Hypothesis Testing | 假设检验考点精讲

📚 Hypothesis Testing | 假设检验考点精讲

Hypothesis testing is a fundamental statistical method used to make inferences or draw conclusions about a population based on sample data. It begins with a claim about a population parameter and uses probability to assess whether the observed sample provides sufficient evidence to support that claim. In the CIE GCSE Mathematics course, hypothesis testing is often introduced through binomial experiments, such as testing the fairness of a coin or the proportion of defective items in a batch.

假设检验是一种基本的统计方法，用于基于样本数据对总体进行推断或得出结论。它从一个关于总体参数的声明开始，利用概率来评估观测到的样本是否提供了足够的证据来支持该声明。在 CIE GCSE 数学课程中，假设检验通常通过二项实验引入，例如检验一枚硬币是否公平，或检查一批产品中次品的比例。

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

In everyday life, we often make decisions based on evidence. For example, if a friend claims they can consistently predict the outcome of a coin toss, you might test this by asking them to predict ten tosses. If they get most predictions right, you might start to believe their claim; if they fail many, you would dismiss it. This is the essence of hypothesis testing: we assume a claim is true unless the data strongly contradicts it.

在日常生活中，我们经常根据证据做决定。例如，如果一个朋友声称他们能持续预测抛硬币的结果，你可能会让他们预测十次来测试。如果他们大部分预测正确，你可能会相信他们的说法；如果他们错了很多，你就会否定这个说法。这就是假设检验的本质：假设一个声明成立，直到数据强烈反对它。

2. Formulating Null and Alternative Hypotheses | 零假设与备择假设的建立

The null hypothesis, denoted H₀, is the default position – the statement being tested. It usually represents “no effect” or “no difference”. The alternative hypothesis, H₁ or Hₐ, is what we accept if there is enough evidence against H₀. For a binomial test on a population proportion p, a typical null hypothesis is H₀: p = p₀, where p₀ is a specific value like 0.5 for a fair coin. The alternative could be H₁: p > p₀, H₁: p < p₀, or H₁: p ≠ p₀.

零假设，记作 H₀，是默认立场——正在被检验的陈述。它通常表示“无效果”或“无差异”。备择假设，记作 H₁ 或 Hₐ，是当有足够证据反对 H₀ 时我们接受的陈述。对于关于总体比例 p 的二项检验，典型的零假设是 H₀：p = p₀，p₀ 是某个特定值，例如公平硬币的 0.5。备择假设可以是 H₁：p > p₀，H₁：p < p₀，或 H₁：p ≠ p₀。

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

A one-tailed test looks for an effect in one direction only. For instance, testing whether a new drug increases recovery rate (H₁: p > 0.7). A two-tailed test looks for any difference, regardless of direction. For example, testing if a coin is biased, we use H₁: p ≠ 0.5. This choice affects how we calculate rejection regions and p-values.

单尾检验只在一个方向上寻找效应。例如，检验一种新药是否提高康复率（H₁：p > 0.7）。双尾检验寻找任何差异，无论方向。例如，检验一枚硬币是否有偏差，我们使用 H₁：p ≠ 0.5。这个选择会影响我们计算拒绝域和 p 值的方式。

4. Significance Level and Critical Region | 显著性水平与拒绝域

The significance level, denoted α (alpha), is the probability threshold for rejecting H₀ when it is actually true. Common values are 0.05 (5%) or 0.01 (1%). The critical region (or rejection region) consists of those values of the test statistic that lead to rejecting H₀. For a binomial test, the critical region is a set of outcomes with total probability ≤ α. If the observed result falls in this region, we reject H₀.

显著性水平，记作 α（alpha），是当 H₀ 实际为真时我们拒绝它的概率阈值。常用值有 0.05（5%）或 0.01（1%）。拒绝域（或临界域）由导致拒绝 H₀ 的检验统计量的值组成。对于二项检验，拒绝域是总概率 ≤ α 的一组结果。如果观测结果落入该区域，我们就拒绝 H₀。

5. Test Statistic: The Binomial Probability | 检验统计量：二项概率

In CIE GCSE hypothesis testing, the test statistic is often the number of successes X in n independent trials, where X ~ B(n, p). We assume p equals the value stated in H₀. For each observed x, we compute P(X ≥ x) or P(X ≤ x) or both tails, depending on H₁.

在 CIE GCSE 假设检验中，检验统计量通常是 n 次独立试验中成功的次数 X，且 X ~ B(n, p)。我们假设 p 等于 H₀ 中所述的值。对于每个观测到的 x，我们根据 H₁ 计算 P(X ≥ x) 或 P(X ≤ x) 或双尾概率。

P(X = k) = ⁿCₖ · pᵏ · (1 − p)ⁿ⁻ᵏ

The binomial coefficient ⁿCₖ counts the number of ways to choose k successes from n trials. Use your calculator’s nCr function for efficiency.

二项式系数 ⁿCₖ 计算从 n 次试验中选取 k 次成功的方法数。为了提高效率，可使用计算器的 nCr 功能。

6. Calculating p-values | 计算p值

The p-value is the probability of obtaining a result at least as extreme as the one observed, assuming H₀ is true. For a right‑tailed test (H₁: p > p₀), the p‑value is P(X ≥ x). For a left‑tailed test, it is P(X ≤ x). For a two‑tailed test, it is twice the smaller tail probability (or the sum of both tail probabilities if not symmetric). If the p‑value is less than α, we reject H₀.

p 值是在 H₀ 为真的条件下，得到至少与观测结果一样极端的结果的概率。对于右尾检验（H₁：p > p₀），p 值是 P(X ≥ x)。对于左尾检验，则是 P(X ≤ x)。对于双尾检验，它是较小尾概率的两倍（或不对称时两尾概率之和）。如果 p 值小于 α，我们拒绝 H₀。

7. Making a Decision: Reject or Not Reject | 做出决策：拒绝与否

There are two possible conclusions: reject H₀ if the evidence is sufficient; otherwise, do not reject H₀. Note that “accept H₀” is technically incorrect — failing to reject H₀ simply means the data are not inconsistent with H₀. Always state your decision in the context of the problem.

有两种可能的结论：如果证据充分则拒绝 H₀；否则，不拒绝 H₀。注意，“接受 H₀” 在术语上是不正确的——未能拒绝 H₀ 仅表示数据与 H₀ 并不矛盾。务必在问题背景下陈述你的决定。

8. 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 H₀ — the probability of this is α. A Type II error occurs when we fail to reject a false H₀ — its probability is denoted β. In GCSE problems, you may be asked to identify or describe these errors in context.

第一类错误发生在拒绝了真实的 H₀ 时——其概率就是 α。第二类错误发生在未能拒绝错误的 H₀ 时——其概率记作 β。在 GCSE 问题中，你可能需要在上下文里识别或描述这些错误。

Decision \ Reality	H₀ True	H₀ False
Reject H₀	Type I Error (α)	Correct
Do not reject H₀	Correct	Type II Error (β)

Understanding this table helps you appreciate the trade‑off between sensitivity and specificity in testing.

理解此表有助于你体会检验中灵敏度与特异度之间的权衡。

9. Contextual Conclusion | 结合上下文得出结论

Always translate the statistical decision back into the language of the original problem. For example: “There is sufficient evidence at the 5% level to suggest that the coin is biased in favour of heads.” Never just write “reject H₀”. Marks are often awarded for a clear, non‑technical summary.

始终将统计决定转译为原始问题的语言。例如：“在 5% 显著性水平下，有充分证据表明这枚硬币偏向正面。” 切勿只写“拒绝 H₀”。清晰的非技术性总结往往能得分。

10. Worked Example: Coin Fairness Test | 案例精讲：硬币公平性检验

A coin is tossed 10 times and lands heads 8 times. Test at the 5% significance level whether the coin is biased towards heads. Let p be the probability of heads. H₀: p = 0.5, H₁: p > 0.5 (one‑tailed). We observe X = 8 from B(10, 0.5). p‑value = P(X ≥ 8) = P(X=8) + P(X=9) + P(X=10). Calculate: P(X=8) = ¹⁰C₈ × 0.5⁸ × 0.5² = 45 × 0.5¹⁰ = 0.0439. P(X=9) = ¹⁰C₉ × 0.5¹⁰ = 10 × 0.0009766 = 0.0098. P(X=10) = 0.5¹⁰ = 0.00098. Total p‑value ≈ 0.0547. Since 0.0547 > 0.05, we do not reject H₀. Conclusion: there is insufficient evidence at the 5% level to say the coin is biased towards heads.

一枚硬币抛掷 10 次，出现 8 次正面。在 5% 显著性水平下检验该硬币是否偏向正面。设 p 为出现正面的概率。H₀：p = 0.5，H₁：p > 0.5（单尾）。我们观测到 X = 8，来自 B(10, 0.5)。p 值 = P(X ≥ 8) = P(X=8) + P(X=9) + P(X=10)。计算：P(X=8) = ¹⁰C₈ × 0.5⁸ × 0.5² = 45 × 0.5¹⁰ = 0.0439。P(X=9) = ¹⁰C₉ × 0.5¹⁰ = 10 × 0.0009766 = 0.0098。P(X=10) = 0.5¹⁰ = 0.00098。总 p 值 ≈ 0.0547。由于 0.0547 > 0.05，我们不拒绝 H₀。结论：在 5% 水平下，没有充分证据表明该硬币偏向正面。

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

Confusing the null and alternative hypotheses. Always set H₀ as the status quo.

混淆零假设与备择假设。始终将 H₀ 设为现状。
Using the wrong tail probability. Draw a diagram and check whether the test is one‑ or two‑tailed.

使用错误的尾概率。画图并检查是单尾还是双尾检验。
Treating the p‑value as the probability that H₀ is true. The p‑value is the probability of the data under H₀, not the reverse.

将 p 值视为 H₀ 为真的概率。p 值是在 H₀ 下获得该数据的概率，而非反过来的概率。
Forgetting to define p in context. Always start by defining the population parameter clearly.

忘记在上下文中定义 p。始终以明确界定总体参数开始。
Misapplying the 5% rule when the p‑value is exactly on the boundary – in such cases, follow the usual comparison rules; some syllabuses allow “borderline” discussion.

当 p 值恰好等于分界线时误用 5% 规则——此时按常规比较规则处理；某些大纲允许讨论“临界情况”。

12. Exam Tips for CIE GCSE | CIE GCSE考试技巧

Read the question carefully to identify the direction of the test. Show all steps: hypotheses, significance level, distribution, calculation of tail probabilities, p‑value or rejection region, decision, and contextual conclusion. Use clear notation and label outcomes. If the question asks for a critical region, list the specific values of X that would lead to rejecting H₀. Practice with different sample sizes and values of p to become comfortable switching between tables, calculator functions, and formulas.

仔细阅读题目以确定检验的方向。展示所有步骤：假设、显著性水平、分布、尾概率的计算、p 值或拒绝域、决策，以及结合上下文的结论。使用清晰的符号并标注结果。如果题目要求拒绝域，列出会导致拒绝 H₀ 的 X 的具体值。练习不同的样本量和 p 值，以便熟练在表格、计算器功能和公式之间切换。

Published by TutorHao | Mathematics Revision Series | aleveler.com

更多咨询请联系16621398022（同微信）

Hypothesis Testing | 假设检验 考点精讲