📚 Minority Rights: The Role of Stratified Sampling in Fair Representation | 少数群体权利：分层抽样在公平代表中的作用
In democratic societies, the rights of minority groups are protected not only through legislation but also through accurate representation in data and surveys. When national statistics, health studies or opinion polls fail to include minority voices, policies can become biased against those who are already vulnerable. For A‑Level Mathematics students, understanding how sampling methods can either undermine or uphold minority rights provides a powerful real‑world application of statistical techniques. This article explores the connection between fair representation and stratified sampling, showing how a mathematical approach can help secure a voice for every group.
在民主社会中，少数群体的权利不仅通过立法得到保护，也通过在数据和调查中得到准确代表来维护。当国家统计、健康研究或民意调查未能纳入少数群体的声音时，政策可能会对原本就脆弱的群体产生偏见。对于A‑Level数学学生而言，理解抽样方法如何可能削弱或维护少数群体权利，是统计技术的一个强有力的现实应用。本文探讨公平代表性与分层抽样之间的联系，展示数学方法如何帮助确保每个群体都能发出自己的声音。
1. Core Principles of Sampling in A‑Level Mathematics | A‑Level数学中的抽样核心原则
Edexcel A‑Level Mathematics introduces sampling as a fundamental concept in statistics. A population is the whole set of items that are of interest, while a sample is a subset of individuals selected from that population. The sampling frame is a list of all sampling units from which the sample is drawn. Common methods include simple random sampling, systematic sampling, stratified sampling and quota sampling. Each method has different properties regarding bias, efficiency and fairness, making the choice of method crucial when minority representation is a concern.
Edexcel A‑Level数学将抽样作为统计学的基本概念引入。总体是感兴趣的全部对象集合，而样本是从该总体中选出的个体子集。抽样框是所有抽样单位的名单。常见的方法包括简单随机抽样、系统抽样、分层抽样和配额抽样。每种方法在偏差、效率和公平性方面具有不同的特性，因此当少数群体代表性问题受到关注时，抽样方法的选择至关重要。
2. The Challenge of Underrepresentation in Simple Random Sampling | 简单随机抽样中代表不足的挑战
Simple random sampling (SRS) gives every member of the population an equal chance of being selected. While this sounds fair, it often works against minority groups. If a minority makes up only 5% of the population, a random sample of size 200 is expected to contain just 10 individuals from that group. However, due to random variation, the actual number could be lower — or even zero. The probability of not picking any member of a minority group of proportion p in a sample of size n 00004, which seems very small. But when surveys are repeated many times or when the subgroup is extremely small (e.g., 0.5%), the risk of exclusion becomes unacceptably high.
简单随机抽样（SRS）让总体中每个成员都有相同的机会被选入样本。这听起来公平，但往往对少数群体不利。如果少数族裔仅占总体的5%，一个容量为200的随机样本预期只包含该群体的10人。然而，由于随机波动，实际人数可能更低——甚至为零。在一个容量为n的样本中，完全没有抽到占比为p的少数群体成员的概率约为(1 − p)ⁿ。当p = 0.05且n = 200时，概率为(0.95)²⁰⁰ ≈ 0.00004，看起来很小。但当调查重复多次，或子群体极其微小（例如0.5%）时，被排除的风险就会高到不可接受。
3. Stratified Sampling: A Fairer Approach | 分层抽样：一种更公平的方法
Stratified sampling overcomes this weakness by dividing the population into distinct, non‑overlapping groups called strata (singular: stratum). Strata are based on characteristics relevant to the study, such as ethnicity, age or income. The sample is then selected so that the proportion of each stratum in the sample matches its proportion in the population. The key formula is:
nh = (Nh ÷ N) × n
where Nh is the size of stratum h, N is the total population size, n is the total sample size, and nh is the number to be sampled from stratum h. This guarantees that every group, including minorities, is represented in exact proportion to its size, giving every stratum an assured voice.
分层抽样通过将总体划分为互不重叠的组（称为层）来克服这一弱点。层的划分基于与研究相关的特征，如种族、年龄或收入。然后按总体中各层比例选取样本，使得样本中每一层的比例与其在总体中的比例一致。关键公式为：
nh = (Nh ÷ N) × n
其中Nh是层h的个体数，N是总体总数，n是样本总数，nh是从层h中抽取的数量。这确保了包括少数群体在内的每一个组都能按其规模获得精确比例的代表，从而保证了每个层都有确定的话语权。
4. Step‑by‑Step Calculation for Minority Strata | 少数群体的逐层计算
Suppose a town has a population of 80,000. Researchers want to survey residents' access to healthcare, paying special attention to an ethnic minority community that numbers 2,400. The total sample size is set at 400. Using stratified sampling with ethnicity as the stratification variable:
Nminority = 2400, N = 80,000, n = 400
nminority = (2400 ÷ 80000) × 400 = 0.03 × 400 = 12
So exactly 12 members of the minority group are selected. In contrast, a simple random sample of 400 would give an expected count of 12, but could easily yield 7 or 8 because of sampling variability, weakening the reliability of estimates for that group.
假设某镇有80,000人。研究者希望调查居民获得医疗保健的情况，并特别关注一个只有2,400人的少数族裔社区。总样本量定为400。使用以族裔为分层变量的分层抽样：
N少数 = 2400，N = 80,000，n = 400
n少数 = (2400 ÷ 80000) × 400 = 0.03 × 400 = 12
因此恰好有12名少数群体成员被选入样本。相比之下，容量为400的简单随机样本预期人数为12，但由于抽样波动很容易只抽到7或8人，从而削弱了对该群体估计的可靠性。
5. Handling Very Small Strata: Minimum Representation | 处理极小层：最小代表性问题
A practical difficulty arises when a stratum is so small that the proportionate allocation gives a number below 1. For example, if a minority group comprises only 50 out of 100,000 and the sample size is 200, the formula gives 0.1. Rounding down to 0 would exclude the group entirely, violating the spirit of fair representation. Statisticians often impose a rule that every stratum must have at least one sampling unit, or they deliberately oversample very small strata to allow meaningful analysis. Oversampling then requires weighting adjustments when computing overall estimates.
当某一层小到按比例分配的结果小于1时，就会出现实际困难。例如，如果少数群体只有100,000人中的50人，而样本量为200，公式给出0.1。向下取整为0会将这个群体完全排除，违背了公平代表的精神。统计学家通常规定每个层至少抽取一个单位，或者故意对极小层进行超额抽样，以便进行有意义的分析。超额抽样后，在计算总体估计值时必须进行加权调整。
nh,adjusted = max(1, round(nh))
6. Stratified vs. Quota Sampling: Implications for Minority Rights | 分层抽样与配额抽样：对少数群体权利的影响
Quota sampling, another non‑probability method, is often used in market research. The interviewer is told to find a certain number of respondents from each subgroup, but the selection within the subgroup is left to convenience or judgement. This can seriously undermine minority rights because interviewers might, consciously or unconsciously, approach the most accessible or most cooperative individuals, introducing bias. Stratified random sampling, by contrast, uses a random mechanism within each stratum, giving every member of the minority group a known, non‑zero probability of selection.
配额抽样是另一种非概率抽样方法，常用于市场研究。访问员被告知要从每个子群中寻找一定数量的受访者，但子群内的选择则交给便利性或判断力。这可能严重损害少数群体权利，因为访问员可能有意无意地选择最容易接触或最合作的人，从而引入偏差。相比之下，分层随机抽样在每个层内采用随机机制，让少数群体的每个成员都有一个已知的、非零的入选概率。
7. Examining Bias and Accuracy in Minority Estimates | 考察少数群体估计中的偏差与准确性
Even with stratified sampling, bias can creep in through an incomplete sampling frame. If the list used to identify minority members (such as electoral registers or medical records) misses undocumented individuals or those who have recently moved, the stratum will be under‑covered. The mathematical consequence is that the estimator for the minority mean becomes biased. Students can quantify this by comparing the expected value of the sample mean with the true population mean. In Edexcel examinations, questions often ask candidates to identify sources of bias and suggest improvements, such as using multiple frames or post‑enumeration surveys.
即使采用分层抽样，不完整的抽样框也可能引入偏差。如果用来识别少数群体成员的名册（如选民登记册或医疗记录）遗漏了无证件者或刚迁居的人，该层就会被覆盖不足。其数学后果是，对少数群体均值的估计量会产生偏差。学生可以通过比较样本均值的期望值与真实总体均值来量化这一点。在Edexcel考试中，试题经常要求考生找出偏差来源并提出改进建议，例如使用多个抽样框或进行事后计数调查。
8. Real‑World Applications: Government Surveys and Policy | 实际应用：政府调查与政策
Many national statistical institutes, including the UK Office for National Statistics (ONS), use stratified designs to ensure that minority ethnic groups, disabled people and other small subpopulations are adequately represented. For instance, the Labour Force Survey stratifies by region and then boosts samples in areas with high minority‑ethnic density. Without such stratification, policy decisions on language support services, healthcare outreach or educational funding would rest on weak data that systematically ignore the needs of the most marginalised.
许多国家统计机构，包括英国国家统计局，采用分层设计以确保少数族裔、残障人士和其他小型子群体得到充分代表。例如，劳动力调查先按地区分层，然后在少数族裔聚居区增加样本。如果缺乏这样的分层，有关语言支持服务、医疗外展服务或教育经费的政策决策就将建立在薄弱的数据之上，系统性地忽视最边缘群体的需求。
9. Exam‑Style Question and Worked Solution | 考试风格问题与详细解答
Question: A school has 1,200 students. The head teacher wishes to survey attitudes to canteen food. The school has 60 students from a refugee background — a group that often has specific dietary requirements. A stratified sample of 150 students is to be taken using year group and refugee status as strata. The refugee students are treated as a separate stratum.
(a) Calculate the number of refugee students to be sampled.
(b) Explain why this stratified design is better than a simple random sample for capturing the needs of the refugee group.
(c) Suggest one practical difficulty in constructing the sampling frame for this survey.
问题：某学校有1200名学生。校长希望调查学生对食堂伙食的态度。学校有60名具有难民背景的学生——该群体常有特殊的饮食需求。拟按年级和难民身份分层，抽取150名学生的分层样本。难民学生作为一个独立的层。
(a) 计算应抽取的难民学生人数。
(b) 解释为什么这种分层设计比简单随机抽样更能体现难民群体的需求。
(c) 提出构建该调查抽样框时的一个实际困难。
Solution (a): Nrefugee = 60, N = 1200, n = 150
nrefugee = (60 ÷ 1200) × 150 = 0.05 × 150 = 7.5
You cannot select half a student, so the number is rounded to 8. (Many mark schemes accept 7 or 8, with justification.)
Solution (b): In a simple random sample, the refugee group might be missed or represented by only 2 or 3 students, making it impossible to draw reliable conclusions about their dietary needs. Stratified sampling guarantees that roughly the same proportion is selected, so their views are definitely included in the survey.
Solution (c): One difficulty is identifying all refugee students while respecting their privacy — some families may not wish to disclose their status, so the sampling frame may be incomplete.
解答(a): N难民 = 60，N = 1200，n = 150
n难民 = (60 ÷ 1200) × 150 = 0.05 × 150 = 7.5
不能选半个学生，故人数取为8。（许多评分标准接受7或8，但需说明理由。）
解答(b): 在简单随机抽样中，难民群体可能被遗漏或者仅仅由2、3名学生代表，这使得无法对其饮食需求得出可靠结论。分层抽样保证了大致相同的比例被选入样本，因此他们的意见必定会被纳入调查。
解答(c): 一个困难是在尊重隐私的前提下识别所有难民学生——有些家庭可能不愿披露其身份，因此抽样框可能不完整。
10. Mathematical Principles Behind Oversampling and Weighting | 超额抽样与加权的数学原理
When a minority stratum is deliberately oversampled — say, sampling at twice its proportional share — the raw sample mean is no longer an unbiased estimator of the population mean. Statisticians apply weights to correct for this. If the minority group has a sampling fraction fh = nh / Nh ll sampling fraction f = n / N, then each individual from that stratum receives a weight wh = (Nh / nh) / (N / n) = (f / fh). The weighted estimator is then
x̄weighted = ( Σ wh xi ) ÷ ( Σ wh )
This technique allows researchers to obtain precise information about a minority without distorting national estimates, striking a balance between depth and fairness.
当对少数层进行故意超额抽样时——例如，抽取比例份额的两倍——原始样本均值就不再是总体均值的无偏估计量。统计学家采用权重进行修正。如果少数群体的抽样比fh = nh / Nh高于总抽样比f = n / N，那么该层每个个体获得的权重为wh = (Nh / nh) / (N / n) = (f / fh)。加权估计量即为
x̄weighted = ( Σ wh xi ) ÷ ( Σ wh )
这一技术使研究者既能够获得关于少数群体的精确信息，又不会扭曲全国性估计值，实现了深度与公平的平衡。
11. Conclusion: Upholding Rights Through Mathematical Design | 结论：通过数学设计维护权利
Minority rights in statistics are not an abstract political concept — they are a direct consequence of sampling design. A‑Level Mathematics equips students with the tools to evaluate and construct sampling strategies that give every voice a chance to be heard. By mastering stratified sampling, oversampling corrections and bias analysis, learners see that numbers can serve as guardians of inclusion, ensuring that the needs of the few are never lost in the noise of the many.
统计中的少数群体权利并非抽象的政治概念——它们是抽样设计的直接结果。A‑Level数学为学生提供了工具，去评价和构建让每一个声音都有机会被听到的抽样策略。通过掌握分层抽样、超额抽样修正和偏差分析，学习者会认识到，数字可以成为包容的守护者，确保少数群体的需求永远不会淹没在多数群体的噪声之中。
Published by TutorHao | Mathematics Revision Series | aleveler.com
更多咨询请联系16621398022（同微信）
或截图后用「扫一扫 → 相册」选择本图