A-Level Edexcel Mathematics: Hypothesis Testing | A-Level Edexcel 数学：假设检验考点精讲

📚 A-Level Edexcel Mathematics: Hypothesis Testing | A-Level Edexcel 数学：假设检验考点精讲

Hypothesis testing is a core statistical method used to make decisions or inferences about population parameters based on sample data. In A-Level Edexcel Mathematics, you are expected to formulate null and alternative hypotheses, conduct tests using binomial, Poisson, and normal distributions, interpret p-values and critical regions, and draw valid conclusions in context. This article breaks down every key concept, process step, and common pitfall to help you master hypothesis testing with confidence.

假设检验是一种核心统计方法，用于根据样本数据对总体参数做出决策或推断。在 A-Level Edexcel 数学课程中，你需要学会建立零假设与备择假设，运用二项分布、泊松分布和正态分布进行检验，解读 p 值与临界区域，并结合背景得出有效结论。本文拆解所有关键概念、操作步骤和常见误区，帮助你扎实掌握假设检验。

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

Hypothesis testing is a formal procedure for checking whether a claim about a population parameter (e.g. proportion, mean, rate) is supported by sample evidence. It starts from a starting assumption, collects data, and then measures how extreme the observed result is under that assumption.

假设检验是一套规范流程，用来检测关于总体参数（如比例、均值、发生率）的声明是否得到样本证据支持。它从一个初始假设出发，收集数据，然后衡量在该假设下观测到的结果有多极端。

In Edexcel exams, you will be given a real-world scenario, a probability distribution, and a significance level. Your job is to follow a structured approach: define hypotheses, calculate probabilities or test statistics, compare with critical values or p-values, and write a conclusion that refers back to the original claim.

在 Edexcel 考试中，你会得到一个真实场景、一个概率分布和一个显著性水平。你的任务是按结构化的步骤：定义假设、计算概率或检验统计量、与临界值或 p 值比较，然后写出一条回扣原始声明的结论。

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

The null hypothesis, denoted H₀, is the default position that assumes no effect, no difference, or that the population parameter equals a stated value. The alternative hypothesis, H₁, represents what the test is designed to detect – a change, an increase, a decrease, or simply a difference.

零假设，记作 H₀，是默认立场，假定没有效应、没有差异，或总体参数等于某个指定值。备择假设 H₁ 则代表检验想要发现的内容——变化、增加、减少或仅仅是有差异。

For example, if a coin is claimed to be fair, H₀: p = 0.5. If you suspect it is biased towards heads, then H₁: p > 0.5 (one-tailed). If you only suspect unfairness without specifying direction, H₁: p ≠ 0.5 (two-tailed).

例如，声称一枚硬币是公平的，则 H₀: p = 0.5。如果你怀疑它倾向于正面，则 H₁: p > 0.5（单尾检验）。如果你只怀疑不公平但不指定方向，则 H₁: p ≠ 0.5（双尾检验）。

Always state hypotheses in terms of population parameters, never sample statistics. Use words and symbols together for clarity in exam answers.

始终用总体参数表述假设，绝不用样本统计量。在考试答案中同时使用文字和符号，以求清晰。

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

A one-tailed test is used when the alternative hypothesis specifies a direction of change, e.g., H₁: p > 0.7 or H₁: μ < 50. The critical region lies entirely in one tail of the distribution.

当备择假设指明了变化方向时，使用单尾检验，如 H₁: p > 0.7 或 H₁: μ < 50。临界区域完全位于分布的一个尾部。

A two-tailed test is used when H₁ states only that the parameter is different from the null value, e.g., H₁: p ≠ 0.7. In this case, the significance level is split equally between both tails.

当 H₁ 仅说明参数与零假设值不同时，使用双尾检验，如 H₁: p ≠ 0.7。这种情况下，显著性水平被平均分配到两个尾部。

Choosing the correct test type is crucial. Read the wording carefully: ‘changed’, ‘different’, or ‘not equal to’ implies two-tailed; ‘greater than’ or ‘less than’ implies one-tailed. Edexcel markschemes penalise using the wrong tail.

选择正确的检验类型至关重要。仔细审题措辞：“变化”、“不同”或“不等于”暗示双尾；“大于”或“小于”暗示单尾。Edexcel 评分标准对尾部选错会扣分。

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

The significance level, α, is the probability of rejecting H₀ when it is actually true. Common values are 5%, 1%, or 10%. In Edexcel, you are usually told the significance level.

显著性水平 α 是当 H₀ 为真时却拒绝它的概率。常用值有 5%、1% 或 10%。Edexcel 题目通常会告知显著性水平。

The critical region (or rejection region) consists of all values of the test statistic that lead to rejection of H₀. The boundary value(s) that separate the critical region from the acceptance region are called critical values.

临界区域（或拒绝域）包含所有导致拒绝 H₀ 的检验统计量的取值。分隔临界区域与接受域的边界值称为临界值。

For discrete distributions like binomial and Poisson, the actual significance level – the exact probability of being in the critical region – is often slightly less than the nominal α because of discreteness. Edexcel expects you to find the critical region such that the tail probability ≤ α.

对于二项分布和泊松分布等离散型分布，实际显著性水平——即落入临界区域的精确概率——由于离散性通常略小于名义 α。Edexcel 要求你找到使尾部概率 ≤ α 的临界区域。

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

The test statistic is a summary calculated from the sample that measures evidence against H₀. For binomial tests, the test statistic is the observed number of successes X. For normal distribution tests, it is often the sample mean x̄, which is then standardised to a Z-score.

检验统计量是从样本计算出的一个总结量，衡量反对 H₀ 的证据。对于二项检验，检验统计量是观测到的成功次数 X。对于正态分布检验，通常是样本均值 x̄，再将其标准化为 Z 分数。

The p-value is the probability, if H₀ is true, of obtaining a test statistic at least as extreme as the observed value. For a one-tailed test (upper tail), p-value = P(X ≥ observed | H₀). If p-value ≤ α, reject H₀. The smaller the p-value, the stronger the evidence against H₀.

p 值是在 H₀ 为真的前提下，获得至少与观测值同样极端的检验统计量的概率。对于上尾单尾检验，p 值 = P(X ≥ 观测值 | H₀)。若 p 值 ≤ α，则拒绝 H₀。p 值越小，反对 H₀ 的证据越强。

Edexcel accepts either the critical-region method or the p-value method. Both lead to the same conclusion, but you must clearly show your working.

Edexcel 接受临界区域法或 p 值法，两者得出相同结论，但你必须清晰展示步骤。

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

When testing a population proportion p using a binomial distribution X ~ B(n, p₀), you calculate tail probabilities directly from the binomial formula or use cumulative probability tables.

当使用二项分布 X ~ B(n, p₀) 检验总体比例 p 时，你直接利用二项公式计算尾部概率，或使用累积概率表。

Example: It is claimed that 30% of seeds germinate. A new treatment is tested on 20 seeds; 10 germinate. Test at 5% significance whether the germination rate has increased.

示例：据称种子发芽率为 30%。对 20 粒种子施用新处理，有 10 粒发芽。在 5% 显著性水平下检验发芽率是否提高。

H₀: p = 0.3, H₁: p > 0.3 (one-tailed)
Test statistic X = 10, under H₀ X ~ B(20, 0.3)
p-value = P(X ≥ 10) = 1 – P(X ≤ 9) = 1 – 0.9520 = 0.0480
Since 0.0480 < 0.05, reject H₀. There is sufficient evidence to suggest the germination rate has increased.

Edexcel提醒：二项分布计算中必须精确使用正确的累积概率，注意尾部的定义，以及实际显著性水平的计算。

7. Hypothesis Testing for Poisson Distribution | 泊松分布假设检验

Poisson hypothesis tests work similarly, using the model X ~ Po(λ₀) under H₀. The test statistic is the observed count X. The p-value or critical region is found using Poisson cumulative probabilities.

泊松假设检验操作类似，在 H₀ 下使用模型 X ~ Po(λ₀)。检验统计量是观测计数 X。p 值或临界区域通过泊松累积概率求得。

Example: A call centre claims calls arrive at a mean rate of 4 per 10 minutes. In a random 10-minute interval, 8 calls are received. Test at 5% significance whether the rate has increased.

示例：某呼叫中心声称每 10 分钟平均呼入 4 通电话。在随机 10 分钟内，收到 8 通。在 5% 显著性水平下检验呼入率是否上升。

H₀: λ = 4, H₁: λ > 4
X ~ Po(4), p-value = P(X ≥ 8) = 1 – P(X ≤ 7) = 1 – 0.9489 = 0.0511
0.0511 > 0.05, so do not reject H₀. Insufficient evidence of increase.

注意：对于离散分布，临界区域的实际概率往往不等于名义 α，Edexcel 评分时要求找出≤α的最大临界区域。

8. Hypothesis Testing for Normal Distribution (Mean) | 正态分布均值检验

When the population is normally distributed, or the sample size is large (Central Limit Theorem applies), we test the population mean μ using the sample mean x̄. If σ² is known, the test statistic is:

当总体服从正态分布，或样本容量较大（中心极限定理适用）时，我们利用样本均值 x̄ 检验总体均值 μ。若 σ² 已知，检验统计量为：

Z = (x̄ – μ₀) / (σ / √n) ~ N(0, 1)

Compare Z with critical Z-values from the standard normal distribution. For a 5% two-tailed test, critical values are ±1.96.

将 Z 与标准正态分布的临界 Z 值比较。对于 5% 双尾检验，临界值为 ±1.96。

If σ² is unknown and the sample size is small, you would use the t-distribution, but Edexcel A-Level Mathematics normally focuses on known variance cases or uses the normal approximation. Always check context.

若 σ² 未知且样本量较小，须使用 t 分布，但 Edexcel A-Level 数学通常聚焦于方差已知的情形或使用正态近似。务必结合题目背景判断。

Example: A machine fills packets with a nominal mean 500 g and σ = 4 g. A sample of 25 packets has mean 498.5 g. Test at 1% significance if the mean has changed.

示例：某机器装袋，标称均值 500 g，σ = 4 g。抽取 25 袋样本的均值为 498.5 g。在 1% 显著性水平下检验均值是否改变。

H₀: μ = 500, H₁: μ ≠ 500 (two-tailed)
Z = (498.5 – 500) / (4/5) = –1.875
Critical values: ±2.576 (1% two-tailed). |Z| = 1.875 < 2.576, do not reject H₀.

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

A Type I error occurs when H₀ is true but is rejected. The probability of a Type I error equals the significance level α. A Type II error occurs when H₀ is false but we fail to reject it. The probability of a Type II error is denoted by β and depends on the true value of the parameter.

第一类错误发生在 H₀ 为真却被拒绝时。其概率等于显著性水平 α。第二类错误发生在 H₀ 为假但未被拒绝时，概率记为 β，并依赖于参数的真值。

Edexcel may ask you to calculate the actual probability of a Type I error from the critical region for discrete distributions (which is the size of the test) or to state which error is applicable in context. For example, if a new drug is incorrectly approved when it does not work, that is a Type I error; if an effective drug is not approved, it is a Type II error.

Edexcel 可能要求你根据离散分布的临界区域计算实际的第一类错误概率（即检验的规模），或结合背景说明适用哪种错误。例如，如果一种无效的新药被错误批准，这是第一类错误；如果有效药物未获批准，则是第二类错误。

10. Writing Conclusions in Context | 结合背景撰写结论

Your final conclusion must be stated in the context of the original problem, not just in statistical terms. Use phrases like “there is sufficient evidence at the 5% level to suggest that… ” or “there is insufficient evidence to reject the claim that…”. Always mention the significance level and refer back to the population parameter and direction.

最终结论必须结合原问题背景表述，而不仅仅是统计术语。应使用诸如“在 5% 显著性水平下，有充分证据表明……”或“没有充分证据拒绝……这一声称”的措辞。务必提及显著性水平，并回扣总体参数及变化方向。

Avoid saying “accept H₀”. Instead, say “do not reject H₀”. Edexcel examiners are strict on this because hypothesis testing can only accumulate evidence against H₀, never prove it true.

应避免说“接受 H₀”，而应说“不拒绝 H₀”。Edexcel 考官对此要求严格，因为假设检验只能积累反对 H₀ 的证据，永远无法证明 H₀ 为真。

Also check that your conclusion matches the p-value or critical region decision. Contradictory statements will lose marks.

还要确保结论与 p 值或临界区域判断一致，自相矛盾的陈述将失分。

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

Confusing one-tailed and two-tailed tests is the single most frequent error. Always underline directional words in the question. Another common mistake is misreading cumulative probability tables – be sure you know whether your table gives P(X ≤ x) or P(X < x).

混淆单尾与双尾是最常见的错误。务必在题目中划出方向性词语。另一个常见错误是读错累积概率表——要确认表格提供的是 P(X ≤ x) 还是 P(X < x)。

When finding critical regions for discrete distributions, remember that the probability of being in the critical region must be ≤ α. If P(X ≤ c) < α but P(X ≤ c+1) > α, the lower critical value is c. For upper tail, choose c such that P(X ≥ c) ≤ α but P(X ≥ c-1) > α. Provide both the critical region and its actual significance level for full marks.

为离散分布寻找临界区域时，记住落入临界区域的概率必须 ≤ α。如果 P(X ≤ c) < α 而 P(X ≤ c+1) > α，则下侧临界值为 c。上尾则选择 c 使得 P(X ≥ c) ≤ α 但 P(X ≥ c-1) > α。给出临界区域及其实际显著性水平才能得满分。

In normal hypothesis testing, always sketch the distribution and mark critical values. This reduces sign errors. When standardising, double-check the denominator: it is the standard error σ/√n, not the population standard deviation alone.

在正态假设检验中，始终画出分布图并标记临界值，能减少符号错误。标准化时，再次检查分母：是标准误 σ/√n，而不仅仅是总体标准差。

12. Linking Hypothesis Tests with Confidence Intervals | 假设检验与置信区间的联系

A two-tailed hypothesis test at significance level α has a direct link with a (1 – α) confidence interval. If the null value lies outside the confidence interval, you reject H₀ at α significance. Conversely, if it lies inside, you do not reject.

显著性水平为 α 的双尾假设检验与 (1 – α) 置信区间直接关联。若零假设值落在置信区间外，则在 α 显著性水平下拒绝 H₀；反之，若落在区间内，则不拒绝。

Edexcel sometimes asks you to use a confidence interval to carry out a test. For example, a 95% confidence interval for a proportion that is (0.12, 0.28) and H₀: p = 0.3: since 0.3 is outside, we reject H₀. This method is acceptable and offers a quick cross-check.

Edexcel 有时要求使用置信区间进行检验。例如，某个比例的 95% 置信区间为 (0.12, 0.28)，而 H₀: p = 0.3，由于 0.3 位于区间外，我们拒绝 H₀。该方法可行，并提供了一种快速的交叉验证途径。

Understanding this relationship deepens your statistical reasoning and prepares you for questions that blend estimation and testing.

理解这层关系能深化你的统计推理，帮助你应对将估计与检验相结合的考题。

Published by TutorHao | Mathematics Revision Series | aleveler.com

更多咨询请联系16621398022（同微信）

A-Level Edexcel Mathematics: Hypothesis Testing | A-Level Edexcel 数学：假设检验 考点精讲