Hypothesis Testing – A-Level OCR Mathematics Key Points | A-Level OCR 数学：假设检验考点精讲

📚 Hypothesis Testing – A-Level OCR Mathematics Key Points | A-Level OCR 数学：假设检验考点精讲

Hypothesis testing is a fundamental tool in A-Level OCR Mathematics that allows you to make inferences about population parameters based on sample data. Whether you are testing a claimed proportion using the binomial distribution or assessing a mean with the normal distribution, understanding the structure of a hypothesis test is essential for success. This revision guide breaks down the key concepts, terminology, and steps required by the OCR specification, ensuring you can confidently tackle exam questions on hypothesis testing.

假设检验是 A-Level OCR 数学中的基本工具，它让你能够根据样本数据推断总体参数。无论你是用二项分布检验某个比例的声明，还是用正态分布评估均值，理解假设检验的结构对于考试成功至关重要。这份复习指南拆解了 OCR 考纲要求的关键概念、术语和步骤，确保你能自信地应对假设检验的考题。

1. Introduction to Hypothesis Testing | 假设检验简介

A hypothesis test is a statistical procedure used to determine whether there is enough evidence in a sample to reject a stated claim about a population parameter. In OCR A-Level, you will encounter tests for a binomial probability p and tests for the mean of a normal distribution. The process always begins by assuming a null hypothesis is true, and then checking if the observed data are too extreme to have occurred by random chance alone.

假设检验是一种统计程序，用于判断样本中是否有足够证据来拒绝关于总体参数的某个声明。在 OCR A-Level 中，你会遇到对二项概率 p 的检验和对正态分布均值的检验。整个过程始终从假设零假设为真开始，然后检查观测数据是否极端到无法仅由随机因素产生。

All hypothesis tests in this module are based on the concept of a test statistic, which summarises the sample information. For a binomial test the test statistic is the number of successes X; for a normal test it is the sample mean X̄ or the standardised Z-score. The decision rule is then applied using either critical values or p-values.

本模块中的所有假设检验都基于检验统计量的概念，它概括了样本信息。二项检验的检验统计量是成功次数 X；正态检验的检验统计量是样本均值 X̄ 或标准化的 Z 分数。然后通过临界值或 p 值来应用决策规则。

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

The null hypothesis, denoted H₀, is the statement being tested and is assumed true until evidence suggests otherwise. In OCR hypothesis tests, H₀ usually takes a precise value, such as H₀: p = 0.2 or H₀: μ = 50. The alternative hypothesis, denoted H₁, is what you conclude if the null is rejected. It can be one‑sided (p > 0.2, p < 0.2) or two‑sided (p ≠ 0.2). The wording of the claim in the question will determine which form to use.

零假设，记作 H₀，是被检验的陈述，在出现相反证据之前假定为真。在 OCR 的假设检验中，H₀ 通常取一个精确值，例如 H₀: p = 0.2 或 H₀: μ = 50。备择假设，记作 H₁，是当零假设被拒绝时你所得出的结论。它可以是单侧的（p > 0.2, p < 0.2）或双侧的（p ≠ 0.2）。题目中声明的措辞将决定采用哪种形式。

It is essential to state both hypotheses clearly in the context of the problem. For instance, if a manufacturer claims that at most 5% of light bulbs are defective, the null would be H₀: p = 0.05 (using the boundary value) and the alternative H₁: p > 0.05, because we only doubt the claim if the defect rate is higher.

必须在问题背景下清晰地表述两个假设。例如，如果一家制造商声称最多有 5% 的灯泡是次品，那么零假设将是 H₀: p = 0.05（使用边界值），备择假设 H₁: p > 0.05，因为只有在次品率更高时我们才会质疑该声明。

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

The significance level, denoted α (alpha), is the probability of rejecting H₀ when it is actually true – a Type I error. In A‑Level OCR exams, typical values are 1%, 5% or 10%, and the question will specify which to use. The critical (or rejection) region is the set of values of the test statistic for which H₀ is rejected. Its size is controlled by α.

显著性水平，记作 α（alpha），是在 H₀ 实际为真时拒绝 H₀ 的概率——即第一类错误。在 A‑Level OCR 考试中，典型取值为 1%、5% 或 10%，题目会指定使用哪一个。临界（或拒绝）区域是使得 H₀ 被拒绝的检验统计量值的集合。其大小由 α 控制。

For a right‑tailed binomial test at α = 0.05, we find the smallest integer c such that P(X ≥ c) ≤ 0.05, and the critical region is X ≥ c. For a two‑tailed test, α is split equally between the two tails, so we look for a lower critical value c₁ and an upper critical value c₂ such that P(X ≤ c₁) ≤ α/2 and P(X ≥ c₂) ≤ α/2.

对于一个显著性水平 α = 0.05 的右侧二项检验，我们要找到最小的整数 c 使得 P(X ≥ c) ≤ 0.05，此时临界区域为 X ≥ c。对于双侧检验，α 被平分到两个尾部，因此我们要寻找下临界值 c₁ 和上临界值 c₂，使得 P(X ≤ c₁) ≤ α/2 且 P(X ≥ c₂) ≤ α/2。

4. One-Tailed vs Two-Tailed Tests | 单尾检验与双尾检验

A one‑tailed test is used when the alternative hypothesis specifies a definite direction: either an increase (H₁: p > value, right‑tailed) or a decrease (H₁: p < value, left‑tailed). The entire significance level α is placed in one tail of the distribution. A two‑tailed test is used when the alternative hypothesis is simply 'not equal to' (H₁: p ≠ value), and α is split equally between both tails.

当备择假设指明了明确的方向时，使用单尾检验：要么增加（H₁: p > 某值，右侧检验），要么减少（H₁: p < 某值，左侧检验）。整个显著性水平 α 被放在分布的一个尾部。当备择假设只是“不等于”（H₁: p ≠ 某值）时，使用双尾检验，此时 α 被平分到两个尾部。

Choosing the correct tail is vital. If a question asks ‘test whether the proportion has increased’, a right‑tailed test is appropriate. If it asks ‘test whether the proportion has changed’, a two‑tailed test must be used. The OCR mark schemes penalise incorrect tail choices, so read the wording carefully.

选择正确的尾部至关重要。如果题目要求“检验比例是否增加”，则应使用右侧检验。如果要求“检验比例是否改变”，则必须使用双尾检验。OCR 的评分方案会惩罚错误的尾部选择，因此要仔细阅读措辞。

5. Test Statistic and the Binomial Distribution | 检验统计量与二项分布

When testing a population proportion, the test statistic is X, the number of successes in n independent trials. Under H₀: p = p₀, X follows a binomial distribution: X ~ B(n, p₀). The observed number of successes, x, is used to calculate probabilities. You will use your calculator’s binomial CD function, or cumulative tables, to find probabilities such as P(X ≥ x) or P(X ≤ x).

当检验一个总体比例时，检验统计量是 X，即 n 次独立试验中的成功次数。在 H₀: p = p₀ 下，X 服从二项分布：X ~ B(n, p₀)。观测到的成功次数 x 被用来计算概率。你将使用计算器的二项累积分布函数或累积表格来计算诸如 P(X ≥ x) 或 P(X ≤ x) 的概率。

For a right‑tailed test, you calculate P(X ≥ observed x). For a left‑tailed test, you calculate P(X ≤ observed x). For a two‑tailed test, you find whichever tail the observed value lies in, compute the tail probability, and then double it to compare with α. Alternatively, you can determine critical values and check whether the observed x falls into the rejection region.

对于右侧检验，计算 P(X ≥ 观测值 x)。对于左侧检验，计算 P(X ≤ 观测值 x)。对于双侧检验，先判断观测值落在哪个尾部，计算该尾部的概率，然后加倍后与 α 比较。或者，你可以先确定临界值，然后检查观测值 x 是否落入拒绝域。

6. Calculating p-values | 计算 p 值

The p‑value is the probability, under H₀, of obtaining a result at least as extreme as the one observed. In a one‑tailed test, p‑value = P(X ≥ x) for a right‑tailed test, and P(X ≤ x) for a left‑tailed test. In a two‑tailed binomial test, the p‑value is twice the probability of the observed tail: p‑value = 2 × min(P(X ≤ x), P(X ≥ x)). If p‑value ≤ α, H₀ is rejected.

p 值是在 H₀ 为真的前提下，得到至少与观测结果一样极端的结果的概率。在单尾检验中，右侧检验的 p 值 = P(X ≥ x)，左侧检验的 p 值 = P(X ≤ x)。在双侧二项检验中，p 值是观测尾部概率的两倍：p 值 = 2 × min(P(X ≤ x), P(X ≥ x))。如果 p 值 ≤ α，则拒绝 H₀。

OCR examiners expect you to compare the p‑value directly with the significance level. A common question asks ‘Does the test provide evidence at the 5% level?’ You answer by checking whether the calculated p‑value is less than 0.05. If it is, the evidence is significant; otherwise it is not.

OCR 考官期望你直接将 p 值与显著性水平进行比较。一个常见的问题是“该检验是否在 5% 的水平上提供了证据？”你通过检查计算出的 p 值是否小于 0.05 来回答。如果是，则证据显著；否则不显著。

7. Critical Values and Rejection Regions | 临界值与拒绝域

Instead of p‑values, you can use the critical region method. For a binomial test, find the critical value c such that the probability in the rejection region is as close to α as possible without exceeding it. For example, at α = 0.05 for B(20, 0.3), you might find P(X ≥ 10) = 0.0480 ≤ 0.05, so the critical region is X ≥ 10. If the observed x is 10 or more, reject H₀.

除了 p 值，你也可以使用临界区域法。对于二项检验，寻找临界值 c，使得拒绝域内的概率尽可能接近 α 且不超过它。例如，对于 α = 0.05 的 B(20, 0.3)，你可能发现 P(X ≥ 10) = 0.0480 ≤ 0.05，因此临界区域为 X ≥ 10。如果观测值 x 为 10 或更多，则拒绝 H₀。

The advantage of the critical region method is that it gives a clear decision rule. However, you must be careful with discrete distributions: the actual significance level (the size of the test) may be slightly smaller than the nominal α. In OCR exams you often need to state the actual significance level of the test.

临界区域法的优点是给出了清晰的决策规则。然而，你必须小心离散分布：实际显著性水平（检验的大小）可能略小于名义上的 α。在 OCR 考试中，你经常需要给出检验的实际显著性水平。

8. Steps for Hypothesis Testing | 假设检验的步骤

A structured approach is essential to gain full marks. Follow these steps for every hypothesis test:

有条理的方法对于获得满分至关重要。每一次假设检验都遵循以下步骤：

Step 1: Define the null and alternative hypotheses using correct notation and context. / 步骤 1： 用正确的符号和背景定义零假设与备择假设。
Step 2: State the significance level α. / 步骤 2： 声明显著性水平 α。
Step 3: Identify the distribution of the test statistic under H₀ (e.g. X ~ B(30, 0.4)). / 步骤 3： 确定检验统计量在 H₀ 下的分布（如 X ~ B(30, 0.4)）。
Step 4: Calculate the probability (p‑value) or determine the critical region. / 步骤 4： 计算概率（p 值）或确定临界区域。
Step 5: Compare the p‑value with α, or the test statistic with the critical value, and make a decision to either reject H₀ or not reject H₀. / 步骤 5： 比较 p 值与 α，或检验统计量与临界值，并做出拒绝 H₀ 或不拒绝 H₀ 的决定。
Step 6: Write a conclusion in context and in non‑technical language. Avoid saying ‘accept H₀’; instead say ‘there is insufficient evidence to reject H₀’ or ‘the evidence supports the alternative hypothesis’. / 步骤 6： 用非技术性语言写出基于背景的结论。避免说“接受 H₀”，而应该说“没有足够证据拒绝 H₀”或“证据支持备择假设”。

9. Normal Distribution Hypothesis Testing | 正态分布假设检验

When testing the population mean of a normally distributed variable with known variance σ², or when using a large sample (central limit theorem), we use a Z‑test. The test statistic is:

当检验已知方差 σ² 的正态分布总体的均值时，或使用大样本（中心极限定理）时，我们采用 Z 检验。检验统计量为：

Z = (X̄ − μ₀) / (σ/√n)

where X̄ is the sample mean, μ₀ is the hypothesised mean under H₀, σ is the population standard deviation, and n is the sample size. Under H₀, Z ~ N(0, 1). For a one‑tailed test, compare Z with the critical z‑value from tables; for a two‑tailed test at α = 0.05, the critical values are ±1.96.

其中 X̄ 为样本均值，μ₀ 为 H₀ 下的假设均值，σ 为总体标准差，n 为样本量。在 H₀ 下，Z ~ N(0, 1)。对于单尾检验，将 Z 与查表得到的临界 z 值比较；对于 α = 0.05 的双侧检验，临界值为 ±1.96。

For OCR, you must also calculate the p‑value from the normal distribution. If the alternative is H₁: μ > μ₀, p‑value = P(Z > z), and if H₁: μ ≠ μ₀, p‑value = 2 × P(Z > |z|). Use the standard normal table and your calculator correctly.

在 OCR 考试中，你还必须计算正态分布的 p 值。如果备择假设是 H₁: μ > μ₀，则 p 值 = P(Z > z)；如果 H₁: μ ≠ μ₀，则 p 值 = 2 × P(Z > |z|)。正确使用标准正态表与计算器。

10. Common Mistakes and Tips | 常见错误与提示

Many marks are lost through avoidable errors. Always state hypotheses in terms of the population parameter, not the sample statistic. For a binomial test, ensure the probability P(X ≥ x) is calculated correctly – some candidates mistakenly use P(X > x). Remember that for a discrete distribution, P(X ≥ x) = 1 − P(X ≤ x − 1).

许多分数因为一些可避免的错误而丢失。一定要用总体参数来陈述假设，而不是样本统计量。对于二项检验，确保正确计算概率 P(X ≥ x)——有些考生错误地使用了 P(X > x)。记住对于离散分布，P(X ≥ x) = 1 − P(X ≤ x − 1)。

In two‑tailed tests, always double the appropriate tail probability or use the symmetric critical region. Never compare a one‑tailed p‑value directly with α/2. Also, when writing a conclusion, refer back to the context: ‘There is significant evidence that the proportion of defective items exceeds 2%’ rather than just ‘Reject H₀’.

在双侧检验中，一定要将对应的尾部概率加倍，或者使用对称的临界区域。绝不要将单尾 p 值直接与 α/2 比较。此外，在撰写结论时，要联系题目背景：“有显著证据表明次品率超过 2%”，而不是仅仅写“拒绝 H₀”。

For normal tests, check whether the population variance is known; if only the sample standard deviation s is given and the sample is large, you may use it, but ideally the question specifies. Always state the distribution of the test statistic clearly.

对于正态检验，要检查总体方差是否已知；如果只给出了样本标准差 s 且样本量很大，你可以使用它，但理想情况下题目会指明。始终清楚地写明检验统计量的分布。

11. Worked Example: Binomial Hypothesis Test | 实例解析：二项分布检验

A company claims that 40% of customers prefer its new product. In a random sample of 25 customers, 15 say they prefer it. Test, at the 5% significance level, whether the true proportion is greater than 0.4.

一家公司声称 40% 的顾客偏好其新产品。在一个包含 25 名顾客的随机样本中，有 15 人表示偏好。在 5% 的显著性水平下，检验真实比例是否大于 0.4。

Step 1: Write hypotheses. H₀: p = 0.4; H₁: p > 0.4 (right‑tailed test). / 步骤 1： 写出假设。H₀: p = 0.4；H₁: p > 0.4（右侧检验）。

Step 2: Significance level α = 0.05. / 步骤 2： 显著性水平 α = 0.05。

Step 3: Under H₀, X ~ B(25, 0.4), where X is the number of customers who prefer the new product. Observed x = 15. / 步骤 3： 在 H₀ 下，X ~ B(25, 0.4)，X 是偏好新产品的顾客人数。观测值 x = 15。

Step 4: Since it is right‑tailed, find the p‑value: P(X ≥ 15) = 1 − P(X ≤ 14). Using calculator, P(X ≤ 14) ≈ 0.9569, so p‑value ≈ 1 − 0.9569 = 0.0431. / 步骤 4： 因为是右侧检验，求 p 值：P(X ≥ 15) = 1 − P(X ≤ 14)。使用计算器，P(X ≤ 14) ≈ 0.9569，所以 p 值 ≈ 1 − 0.9569 = 0.0431。

Step 5: Compare p‑value (0.0431) with α (0.05). Since 0.0431 < 0.05, we reject H₀. / 步骤 5： 比较 p 值（0.0431）与 α（0.05）。因为 0.0431 < 0.05，我们拒绝 H₀。

Step 6: Conclusion: There is sufficient evidence at the 5% level to suggest that the true proportion of customers who prefer the new product is greater than 0.4. / 步骤 6： 结论：在 5% 的显著性水平上有足够证据表明，真正偏好新产品的顾客比例大于 0.4。

Alternatively, using the critical region method: find c such that P(X ≥ c) ≤ 0.05. P(X ≥ 14) = 1 − P(X ≤ 13) ≈ 0.060; P(X ≥ 15) = 0.0431 ≤ 0.05. So the critical region is X ≥ 15. Since the observed value 15 lies in this region, we again reject H₀.

或者，使用临界区域法：寻找 c 使得 P(X ≥ c) ≤ 0.05。P(X ≥ 14) = 1 − P(X ≤ 13) ≈ 0.060；P(X ≥ 15) = 0.0431 ≤ 0.05。因此临界区域为 X ≥ 15。由于观测值 15 落在此区域内，我们同样拒绝 H₀。

Published by TutorHao | Mathematics Revision Series | aleverer.com

更多咨询请联系16621398022（同微信）

Hypothesis Testing – A-Level OCR Mathematics Key Points | A-Level OCR 数学：假设检验 考点精讲