Hypothesis Testing: Key Points for Edexcel Maths | 假设检验考点精讲

📚 Hypothesis Testing: Key Points for Edexcel Maths | 假设检验考点精讲

Hypothesis testing is a formal statistical procedure used to decide whether there is enough evidence in a sample of data to infer that a certain condition holds for the entire population. In the Edexcel A Level Mathematics specification, hypothesis tests are applied to binomial distributions, normal distributions, and occasionally Poisson distributions. Mastering this topic requires a clear understanding of null and alternative hypotheses, significance levels, critical regions, p-values, and the ability to interpret conclusions in context.

假设检验是一种正式统计程序，用于判断样本数据中是否有足够证据推断总体满足某一条件。在 Edexcel A Level 数学大纲中，假设检验应用于二项分布、正态分布，有时也涉及泊松分布。掌握本主题需要清晰理解零假设与备择假设、显著性水平、临界域、p 值，并能在情境中正确解读结论。

1. What is Hypothesis Testing? | 什么是假设检验？

Hypothesis testing begins with a statement about a population parameter, such as the mean or proportion. We collect sample data and use probability theory to assess how likely the observed result is, assuming the initial statement is true. If the result is highly improbable, we reject the initial statement in favour of an alternative explanation.

假设检验从关于总体参数（如均值或比例）的陈述开始。我们采集样本数据，并利用概率理论评估在最初陈述为真的情况下观察到当前结果的可能性。如果结果极不可能发生，我们就拒绝最初的陈述，转而接受另一种解释。

It is essential to remember that hypothesis testing never proves a hypothesis to be true; it only provides evidence against the null hypothesis. The conclusion must be framed in terms of the original problem, using phrases such as ‘there is sufficient evidence to suggest that…’ or ‘there is insufficient evidence to reject…’.

必须记住，假设检验永远不能证明一个假设为真；它只提供反对零假设的证据。结论必须结合原始问题，使用诸如“有充分证据表明……”或“没有足够证据拒绝……”这样的措辞。

2. Null and Alternative Hypotheses | 零假设与备择假设

The null hypothesis, denoted H₀, is the statement being tested. It typically represents ‘no effect’ or ‘no difference’, and is assumed to be true until evidence suggests otherwise. The alternative hypothesis, denoted H₁ or Hₐ, represents the claim we are trying to find evidence for.

零假设，记作 H₀，是被检验的陈述。它通常表示“无效果”或“无差异”，在证据表明相反之前假设为真。备择假设，记作 H₁ 或 Hₐ，代表我们试图寻找证据支持的论断。

For example, if a company claims that no more than 5% of its products are defective, we set H₀: p = 0.05 and H₁: p > 0.05 for a one-tailed test. In a two-tailed test, H₁ would be p ≠ 0.05. The hypotheses must be stated precisely in terms of the population parameter, using proper notation.

例如，如果一家公司声称其产品不合格率不超过 5%，则对于单尾检验，我们设 H₀: p = 0.05，H₁: p > 0.05。对于双尾检验，H₁ 应为 p ≠ 0.05。假设必须使用正确记号，以总体参数的形式精确陈述。

3. Significance Level and Critical Region | 显著性水平与临界域

The significance level, typically denoted by α, is the probability of rejecting the null hypothesis when it is actually true. Common values are 0.05, 0.01, and 0.10. In Edexcel exam questions, the significance level is usually given, and students must identify the critical region accordingly.

显著性水平，通常记作 α，是当零假设实际为真时拒绝它的概率。常用取值有 0.05、0.01 和 0.10。在 Edexcel 考试题中，显著性水平通常会给出，学生必须据此确定临界域。

The critical region (or rejection region) is the set of values of the test statistic for which the null hypothesis is rejected. For a binomial test with n = 20 and H₁: p > 0.5 at α = 0.05, the critical region might be {15, 16, …, 20} if P(X ≥ 15) ≤ 0.05. The boundary value that separates the critical region from the acceptance region is called the critical value.

临界域（或拒绝域）是检验统计量的一组取值，当统计量落入该范围时拒绝零假设。对于 n = 20，H₁: p > 0.5，α = 0.05 的二项检验，如果 P(X ≥ 15) ≤ 0.05，临界域可能是 {15, 16, …, 20}。将临界域与接受域分开的边界值称为临界值。

4. Test Statistic and p-Value | 检验统计量与 p 值

A test statistic is a single value calculated from the sample data, used to measure the degree of agreement between the data and the null hypothesis. In the binomial setting, the test statistic is simply the observed number of successes. In a normal test, it might be the sample mean or the z-statistic.

检验统计量是根据样本数据计算出的单个数值，用于衡量数据与零假设之间的一致性程度。在二项分布情境下，检验统计量就是观察到的成功次数。在正态检验中，它可能是样本均值或 z 统计量。

The p-value is the probability of obtaining a test statistic at least as extreme as the one observed, assuming the null hypothesis is true. If the p-value is less than or equal to α, we reject H₀. The p-value approach is particularly useful when critical values are not readily available, and it is increasingly emphasised in Edexcel mark schemes.

p 值是在零假设为真的条件下，得到至少与观察值同样极端的检验统计量的概率。如果 p 值小于或等于 α，我们就拒绝 H₀。当临界值不易获取时，p 值法尤其有用，且在 Edexcel 评分方案中越来越受重视。

5. One-Tailed vs Two-Tailed Tests | 单尾与双尾检验

A one-tailed test is used when the alternative hypothesis specifies a direction of interest, such as ‘greater than’ or ‘less than’. The entire significance level is placed in one tail of the distribution. For instance, H₁: p < 0.3 focuses only on the lower tail.

单尾检验用于备择假设指明了关注方向的情况，例如“大于”或“小于”。整个显著性水平被置于分布的单一尾部。例如，H₁: p < 0.3 仅关注下尾。

A two-tailed test is appropriate when H₁ indicates any difference from the hypothesised value, e.g., H₁: p ≠ 0.3. The significance level is split equally between the two tails. In a normal test, this means finding two symmetrical critical values. Students often need to halve the significance level when looking up critical values for a two-tailed test.

双尾检验适用于 H₁ 表示与假设值有任何差异的情况，如 H₁: p ≠ 0.3。显著性水平被平分到两个尾部。在正态检验中，这意味着要找出两个对称的临界值。进行双尾检验查表时，学生常需将显著性水平减半来寻找临界值。

6. Hypothesis Testing for Binomial Distribution | 二项分布的假设检验

This is one of the most common hypothesis test scenarios in Edexcel A Level Maths. The binomial distribution B(n, p) models the number of successes in n independent trials, each with constant probability p. The population parameter under test is p, the probability of success.

这是 Edexcel A Level 数学中最常见的假设检验情境之一。二项分布 B(n, p) 描述了在 n 次独立试验中成功的次数，每次试验具有恒定的成功概率 p。被检验的总体参数是成功概率 p。

To perform the test, find the probability of obtaining the observed result, or a more extreme one, using cumulative binomial probabilities. For a one-tailed test with H₁: p > value, calculate P(X ≥ observed value) under H₀. Compare this p-value with α, or identify the critical region by finding the smallest x such that P(X ≥ x) ≤ α.

进行检验时，使用累积二项概率求出获得观察结果或更极端结果的概率。对于 H₁: p > 某值的单尾检验，计算 H₀ 下的 P(X ≥ 观察值)。将此 p 值与 α 比较，或通过找出使 P(X ≥ x) ≤ α 的最小 x 来确定临界域。

Edexcel often asks for a conclusion in context. For example: ‘Since 0.021 < 0.05, we reject H₀. There is sufficient evidence at the 5% significance level to suggest that the proportion of defective items has increased.' Always mention the significance level and direction of effect.

Edexcel 常要求在情境中写出结论。例如：“由于 0.021 < 0.05，我们拒绝 H₀。在 5% 显著性水平下有充分证据表明次品率已经上升。”务必提及显著性水平和效应方向。

7. Hypothesis Testing for Normal Distribution | 正态分布的假设检验

When the population is normally distributed and the population variance σ² is known, we use the standard normal distribution to test hypotheses about the population mean μ. The test statistic is Z = (x̄ – μ₀) / (σ/√n), where μ₀ is the hypothesised mean under H₀.

当总体服从正态分布且总体方差 σ² 已知时，我们使用标准正态分布来检验关于总体均值 μ 的假设。检验统计量为 Z = (x̄ – μ₀) / (σ/√n)，其中 μ₀ 是 H₀ 下的假设均值。

Z = (x̄ – μ₀) / (σ/√n) ~ N(0, 1) under H₀

If the population variance is unknown and the sample size is large, the sample variance s² can be used, and the normal approximation is still valid. Critical values are obtained from the normal distribution table. For a 5% one-tailed test, the critical z-value is 1.645; for a two-tailed test, it is ±1.96.

如果总体方差未知但样本量较大，可使用样本方差 s²，正态近似仍然有效。临界值从正态分布表中获得。对于 5% 单尾检验，临界 z 值为 1.645；对于双尾检验，为 ±1.96。

Edexcel questions may provide summary statistics and expect you to calculate the test statistic, compare it to the critical value, or find the p-value using the standard normal table. Always state whether the result is significant and relate it back to the claim.

Edexcel 考题可能提供汇总统计量，要求你计算检验统计量，与临界值比较，或利用标准正态表求出 p 值。务必说明结果是否显著，并将其与原始论断联系起来。

8. Hypothesis Testing for Poisson Distribution | 泊松分布的假设检验

Although less frequent, hypothesis tests for a Poisson distribution appear in some Edexcel Further Statistics or advanced S2 questions. The Poisson distribution Po(λ) models the number of events in a fixed interval, where λ is the average rate. The test concerns the parameter λ.

虽然较少见，泊松分布的假设检验出现在某些 Edexcel 进阶统计或 S2 题目中。泊松分布 Po(λ) 描述了固定区间内事件发生的次数，其中 λ 是平均发生率。检验围绕参数 λ 展开。

To test H₀: λ = λ₀ against H₁: λ > λ₀, calculate P(X ≥ x) using the cumulative Poisson distribution with parameter λ₀. The p-value method is employed similarly to the binomial case. Edexcel mark schemes accept the use of either critical values or p-values, as long as the reasoning is clearly shown.

要检验 H₀: λ = λ₀ 对 H₁: λ > λ₀，使用参数为 λ₀ 的累积泊松分布计算 P(X ≥ x)。p 值法类似二项情形使用。Edexcel 评分方案接受临界值法或 p 值法，只要推理过程清晰展示即可。

9. Type I and Type II Errors | 第一类错误与第二类错误

A Type I error occurs when we reject a true null hypothesis. The probability of making a Type I error is exactly the significance level α. By choosing α = 0.05, we accept a 5% chance of incorrectly rejecting H₀. This error is controlled by the experimenter.

第一类错误发生在我们拒绝了真实的零假设时。犯第一类错误的概率就是显著性水平 α。选择 α = 0.05 意味着我们接受 5% 错误拒绝 H₀ 的风险。此错误由实验者控制。

A Type II error occurs when we fail to reject a false null hypothesis. The probability of a Type II error is denoted β, and its complement (1 – β) is the power of the test. Edexcel questions sometimes ask students to calculate P(Type II error) for a specific alternative value, usually involving the normal distribution.

第二类错误发生在我们未能拒绝一个错误的零假设时。第二类错误的概率记作 β，其补集 (1 – β) 是检验的功效。Edexcel 考题有时要求学生针对特定的备择值计算 P(第二类错误)，通常涉及正态分布。

Understanding these errors helps appreciate why we say ‘do not reject H₀’ rather than ‘accept H₀’, as the latter may lead to a Type II error. The balance between the two error types depends on sample size and effect size.

理解这些错误有助于领会为什么我们常说“不拒绝 H₀”而不是“接受 H₀”，因为后者可能导致第二类错误。两类错误之间的平衡取决于样本量和效应大小。

10. Critical Value Method vs p-Value Method | 临界值法与 p 值法

Two approaches are accepted for drawing conclusions. In the critical value method, the test statistic is compared with the critical value(s) that define the boundary of the critical region. If the test statistic falls into that region, H₀ is rejected.

得出检验结论可接受两种方法。在临界值法中，将检验统计量与界定临界域边界的临界值进行比较。如果检验统计量落入该区域，则拒绝 H₀。

In the p-value method, the probability of obtaining a result at least as extreme as the observed one is computed and compared directly with α. If p-value ≤ α, reject H₀. For two-tailed tests, the p-value is often doubled when using symmetrical distributions; with discrete distributions, care is needed in defining ‘more extreme’.

在 p 值法中，计算得到至少与观察结果同样极端的概率，并直接与 α 比较。若 p 值 ≤ α，拒绝 H₀。对于双尾检验，使用对称分布时 p 值常需翻倍；对于离散分布，界定“更极端”需要谨慎。

Edexcel mark schemes often show both methods as valid alternatives. Students should master at least one method and be able to present their reasoning step by step.

Edexcel 评分方案通常将两种方法都列为可接受的替代方案。学生应至少掌握一种方法，并能逐步展示推理过程。

11. Steps to Conduct a Hypothesis Test | 进行假设检验的步骤

A structured approach helps secure full marks. The recommended sequence is:

结构化的方法有助于确保获得满分。推荐的顺序如下：

State the null and alternative hypotheses clearly in symbols and words.

用符号和文字清晰陈述零假设与备择假设。
Specify the significance level α and define the distribution of the test statistic under H₀.

明确显著性水平 α，并定义 H₀ 下检验统计量的分布。
Calculate the test statistic from the sample data.

根据样本数据计算检验统计量。
Determine the critical region or compute the p-value.

确定临界域或计算 p 值。
Compare and make a decision: reject H₀ if the test statistic lies in the critical region, or if p-value ≤ α.

进行比较并作出决定：如果检验统计量落在临界域内，或 p 值 ≤ α，则拒绝 H₀。
Write a conclusion in context, referencing the evidence and the significance level.

在情境中写出结论，引用证据和显著性水平。

Following these steps not only organises your work but also demonstrates to the examiner that you understand the statistical process, which is crucial for the reasoning marks.

遵循这些步骤不仅能整理你的作答，还能向考官展示你理解了统计过程，这对于获得推理分数至关重要。

12. Common Mistakes and Exam Tips | 常见错误与考试技巧

Many candidates lose marks by failing to interpret the p-value correctly or by writing a conclusion that does not refer to the original claim. Always use the precise wording from the question. For instance, if the question asks about ‘a change’, a two-tailed test is implied.

许多考生因未能正确解读 p 值或写出的结论未提及原始论断而失分。务必使用题目中的准确措辞。例如，如果题目问及“是否有变化”，则暗示了双尾检验。

Another frequent error is using the wrong tail or forgetting to halve the significance level for a two-tailed normal test. When working with discrete distributions like binomial, remember that the critical region must satisfy the inequality exactly; using a value that gives a probability just above α is not acceptable.

另一个常见错误是使用了错误的尾部，或忘记在双尾正态检验中将显著性水平减半。处理离散分布（如二项分布）时，切记临界域必须恰好满足不等式；使用给出概率略高于 α 的值是不可接受的。

Lastly, always check whether the population variance is known before choosing between a z-test and a t-test (though t-tests are not in core Edexcel Maths, they may appear in further modules). For large samples, normal approximation is fine. Practise past papers to become familiar with the phrasing typical of Edexcel examiners.

最后，在选择 z 检验或 t 检验（尽管 t 检验不在 Edexcel 核心数学中，但可能出现于进阶模块）之前，务必检查总体方差是否已知。对于大样本，正态近似是可行的。多练习历年真题，熟悉 Edexcel 考官典型的问题措辞。

Published by TutorHao | Mathematics Revision Series | aleveler.com

更多咨询请联系16621398022（同微信）

Hypothesis Testing: Key Points for Edexcel Maths | 假设检验考点精讲