有足够证据表明保守党选民的真实比例不同于 45%。
11. Common Mistakes | 常见错误
- Forgetting to use p₀ from the null hypothesis in the standard error when testing a proportion.
- 在检验比例时，忘记在标准误差中使用原假设中的 p₀。
- Using p̂ instead of p₀ in the denominator of the z statistic.
- 在 z 统计量的分母中使用 p̂ 而不是 p₀。
- Not checking the conditions np > 5 and n(1 − p) > 5 before using the normal approximation.
- 在使用正态近似之前，没有检查条件 np > 5 和 n(1 − p) > 5。
- Confusing one-tailed and two-tailed critical values.
- 混淆单尾和双尾检验的临界值。
- Stating a conclusion that says ‘accept H₀’ instead of ‘do not reject H₀’.
- 结论写成“接受 H₀”而不是“不拒绝 H₀”。
12. Key Takeaways | 核心要点
The Conservative Party polling context provides a realistic setting for Edexcel A-Level Statistics. Remember that the sample proportion p̂ is a random variable, the binomial model underpins vote counts, and the normal approximation enables fast inference for large samples.
保守党民调背景为 Edexcel A-Level 统计提供了现实情境。请记住，样本比例 p̂ 是一个随机变量，二项模型是投票计数的基础，而正态近似使大样本的快速推断成为可能。
Confidence intervals quantify uncertainty, while hypothesis tests let you decide whether observed data contradict a claimed value. Always check the conditions, use the correct standard error, and write conclusions in the context of the original problem.
置信区间用于量化不确定性，而假设检验让你判断观测数据是否与声称的值相矛盾。务必检查条件、使用正确的标准误差，并在原始问题情境中撰写结论。
Published by TutorHao | Mathematics Revision Series | aleveler.com
更多咨询请联系16621398022（同微信）
屏轩国际教育cambridge primary/secondary checkpoint, cat4, ukiset,ukcat,igcse,alevel,PAT,STEP,MAT, ibdp,ap,ssat,sat,sat2课程辅导，国外大学本科硕士研究生博士课程论文辅导Cancel reply