Hypothesis Testing for GCSE Mathematics | GCSE 数学：假设检验 考点精讲

📚 Hypothesis Testing for GCSE Mathematics | GCSE 数学：假设检验 考点精讲

Hypothesis testing is a fundamental tool in statistics that allows you to use sample data to decide whether a claim about a population parameter is likely to be true. In GCSE Mathematics, you will learn to perform hypothesis tests for a population proportion using the binomial distribution, applying concepts such as null and alternative hypotheses, significance levels, and critical regions. This article provides a focused revision guide to help you master every key aspect of hypothesis testing and tackle exam questions with confidence.

假设检验是统计学中的基本工具，它让你能够利用样本数据判断关于总体参数的某个声明是否可能成立。在 GCSE 数学中，你将学习使用二项分布对总体比例进行假设检验，并运用零假设、备择假设、显著性水平和临界区域等概念。本文提供了一份重点复习指南，帮助你掌握假设检验的各个关键环节，自信地应对考试题目。

1. Understanding the Concept of a Hypothesis Test | 理解假设检验的概念

In statistics, a hypothesis test is a formal procedure for deciding whether there is enough evidence to reject a particular belief about a population parameter, such as a proportion. The starting point is always an assumption, which we then test against observed data.

在统计学中，假设检验是一个正式程序，用于判断是否有足够证据拒绝关于总体参数（例如比例）的某种特定看法。出发点始终是一个假设，然后我们用观测到的数据来检验它。

We never prove a hypothesis to be true; instead, we look for evidence against it. If the observed data are highly unlikely under the assumed model, we have reason to doubt the original assumption.

我们永远无法证明一个假设为真；相反，我们要寻找反对它的证据。如果观测数据在假设模型下非常不可能出现，我们就有理由怀疑最初的假定。

Think of hypothesis testing like a courtroom trial: the defendant is assumed innocent until the evidence convinces the jury to find them guilty. Similarly, we start with a null hypothesis and only reject it if the sample provides strong enough evidence.

可以把假设检验想象成法庭审判：被告在被证据说服陪审团认定有罪之前，都被假定为无罪。同样，我们从零假设出发，只有当样本提供了足够有力的证据时，才拒绝它。

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

Every hypothesis test involves two competing statements: the null hypothesis, denoted H₀, and the alternative hypothesis, denoted H₁. H₀ is the statement of no effect or no change, while H₁ represents the claim we wish to test for evidence of.

每个假设检验都涉及两个互相对立的陈述：零假设，记为 H₀，以及备择假设，记为 H₁。H₀ 表示没有效果或没有变化，而 H₁ 代表我们希望寻找证据支持的声明。

alternative hypothesis can be one of three forms, depending on the wording of the problem.

对于二项比例 p 的检验，H₀ 总是写成 p = 某个值的形式，例如 H₀: p = 0.3。备择假设根据题目表述可以取三种形式之一。

The three possible forms of H₁ are: p > value (right‑tailed), p < value (left‑tailed), or p ≠ value (two‑tailed). Deciding which one to use requires careful reading of phrases like 'more than', 'less than', 'changed', or 'different from'.

H₁ 的三种可能形式是：p > 值（右侧检验），p < 值（左侧检验），或者 p ≠ 值（双侧检验）。选择哪一种需要仔细阅读题目中的措辞，比如“大于”、“小于”、“发生了变化”或“不同于”。

For instance, if a company claims that 30% suspect the true proportion is actually higher, you would set H₀: p = 0.3 and H₁: p > 0.3.

例如，如果一家公司声称有30%的顾客偏爱一种新饮料，而你怀疑真实比例实际上更高，就应该设定 H₀: p = 0.3 和 H₁: p > 0.3。

3. One‑Tailed and Two‑Tailed Tests | 单尾检验与双尾检验

When the alternative hypothesis specifies a direction (greater than or less than), we conduct a one‑tailed test. The critical region, where we would reject H₀, lies entirely in one tail of the binomial distribution.

当备择假设指明了方向（大于或小于）时，我们进行单尾检验。拒绝 H₀ 的临界区域完全位于二项分布的一个尾部。

If H₁ takes the form p ≠ value, the test is two‑tailed. Here the evidence against H₀ can come from either extreme of the distribution, so the significance level is split equally between the two tails.

如果 H₁ 的形式为 p ≠ 值，检验就是双尾的。此时，反对 H₀ 的证据可以来自分布的两个极端，因此显著性水平被等分到两个尾部。

The decision between one‑tailed and two‑tailed tests depends on the question's wording. Look for keywords: 'increase', 'decrease', 'better', or 'worse' suggest one‑tailed; 'different', 'changed', 'not equal' point to a two‑tailed test.

单尾还是双尾取决于题目的措辞。注意关键词：“增加”、“减少”、“更好”或“更差”暗示单尾；“不同”、“改变”、“不等于”则指向双尾检验。

4. Significance Level and Critical Values | 显著性水平与临界值

The significance level, denoted by α, is the maximum probability of making a Type I error — that is, rejecting a true null hypothesis. In GCSE problems, α is usually given as 5% (0.05) or sometimes 10% (0.10) or 1% (0.01).

显著性水平，记作 α，是犯第一类错误的最大概率——即拒绝了一个真实的零假设。在 GCSE 题目中，α 通常给出为 5% (0.05)，有时是 10% (0.10) 或 1% (0.01)。

Critical values split the range of the test statistic into the acceptance region and the critical (rejection) region. For a binomial test, critical values are specific numbers of successes; if the observed number falls in the critical region, we reject H₀.

临界值将检验统计量的范围划分为接受区域和临界（拒绝）区域。对于二项检验，临界值是具体的成功次数；如果观测到的次数落在临界区域内，我们就拒绝 H₀。

For example, in a one‑ es is 12 or more. The probability of getting 12 or more successes under H₀ must be ≤ 0.05.

例如，在一个 H₀: p = 0.4、n = 20、α = 0.05 的单尾检验中，假设临界值是 12。这意味着如果成功次数达到 12 或以上，我们就拒绝 H₀。在 H₀ 成立的条件下，得到 12 次或更多成功的概率必须 ≤ 0.05。

It is essential to remember that the critical region is always chosen so that the probability of falling in it, when H₀ is true, is as close as possible to α without exceeding it. This can be found using cumulative binomial tables or a calculator.

必须记住，临界区域的选择总是使得当 H₀ 成立时落入该区域的概率尽可能接近 α 但不超过 α。这可以通过累积二项分布表或计算器求得。

5. The Binomial Distribution in Hypothesis Testing | 假设检验中的二项分布

In GCSE hypothesis tests, the test statistic is the number of successes in a fixed number of independent trials, so it follows a binomial distribution: X ~ B(n, p). Under H₀, p is set to the claimed proportion.

在 GCSE 的假设检验中，检验统计量是固定次数独立试验中的成功次数，因此它服从二项分布：X ~ B(n, p)。在 H₀ 下，p 被设定为声称的比例。

We use the binomial distribution to calculate the probability of obtaining a result at least as extreme as the one observed, assuming H₀ is true. This probability is called the p‑value, though GCSE often focuses on comparing the observed count with a critical value.

我们利用二项分布，在假设 H₀ 成立的条件下，计算得到一个至少与实际观测结果同样极端的结果的概率。这个概率称为 p 值，不过 GCSE 通常更侧重于将观测计数与临界值进行比较。

When calculating probabilities, always check whether the test is one‑tailed or two‑tailed. For a one‑ , you calculate P(X ≥ observed value). For H₁: p < 0.3, it would be P(X ≤ observed value).

计算概率时，务必确认检验是单尾还是双尾。对于 H₁: p > 0.3 这样的单尾检验，需要计算 P(X ≥ 观测值)。对于 H₁: p < 0.3，则计算 P(X ≤ 观测值)。

For a two‑tailed test, you consider extremes in both directions. Usually you compare the observed count against two critical values, or double the smaller tail probability and compare it with α.

对于双尾检验，你需要考虑两个方向上的极端情况。通常是将观测计数与两个临界值比较，或者将较小的尾部概率加倍后与 α 比较。

6. Steps for a Hypothesis Test | 假设检验的步骤

A clear, structured approach is the best way to earn full marks in an exam. Always follow these six steps when performing a binomial hypothesis test.

清晰、有条理的方法是在考试中获得满分的最佳途径。进行二项假设检验时，请始终遵循以下六个步骤。

g correct notation, e.g. H₀: p = 0.5, H₁: p < 0.5. Step 2: State the significance level α.

第 1 步：用正确的记号定义零假设和备择假设，例如 H₀: p = 0.5, H₁: p < 0.5。第 2 步：写明显著性水平 α。

Step 3: Identify the distribution of the test statistic under H₀: X ~ B(n, p). Step 4: Determine the critical region or calculate the p‑value for the observed result.

第 3 步：确定在 H₀ 下检验统计量的分布：X ~ B(n, p)。第 4 步：确定临界区域或计算观测结果的 p 值。

Step 5: Compare the observed value with the critical region or compare the p‑value with α. Step 6: Write a conclusion in the context of the question, stating clearly whether there is sufficient evidence to reject H₀ and what that means in plain English.

第 5 步：将观测值与临界区域进行比较，或将 p 值与 α 进行比较。第 6 步：结合题目背景写下结论，明确说明是否有足够证据拒绝 H₀，并用通俗语言解释其含义。

For example, if the observed number falls inside the critical region, you might write: "There is sufficient evidence at the 5% level to reject the null hypothesis. This suggests that the true proportion of customers who prefer the new drink is indeed greater than 0.3."

例如，如果观测数量落在临界区域内，可以这样写：“在 5% 显著性水平下有足够证据拒绝零假设。这表明偏爱新饮料的顾客真实比例确实大于 0.3。”

7. Finding Critical Regions | 求临界区域

To find the critical region for a one‑ t possible number of successes and work downwards until the cumulative probability is ≤ α.

要找到 H₁: p > 声称值的单尾检验临界区域，从最大可能的成功次数开始，向下累加概率，直到累积概率 ≤ α。

广告 / Sponsored

gion is X ≥ 10 because it is the smallest set that gives a probability not exceeding 0.05.

例如，假设 n = 15，H₀ 下 p = 0.4，α = 0.05。计算 P(X ≥ 10) = 1 − P(X ≤ 9)。如果 P(X ≥ 10) ≈ 0.033 而 P(X ≥ 9) ≈ 0.095，那么临界区域就是 X ≥ 10，因为这是概率不超过 0.05 的最小集合。

For a two‑ er critical value c₂ such that P(X ≥ c₂) ≤ 0.025. The critical region is X ≤ c₁ or X ≥ c₂.

对于 α = 0.05 的双尾检验，每个尾部各得 0.025。找到下临界值 c₁ 使得 P(X ≤ c₁) ≤ 0.025，以及上临界值 c₂ 使得 P(X ≥ c₂) ≤ 0.025。临界区域就是 X ≤ c₁ 或 X ≥ c₂。

When using tables, you often need to look up cumulative binomial probabilities and choose the nearest values without exceeding the tail probability. Practise this with a variety of n and p values until it becomes second nature.

使用数表时，通常需要查找累积二项概率，并选择不超过尾部概率的最近数值。用不同的 n 和 p 值多加练习，直到运用自如。

8. Interpreting p‑values | 解读 p 值

Although GCSE exam boards often emphasise the critical region method, understanding the p‑value deepens your grasp of hypothesis testing. The p‑value is the probability, under H₀, of obtaining a result at least as extreme as the one observed.

尽管 GCSE 考试局通常强调临界区域法，理解 p 值可以加深你对假设检验的掌握。p 值是在 H₀ 下，得到一个至少与实际观测结果同样极端的概率。

If the p‑value is less than or equal to α, we reject H₀. For a one‑tailed test with H₁: p > 0.5, and an observed count of 14 out of 20, the p‑value is P(X ≥ 14). If this probability is 0.03 and α = 0.05, we reject H₀.

如果 p 值小于或等于 α，我们就拒绝 H₀。对于 H₁: p > 0.5 的单尾检验，如果在 20 次试验中观察到 14 次成功，p 值就是 P(X ≥ 14)。如果这个概率是 0.03 而 α = 0.05，我们就拒绝 H₀。

In a two‑tailed test, the p‑ lue, calculate P(X ≥ observed) and double it. Then compare this doubled probability with α.

在双侧检验中，p 值通常取观测值所在尾部的概率的两倍。例如，如果观测值高于期望值，计算 P(X ≥ 观测值) 并将其加倍。然后将加倍后的概率与 α 比较。

Be careful: some exam specifications may handle two‑tailed tests slightly differently, but the underlying logic remains the same — the smaller the p‑value, the stronger the evidence against H₀.

注意：某些考试大纲可能对双尾检验的处理略有不同，但底层逻辑是一样的——p 值越小，拒绝 H₀ 的证据就越强。

9. Common Mistakes to Avoid | 常见错误及避免方法

hypotheses about the sample result rather than the population parameter. Never write H₀: p̂ = 0.3; always use the population proportion p.

一个常见错误是围绕样本结果而不是总体参数来写假设。千万不要写 H₀: p̂ = 0.3；一定要用总体比例 p。

Another mistake is misidentifying the direction of the test. Carefully underline words like 'increase', 'decrease', or 'change' in the question to guide your choice of H₁.

另一个错误是误判检验的方向。仔细圈出题目中的“增加”、“减少”或“改变”等词语，以此指导你对 H₁ 的选择。

Many students forget to state the conclusion in context. Always relate your decision back to the original problem: "We reject the null hypothesis. There is evidence that the new drug reduces the side‑effect rate compared to the standard 10%."

许多学生忘记结合背景陈述结论。务必把你的决定与最初的问题联系起来：“我们拒绝零假设。有证据表明新药将副作用发生率相较于标准的 10% 降低了。”

Avoid using the phrase 'accept H₀'. In hypothesis testing, we only ever 'do not reject H₀' because lack of evidence against it does not prove it true. Exam mark schemes are strict on this wording.

避免使用“接受 H₀”这种说法。在假设检验中，我们只说“不能拒绝 H₀”，因为缺乏反对证据并不能证明它为真。考试评分方案对这个措辞要求很严格。

When calculating binomial probabilities, do not round prematurely. Keep intermediate values to at least four decimal places, and only round the final probability or critical value at the end.

计算二项概率时，不要过早四舍五入。中间值至少保留四位小数，只在最后才对最终的概率或临界值进行舍入。

10. Worked Example Step by Step | 逐步示例解析

A school claims that 70% of its students walk to school. A student representative believes the true proportion is lower. He surveys 30 students and finds 16 walk to school. Test at the 5% significance level whether the proportion is actually less than 0.7.

一所学校声称 70% 的学生步行上学。一名学生代表认为真实比例更低。他调查了 30 名学生，发现 16 人步行上学。在 5% 显著性水平下检验比例是否真的小于 0.7。

Step 1: H₀: p = 0.7, H₁: p < 0.7. Step 2: α = 0.05. Step 3: Under H₀, X ~ B(30, 0.7), where X is the number walking to school.

第 1 步：H₀: p = 0.7, H₁: p < 0.7。第 2 步：α = 0.05。第 3 步：在 H₀ 下，X ~ B(30, 0.7)，其中 X 表示步行上学的人数。

ng B(30,0.7), P(X ≤ 16) ≈ 0.0446 and P(X ≤ 17) ≈ 0.1081. So the critical region is X ≤ 16.

第 4 步：求临界区域。对于 H₁: p < 0.7，我们需要找到最大的 c 使得 P(X ≤ c) ≤ 0.05。利用 B(30,0.7)，P(X ≤ 16) ≈ 0.0446，P(X ≤ 17) ≈ 0.1081。因此临界区域是 X ≤ 16。

ritical region. Step 6: Since 16 ≤ 16, we reject H₀. There is sufficient evidence at the 5% level to suggest that the true proportion of students walking to school is less than 70%.

第 5 步：观测值 16 落在临界区域内。第 6 步：由于 16 ≤ 16，我们拒绝 H₀。在 5% 水平下有足够证据表明步行上学的真实学生比例小于 70%。

11. Using Tables and Technology Efficiently | 高效使用数表与工具

n locate the probability column under p = H₀ value, and read off cumulative probabilities.

考试中，可能会提供累积二项分布表。学会快速查阅：找到对应的 n，然后在 p = H₀ 值的概率列中定位，并读出累积概率。

ical values, look for the largest X where the cumulative probability is ≤ 0.05 (for a left‑iled test) or the smallest X where the complementary probability exceeds 0.95 (for a right‑tailed test).

对于求临界值，寻找使得累积概率 ≤ 0.05 的最大 X（左侧检验），或者使得互补概率超过 0.95 的最小 X（右侧检验）。

ators have functions like Binomial CD to compute cumulative probabilities directly. Double‑check whether your calculator expects P(X ≤ x) or P(X < x), and adjust accordingly.

许多计算器都有 Binomial CD 等函数来直接计算累积概率。务必核实你的计算器是计算 P(X ≤ x) 还是 P(X < x)，并做相应调整。

When using technology, always write down the distribution, the probability you are calculating, and the result, so the examiner can follow your working even if the final answer is slightly off.

使用工具时，一定要写下分布、你正在计算的概率以及结果，这样即使最终答案略有偏差，阅卷老师也能看懂你的解题过程。

12. Linking Hypothesis Testing to Real‑World Contexts | 将假设检验联系实际情境

Hypothesis testing is not just an exam technique; it is a key part of the scientific method. Medical trials, quality control in manufacturing, and opinion polls all rely on the same logic you are learning now.

假设检验不只是一项考试技巧，更是科学方法的关键组成部分。医学试验、制造业的质量控制和民意调查都依赖你正在学习的这一逻辑。

Whenever you see a claim in newspapers or advertisements – for example, '8 out of 10 cats prefer Whiskers' – hypothesis: H₀: p = 0.8, and ask what sample size and evidence would be needed to test it.

每当你看到报纸或广告中的声明时——例如，“10 只猫中有 8 只偏爱伟嘉”——你都可以用零假设的角度来思考：H₀: p = 0.8，并问自己需要多大的样本量和怎样的证据才能检验它。

Understanding hypothesis testing empowers you to critically evaluate data‑based arguments. It teaches you to ask: 'Could this result have occurred by chance alone?' – a question at the heart of all statistical inference.

理解假设检验让你有能力批判性地评价基于数据的论证。它教会你提问：“这个结果是否有可能只是偶然而已？”——这正是所有统计推断的核心问题。

As you revise, try to connect each worked example to a real‑world scenario. This will not only make the mathematics more memorable but also sharpen your ability to set up hypotheses correctly in unfamiliar contexts.

复习时，尝试将每个例题与实际场景联系起来。这不仅会让数学知识更难忘，还能让你在不熟悉的情境中更准确地设立假设。

---

Published by TutorHao | Mathematics Revision Series | aleveler.com

更多咨询请联系16621398022（同微信）