Class-based Voting and Other Influences on Voting Patterns: A Statistical Approach | 阶级投票与其它影响投票模式的因素：统计学方法

📚 Class-based Voting and Other Influences on Voting Patterns: A Statistical Approach | 阶级投票与其它影响投票模式的因素：统计学方法

In A-Level Edexcel Mathematics, statistical methods provide powerful tools for analysing real-world data such as voting patterns. Understanding class-based voting and other influences requires techniques like sampling, contingency tables, chi-squared tests, and measures of association. This article explores how these mathematical approaches can be applied to measure and interpret the relationship between social class and voting behaviour, while also considering other variables such as age, region, and education. By linking statistical concepts directly to the study of voting influences, you will see how the Edexcel syllabus equips you to handle complex sociological data with mathematical rigour.

在A-Level Edexcel数学中，统计学方法为分析投票模式等现实数据提供了有力工具。理解阶级投票及其它影响因素需要运用抽样、列联表、卡方检验和关联性度量等技术。本文将探讨如何应用这些数学方法衡量和解读社会阶级与投票行为之间的关系，同时考虑年龄、地区和受教育程度等其它变量。通过将统计概念与投票影响研究直接联系起来，你将看到Edexcel课程如何使你能够以数学严谨性处理复杂的社会数据。

---

1. Voting Patterns as Categorical Data | 作为类别数据的投票模式

Voting behaviour is typically recorded as categorical (nominal) data, with categories such as ‘Party A’, ‘Party B’, or ‘Abstained’. Social class is often classified into discrete groups like ‘Working class’, ‘Middle class’, and ‘Upper class’. In mathematics, we handle such data using frequency tables and two-way contingency tables, which form the basis for analysing class-based voting. Each cell in a contingency table represents the number of individuals belonging to a specific class and voting choice, allowing us to investigate whether the two variables are independent.

投票行为通常被记录为类别（名义）数据，其类别如“政党A”、“政党B”或“弃权”。社会阶级通常被划分为离散的群体，如“工人阶级”、“中产阶级”和“上层阶级”。在数学中，我们使用频数表和双向列联表来处理这类数据，这些表格构成了分析阶级投票的基础。列联表中的每个单元格代表属于特定阶级并做出特定投票选择的人数，使我们能够研究这两个变量是否独立。

A simple example: a sample of 200 voters might yield a 2×3 contingency table with rows ‘Working class’ and ‘Middle class’ and columns ‘Left party’, ‘Right party’, ‘Other’. This table can then be used to calculate expected frequencies under the assumption of no association between class and voting—a key step in the chi-squared test for independence, which is part of the Edexcel Statistics specification.

一个简单例子：一个包含200名选民的样本可能产生一个2×3列联表，行是“工人阶级”和“中产阶级”，列是“左翼政党”、“右翼政党”、“其他”。该表格随后可用于计算在阶级与投票无关联的假设下的期望频数——这是独立性卡方检验的关键步骤，属于Edexcel统计学大纲内容。

---

2. Sampling Methods in Election Polling | 选举民调中的抽样方法

To draw conclusions about class-based voting, statisticians must first collect data through appropriate sampling techniques. The Edexcel syllabus covers simple random sampling, stratified sampling, quota sampling, and cluster sampling. When studying voting patterns, stratified sampling is particularly useful: the population is first divided into strata according to social class (and perhaps other factors), and then a random sample is taken from each stratum. This ensures that each class is adequately represented, reducing bias when estimating voting proportions.

要得出关于阶级投票的结论，统计人员必须首先通过适当的抽样技术收集数据。Edexcel大纲涵盖了简单随机抽样、分层抽样、配额抽样和整群抽样。在研究投票模式时，分层抽样尤其有用：先根据社会阶级（可能还有其它因素）将总体分为若干层，然后从每一层中随机抽取样本。这确保了每个阶级都得到充分代表，从而在估计投票比例时减少偏差。

Quota sampling is often used by polling organisations because it is cheaper and faster. An interviewer might be told to find 100 working-class voters and 100 middle-class voters. However, it may introduce selection bias as the choice of whom to interview within each quota is left to the interviewer. From a mathematical perspective, the reliability of any inference about class-based voting depends heavily on the sampling method and sample size, as determined by standard error and confidence intervals for proportions.

配额抽样常被民调机构所采用，因为它更便宜、更快。访员可能被告知要找到100名工人阶级选民和100名中产阶级选民。然而，这可能会引入选择偏差，因为每个配额内选择谁进行采访取决于访员。从数学角度看，关于阶级投票的任何推断的可靠性都高度依赖于抽样方法和样本容量，而这可以通过比例的标准误差和置信区间来确定。

---

3. Contingency Tables and Frequency Distributions | 列联表与频数分布

Once data on class and voting are collected, the first step is to summarise them using a two-way contingency table. This table displays the joint frequency distribution of the two categorical variables. For instance, a table may show that out of 150 working-class respondents, 80 voted for a left-wing party, 50 for a right-wing party, and 20 for others, while out of 100 middle-class respondents, the numbers were 30, 60, and 10 respectively. These observed frequencies can be used to calculate row percentages, column percentages, and overall proportions, giving an initial visual indication of class-based voting tendencies.

一旦收集了阶级和投票数据，第一步是使用双向列联表进行汇总。该表格显示了两个类别变量的联合频数分布。例如，一个表格可能显示，在150名工人阶级受访者中，80人投票给左翼政党，50人投票给右翼政党，20人投给其他政党；而在100名中产阶级受访者中，数字分别为30、60和10。这些观测频数可用于计算行百分比、列百分比和总体比例，从而初步直观地揭示阶级投票倾向。

In Edexcel A-Level, you will learn to construct these tables and interpret conditional distributions. For example, the proportion of working-class voters supporting left-wing parties is 80/150 ≈ 0.533, while for the middle class it is 30/100 = 0.300. Such disparities suggest that class may be an influence, but statistical testing is needed to determine if the association is significant or could have arisen by chance.

在Edexcel A-Level中，你将学习构建这些表格并解释条件分布。例如，工人阶级选民支持左翼政党的比例为80/150 ≈ 0.533，而中产阶级为30/100 = 0.300。这种差异表明阶级可能有影响，但需要进行统计检验来确定这种关联是显著的还是可能由偶然因素造成。

---

4. Chi-Squared Test for Independence | 卡方独立性检验

The chi-squared (χ²) test is the standard method for testing whether two categorical variables, such as class and voting choice, are independent. The test statistic is calculated as

χ² = Σ (O – E)² / E

where O is the observed frequency and E is the expected frequency under the null hypothesis of independence. The expected frequency for each cell is (Row Total × Column Total) / Grand Total. The test statistic is then compared to a critical value from the χ² distribution with (r−1)(c−1) degrees of freedom, where r and c are the numbers of rows and columns.

卡方（χ²）检验是检验两个类别变量（如阶级和投票选择）是否独立的标准方法。检验统计量计算公式为

χ² = Σ (O – E)² / E

其中O为观测频数，E为在独立性零假设下的期望频数。每个单元格的期望频数为（行合计 × 列合计）/ 总合计。然后将检验统计量与自由度为(r−1)(c−1)的χ²分布的临界值进行比较，其中r和c分别为行数和列数。

Suppose a contingency table has 2 rows and 3 columns. The degrees of freedom are (2−1)(3−1) = 2. At a 5% significance level, the critical value is 5.991. If computed χ² = 12.4, we reject the null hypothesis, concluding there is a significant association between class and voting. This is the mathematical foundation for claiming that class-based voting exists within that population.

假设一个列联表有2行3列。自由度为(2−1)(3−1) = 2。在5%显著性水平下，临界值为5.991。如果计算出的χ² = 12.4，我们拒绝零假设，得出结论：阶级与投票之间存在显著关联。这是声称在该总体中存在阶级投票的数学基础。

---

5. Measuring the Strength of Association | 度量关联强度

While the chi-squared test tells us whether an association exists, it does not indicate the strength of that association. For this, we can use the phi coefficient (Φ) in 2×2 tables, or Cramér’s V for larger tables. For a 2×2 table, Φ = √(χ² / n), where n is the total sample size. The value lies between 0 (no association) and 1 (perfect association). In larger tables, Cramér’s V is used: V = √[χ² / (n × min(r−1, c−1))].

虽然卡方检验能告诉我们关联是否存在，但它并不表明关联的强度。为此，在2×2表中我们可以使用phi系数（Φ），在更大的表中使用Cramér’s V。对于2×2表，Φ = √(χ² / n)，其中n为总样本量。该值在0（无关联）到1（完全关联）之间。在更大的表中，使用Cramér’s V：V = √[χ² / (n × min(r−1, c−1))]。

For example, if χ² = 12.4 with n = 250 and min(r−1, c−1) = 1, then Φ = √(12.4/250) ≈ 0.223. This suggests a small to moderate effect of class on voting. These measures allow us to compare the influence of class with other factors such as age or education, by computing separate coefficients for each. The Edexcel syllabus includes the concept of correlation for quantitative data; these are the categorical analogues.

例如，如果χ² = 12.4，n = 250，且min(r−1, c−1) = 1，则Φ = √(12.4/250) ≈ 0.223。这表明阶级对投票有微弱到中等程度的影响。这些度量允许我们通过计算各自的系数来比较阶级与年龄或教育等其它因素的影响。Edexcel大纲包含了定量数据相关性的概念；这些正是类别数据的对应物。

---

6. Age, Region, and Education as Confounding Variables | 年龄、地区和受教育程度作为混杂变量

Class is not the sole influence on voting. Other variables—such as age, geographic region, and level of education—often correlate with both class and voting choice, potentially confounding the observed relationship. Mathematically, we can control for such variables by stratifying the data. For example, we might create separate contingency tables for different age groups (<30, 30–50, >50) and then perform chi-squared tests or compute Cramér’s V within each stratum. This reveals whether class-based voting persists after removing the age effect.

阶级并非投票的唯一影响因素。其它变量——如年龄、地理区域和受教育程度——通常与阶级和投票选择都相关，可能会混淆观察到的关系。在数学上，我们可以通过分层数据来控制这些变量。例如，我们可以为不同年龄组（<30岁、30–50岁、>50岁）创建单独的列联表，然后在每个层内进行卡方检验或计算Cramér’s V。这将揭示在剔除年龄效应后阶级投票是否仍然存在。

Alternatively, logistic regression (beyond the scope of A-Level but conceptually linked) builds a model predicting the probability of voting for a particular party based on several predictors simultaneously. In A-Level terms, we can use descriptive statistics and hypothesis tests to examine each factor in turn, always mindful that correlation does not imply causation. A scatter diagram of constituency-level data might show a correlation between the proportion of working-class residents and the vote share for a certain party, but this could be influenced by regional economic factors.

或者，逻辑回归（超出A-Level范围但概念上有联系）建立一个基于多个预测变量同时预测投票给某一政党概率的模型。用A-Level的说法，我们可以使用描述性统计和假设检验逐一考察每个因素，始终牢记相关并不意味着因果。选区层面的数据散点图可能显示工人阶级居民比例与某政党得票率之间的相关性，但这可能受到地区经济因素的影响。

---

7. Normal Distribution and Polling Error | 正态分布与民调误差

When opinion polls estimate the proportion of a class voting for a given party, the sample proportion is a point estimate. According to the Central Limit Theorem, for large samples, the sampling distribution of a proportion is approximately normal. This allows us to construct confidence intervals and margins of error. The standard error for a proportion p̂ from a sample of size n is √[p̂(1−p̂)/n]. A 95% confidence interval is p̂ ± 1.96 × SE.

当民调估计某阶级投票给特定政党的比例时，样本比例是一个点估计。根据中心极限定理，对于大样本，比例的抽样分布近似正态。这使得我们能够构建置信区间和误差幅度。来自容量为n的样本的比例p̂的标准误差为√[p̂(1−p̂)/n]。95%置信区间为p̂ ± 1.96 × SE。

For example, if a poll of 400 working-class voters finds 55% supporting Party X, the standard error is √(0.55×0.45/400) ≈ 0.0249. The margin of error is 1.96×0.0249 ≈ 0.0488, so we are 95% confident that the true proportion lies between 50.1% and 59.9%. This statistical reasoning is essential when journalists and analysts discuss class-based voting trends—the apparent lead could simply be sampling variation.

例如，如果一项对400名工人阶级选民的民调发现55%支持X党，标准误差为√(0.55×0.45/400) ≈ 0.0249。误差幅度为1.96×0.0249 ≈ 0.0488，因此我们有95%的信心认为真实比例在50.1%至59.9%之间。当记者和分析人士讨论阶级投票趋势时，这种统计推理是不可或缺的——表面的领先可能仅仅是抽样波动造成的。

---

8. Hypothesis Testing for a Single Proportion in a Class | 单一阶级比例的假设检验

We might wish to test whether the proportion of working-class voters supporting a particular party has changed from a historical value. This is a two-tailed (or one-tailed) hypothesis test for a binomial proportion using the normal approximation. The test statistic is

z = (p̂ − p₀) / √[p₀(1−p₀)/n]

where p₀ is the hypothesised proportion. In Edexcel, you must check that np₀ and n(1−p₀) are both greater than 10 to use the normal approximation.

我们可能希望检验工人阶级选民支持某特定政党的比例是否较历史值发生了变化。这是一个使用正态近似的针对二项分布比例的双尾（或单尾）假设检验。检验统计量为

z = (p̂ − p₀) / √[p₀(1−p₀)/n]

其中p₀为假设比例。在Edexcel中，必须检查np₀和n(1−p₀)都大于10方可使用正态近似。

Suppose historically 40% of middle-class voters supported a conservative party. In a recent survey of 200 middle-class voters, 100 supported that party, giving p̂ = 0.5. The test statistic is z = (0.5−0.4)/√(0.4×0.6/200) ≈ 2.89. At the 5% significance level (two-tailed critical value 1.96), we reject the null hypothesis and conclude there is evidence of a shift in middle-class voting behaviour. This method can be applied to any subgroup defined by class or another factor.

假设历史上40%的中产阶级选民支持一个保守党。在最近一次对200名中产阶级选民的调查中，100人支持该党，得到p̂ = 0.5。检验统计量为z = (0.5−0.4)/√(0.4×0.6/200) ≈ 2.89。在5%显著性水平下（双尾临界值1.96），我们拒绝零假设，得出结论：有证据表明中产阶级的投票行为发生了转变。该方法适用于任何由阶级或其它因素定义的子群体。

---

9. Comparing Two Proportions across Classes | 跨阶级比例比较

Another common question is whether the difference in voting proportions between two classes is statistically significant. For two independent samples of sizes n₁ and n₂, with sample proportions p̂₁ and p̂₂, the test statistic under the null hypothesis of equal proportions uses a pooled estimate p̂ = (x₁ + x₂)/(n₁ + n₂) and standard error √[p̂(1−p̂)(1/n₁ + 1/n₂)]. The test statistic z = (p̂₁ − p̂₂) / SE follows a standard normal distribution when conditions are met.

另一个常见问题是，两个阶级之间投票比例的差异是否具有统计显著性。对于两个容量分别为n₁和n₂的独立样本，以及样本比例p̂₁和p̂₂，在等比例的零假设下，检验统计量使用合并估计值p̂ = (x₁ + x₂)/(n₁ + n₂)及标准误差√[p̂(1−p̂)(1/n₁ + 1/n₂)]。当条件满足时，检验统计量z = (p̂₁ − p̂₂) / SE服从标准正态分布。

Imagine a sample of 250 working-class voters shows 60% support for a policy, while a sample of 180 middle-class voters shows 45%. Pooled p̂ = (0.6×250 + 0.45×180)/(250+180) = (150+81)/430 ≈ 0.5372. SE = √[0.5372×0.4628×(1/250+1/180)] ≈ 0.0527. z = (0.6−0.45)/0.0527 ≈ 2.85, significant at 5%. This method—comparing two population proportions—is extensively applied in analysing class-based voting differences.

假设一个250名工人阶级选民的样本显示60%支持某政策，而一个180名中产阶级选民的样本显示45%的支持率。合并p̂ = (0.6×250 + 0.45×180)/(250+180) = (150+81)/430 ≈ 0.5372。SE = √[0.5372×0.4628×(1/250+1/180)] ≈ 0.0527。z = (0.6−0.45)/0.0527 ≈ 2.85，在5%水平上显著。这种比较两个总体比例的方法广泛用于分析阶级投票差异。

---

10. Correlation and Regression: Turnout and Class Indicators | 相关与回归：投票率与阶级指标

Voting turnout (a continuous variable) can also be related to class indicators such as average income or percentage of manual workers in a constituency. The Edexcel Statistics module covers Pearson’s product-moment correlation coefficient r and linear regression. For bivariate data, we calculate r to measure the strength and direction of a linear relationship between, say, constituency turnout and the proportion of working-class residents. The hypothesis test for zero correlation uses the test statistic t = r√[(n−2)/(1−r²)] with n−2 degrees of freedom.

投票率（一个连续变量）也可以与阶级指标相关，如选区内的平均收入或体力工人比例。Edexcel统计模块涵盖了皮尔逊积矩相关系数r和线性回归。对于双变量数据，我们计算r来衡量选区投票率与工人阶级居民比例之间线性关系的强度和方向。零相关假设检验使用检验统计量t = r√[(n−2)/(1−r²)]，自由度为n−2。

If r = −0.45 for n = 30 constituencies, t = −0.45√(28/(1−0.2025)) ≈ −2.67. Comparing to a critical t-value with 28 df (two-tail 5% ≈ 2.048), we conclude there is a significant negative linear relationship: higher working-class proportion is associated with lower turnout. However, many other influences—such as marginality of the seat, education levels—can also be incorporated by using multiple regression conceptually, though A-Level focuses on simple linear regression.

如果对30个选区有r = −0.45，t = −0.45√(28/(1−0.2025)) ≈ −2.67。与自由度为28时的临界t值（双尾5% ≈ 2.048）比较，我们得出结论：存在显著的负线性关系：工人阶级比例越高，投票率越低。然而，许多其它影响因素——如席位的边缘性、教育水平——也可以通过概念上的多元回归予以纳入，尽管A-Level侧重于简单线性回归。

---

11. Limitations and Validity of Statistical Conclusions | 统计结论的局限性与有效性

Mathematical analyses of voting patterns must acknowledge limitations. Data may exhibit ecological fallacy if constituency-level data is used to infer individual behaviour. Sampling errors, non-response bias, and misclassification of social class all affect the validity. In hypothesis testing, the significance level (α) controls the risk of Type I error, but with multiple tests, the risk of false positives increases. Moreover, a significant association does not prove that class causes voting behaviour; other hidden factors may drive both.

对投票模式的数学分析必须承认其局限性。如果使用选区层面数据推断个体行为，数据可能表现出生态学谬误。抽样误差、无应答偏差和社会阶级的错误分类均会影响有效性。在假设检验中，显著性水平（α）控制了第一类错误的风险，但若进行多次检验，误报的风险会上升。此外，显著关联并不证明阶级导致投票行为；其它隐藏因素可能同时驱使两者。

In your Edexcel exam, you may be asked to comment on the reliability of data or the appropriateness of a statistical model. For instance, the normal approximation for binomial proportion tests requires sufficient sample size, and the chi-squared test requires that no more than 20% of expected frequencies fall below 5 and none below 1. Real voting data often violates these assumptions when subgroups are small, requiring use of Yates’ correction or Fisher’s exact test—though only the basic chi-squared test is required for Edexcel.

在Edexcel考试中，你可能需要评论数据的可靠性或统计模型的适当性。例如，二项比例检验的正态近似需要足够的样本量，而卡方检验要求期望频数低于5的单元格比例不超过20%，且没有一个低于1。当子群体较小时，真实的投票数据常常违反这些假设，需使用Yates矫正或Fisher确切检验——不过Edexcel仅要求基本的卡方检验。

---

12. Summary and Exam Relevance | 总结与考试关联

Class-based voting and other influences on voting patterns offer a rich context for applying A-Level Edexcel Statistics. From designing surveys with stratified sampling, to analysing contingency tables with chi-squared tests, to testing proportions and correlations, the mathematical tools directly address questions that political scientists ask. By framing these sociological problems in statistical terms, you not only prepare for exam questions on data handling and hypothesis testing but also appreciate the power and limitations of numerical evidence in shaping our understanding of society.

阶级投票及其它影响投票模式的因素为应用A-Level Edexcel统计知识提供了丰富的背景。从设计分层抽样调查，到运用卡方检验分析列联表，再到检验比例和相关关系，数学工具直接回答了政治学家提出的问题。通过用统计术语构架这些社会学问题，你不仅为数据处理和假设检验的考试题目做好了准备，也能体会到数值证据在塑造我们对社会的理解方面的力量和局限。

Remember to clearly state hypotheses, check conditions, calculate expected values, and interpret p-values or critical values correctly. Use contingency tables, confidence intervals, and correlation coefficients to quantify class-based voting. Mastering these techniques will enhance your data analysis skills and enable you to tackle any Edexcel question involving categorical or quantitative data in a social science context.

请记住要清晰陈述假设，检查条件，计算期望值，并正确解读p值或临界值。使用列联表、置信区间和相关系数来量化阶级投票。掌握这些技术将提升你的数据分析技能，使你能够处理Edexcel中任何涉及社会科学背景下的类别或定量数据的问题。

Published by TutorHao | Mathematics Revision Series | aleveler.com

更多咨询请联系16621398022（同微信）