📚 GCSE CCEA Maths: Hypothesis Testing Essentials | 假设检验考点精讲
Hypothesis testing is a formal statistical procedure used to decide whether to accept or reject a claim about a population parameter based on sample data. For GCSE CCEA mathematics, this topic focuses on testing a proportion using the binomial distribution. You will be expected to set up null and alternative hypotheses, choose a significance level, calculate probabilities from binomial tables, and draw a conclusion in context. Understanding this logical framework not only helps you secure marks in the T6 paper but also builds a foundation for A‑level statistics.
假设检验是一种正式的统计步骤，用于根据样本数据决定是接受还是拒绝一个关于总体参数的声明。在 GCSE CCEA 数学中，本专题的重点是利用二项分布检验一个比例。你需要设定零假设和备择假设、选择显著性水平、根据二项分布表计算概率，并结合实际情境得出结论。理解这个逻辑框架不仅能帮你在 T6 试卷中拿到分数，也为 A‑level 统计学习打下基础。
1. What Is Hypothesis Testing? | 什么是假设检验？
Hypothesis testing is a method of statistical inference that allows us to test an assumption about a population parameter. The assumption is called the null hypothesis, and we use sample data to judge whether it is likely to be true. The process is similar to a court trial: we assume innocence (the null) unless there is enough evidence to prove guilt (the alternative). In GCSE CCEA problems, the population parameter is usually the probability of success p in a binomial distribution, and the sample is a fixed number of trials.
假设检验是一种统计推断方法，允许我们检验关于总体参数的某个假定。这个假定称为零假设，我们用样本数据来判断它是否有可能成立。整个过程类似于法庭审判：我们先假定无罪（零假设），除非有足够证据证明有罪（备择假设）。在 GCSE CCEA 题目中，总体参数通常是二项分布中的成功概率 p，样本则是一组固定次数的试验。
2. Null and Alternative Hypotheses | 零假设与备择假设
The null hypothesis, denoted H₀ ggests otherwise. It always contains an equals sign (p = ...). The alternative hypothesis, H₁, is what we are trying to find evidence for. It uses a strict inequality: p < ..., p > ... or p ≠ ... . In CCEA exams, you must write the hypotheses using the probability symbol p and define it in words if needed.
零假设，记作 H₀，是我们假定为真、除非有令人信服的证据才拒绝的陈述。它总是包含等号（p = …）。备择假设 H₁ 是我们试图寻找证据支持的观点，使用严格不等号：p < …、p > … 或 p ≠ …。在 CCEA 考试中，你必须用概率符号 p 写出假设，并在必要时用文字定义它。
Example: A manufacturer claims that no more than 10% of its light bulbs are defective. If we suspect the true proportion is higher, we write:
例如：一家制造商声称其灯泡的不合格率不超过 10%。如果我们怀疑实际比例更高，则写出：
H₀: p = 0.10 H₁: p > 0.10
3. Significance Level | 显著性水平
The significance level, usually denoted by the Greek letter α (alpha), is the probability of rejecting a true null hypothesis. It is the maximum risk of a Type I error we are willing to accept. In GCSE CCEA questions, the significance level is almost always given as a percentage such as 5% or 1%. You will be told to 'test at the 5% significance level' or similar.
显著性水平，通常用希腊字母 α 表示，是当零假设为真时我们拒绝它的概率，也就是我们愿意接受的 I 类错误的最大风险。在 GCSE CCEA 题目中，显著性水平几乎总是以百分数给出的，如 5% 或 1%。题目会要求你 “在 5% 的显著性水平下检验” 或类似表述。
A smaller significance level means we need stronger evidence to reject H₀. For a 5% level, the critical region covers the most extreme 5% of outcomes under the null hypothesis.
显著性水平越小，意味着我们需要越强的证据才能拒绝 H₀。在 5% 水平下，拒绝域包含了零假设成立时最极端的 5% 的结果。
4. 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. For example, H₁: p > 0.3 is right‑tailed, and H₁: p < 0.3 is left‑tailed. The significance level is placed entirely in one tail of the distribution. A two‑tailed test is used when H₁: p ≠ 0.3: we are simply looking for a difference in either direction. In this case, the significance level is split equally between both tails (2.5% each side for a 5% test).
当备择假设带有方向性时，使用单尾检验。例如，H₁: p > 0.3 是右尾检验，H₁: p < 0.3 是左尾检验。显著性水平全部放在分布的一个尾部。当备择假设为 H₁: p ≠ 0.3 时，我们只关心是否存在任何方向的差异，这时使用双尾检验，显著性水平平分到两个尾部（5% 检验中每侧 2.5%）。
In CCEA questions, one‑tailed tests are more common, but you must read the wording carefully. Words like 'greater than', 'increased', 'higher' suggest a right‑tailed test; 'less than', 'decreased' suggest a left‑tailed test; 'changed', 'different' suggest a two‑tailed test.
在 CCEA 试题中，单尾检验更常见，但你必须仔细阅读措辞。“greater than”、“increased”、“higher” 暗示右尾检验；“less than”、“decreased” 暗示左尾检验；“changed”、“different” 暗示双尾检验。
5. Test Statistic and Binomial Distribution | 检验统计量与二项分布
The test statistic is the observed count of successes in the sample, denoted by X. Under the null hypothesis, X follows a binomial distribution: X ~ B(n, p₀), where n is the sample size and p₀ is the value specified in H₀. All probability calculations are based on this distribution. You will typically use cumulative binomial tables or a calculator to find probabilities.
检验统计量是样本中观察到的成功次数，记作 X。在零假设下，X 服从二项分布：X ~ B(n, p₀)，其中 n 是样本大小，p₀ 是 H₀ 中指定的概率值。所有概率计算都基于这个分布。你通常需要使用二项累积分布表或计算器来求概率。
For example, if H₀: p = 0.5 and n = 20, then under H₀, X ~ B(20, 0.5). You can then work out P(X ≥ observed value) or P(X ≤ observed value).
例如，若 H₀: p = 0.5 且 n = 20，那么在 H₀ 下 X ~ B(20, 0.5)。你可以接着计算 P(X ≥ 观察值) 或 P(X ≤ 观察值)。
6. Critical Region Approach | 临界区域法
ical region, you find the largest or smallest values of X such that P(X in that region) ≤ α. For a right‑tailed test, you look for the smallest value k such that P(X ≥ k) ≤ α; for a left‑ that P(X ≤ k) ≤ α. If the observed test statistic falls in the critical region, reject H₀.
拒绝域是使我们拒绝 H₀ 的检验统计量的所有取值组成的集合，其边界值称为临界值。找到拒绝域的方法是：求满足 P(X 落在该区域) ≤ α 的最大或最小的 X 值。对于右尾检验，寻找最小的 k 使得 P(X ≥ k) ≤ α；对于左尾检验，寻找最大的 k 使得 P(X ≤ k) ≤ α。如果观察到的检验统计量落在拒绝域内，则拒绝 H₀。
For a two‑ tical regions: one at the lower end and one at the upper end, each with probability ≤ α/2.
对于双尾检验，你需要找到两个拒绝域：一个在下尾，一个在上尾，各自的概率 ≤ α/2。
7. p‑Value Approach | p 值法
The p‑ a test statistic at least as extreme as the observed one, assuming H₀ is true. For a right‑tailed test, p‑value = P(X ≥ observed value); for a left‑tailed test, p‑value = P(X ≤ observed value); for a two‑tailed test, p‑ e) if the observed value is above the mean, or similarly for the lower tail. Compare the p‑value with α: if p‑value ≤ α, reject H₀; if p‑value > α, do not reject H₀.
p 值是在 H₀ 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。对于右尾检验，p 值 = P(X ≥ 观察值)；对于左尾检验，p 值 = P(X ≤ 观察值)；对于双尾检验，p 值 = 2 × P(X ≥ 观察值)（若观察值高于均值，或在低尾类似处理）。将 p 值与 α 比较：若 p 值 ≤ α，则拒绝 H₀；若 p 值 > α，则不拒绝 H₀。
Many CCEA mark schemes accept either the critical region or the p‑value method. The p‑value method is often easier because you directly compare a probability with α.
很多 CCEA 评分方案两种方法都接受。p 值法通常更简单，因为你直接把一个概率与 α 比较。
8. Step‑by‑Step Procedure | 完整步骤
Here is a reliable sequence for any hypothesis test question in CCEA GCSE:
以下是 CCEA GCSE 中任何假设检验问题都适用的可靠步骤：
Step 1: Define the population parameter p and state the hypotheses clearly.
第 1 步：定义总体参数 p，并清晰地陈述假设。
Step 2: Write down the significance level α.
第 2 步：写下显著性水平 α。
Step 3: State the distribution of the test statistic under H₀: X ~ B(n, p₀).
第 3 步：陈述在 H₀ 下检验统计量的分布：X ~ B(n, p₀)。
Step 4: Record the observed value of X from the sample.
第 4 步：记录样本中 X 的观察值。
Step 5: Calculate either the critical region or the p‑value.
第 5 步：计算拒绝域或 p 值。
Step 6: Compare with α (or check if observed X lies in the critical region).
第 6 步：与 α 比较（或检查观察值 X 是否落在拒绝域内）。
Step 7: Write a conclusion: 'reject H₀' or 'do not reject H₀'. Always relate your conclusion to the original problem.
第 7 步：写出结论：“拒绝 H₀” 或 “不拒绝 H₀”。始终使结论与原始问题相联系。
9. Worked Example 1 – Testing a Coin | 例题 1 – 检验一枚硬币
Problem: A coin is tossed 20 times, resulting in 15 heads. Test at the 5% significance level whether the coin is biased towards heads.
问题：一枚硬币抛掷 20 次，得到 15 次正面。在 5% 的显著性水平下检验该硬币是否偏向正面。
Solution: Let p = probability of heads.
解：令 p = 得到正面的概率。
H₀: p = 0.5 H₁: p > 0.5 (right‑tailed test)
Significance level α = 0.05.
Under H₀, number of heads X ~ B(20, 0.5).
Observed value: X = 15.
p‑value = P(X ≥ 15) = 1 − P(X ≤ 14).
From cumulative binomial tables, P(X ≤ 14) = 0.9793.
So p‑value = 1 − 0.9793 = 0.0207.
Since 0.0207 ≤ 0.05, we reject H₀.
Conclusion: There is sufficient evidence at the 5% level to suggest that the coin is biased towards heads.
因为 0.0207 ≤ 0.05，我们拒绝 H₀。
结论：有充分证据表明，在 5% 水平下该硬币偏向正面。
10. Worked Example 2 – Testing a Claim about a Proportion | 例题 2 – 检验关于比例的声明
Problem: A company claims that at least 80% of its customers are satisfied. A survey of 15 customers finds that 10 are satisfied. Test the claim at the 5% significance level.
问题：一家公司声称至少有 80% 的顾客满意。一项对 15 名顾客的调查发现 10 名满意。在 5% 的显著性水平下检验该声明。
Solution: tomers. The claim is p ≥ 0.80. We are testing if the true proportion is less than claimed.
解：令 p = 满意顾客的比例。声明为 p ≥ 0.80。我们检验真实比例是否低于声明。
H₀: p = 0.80 H₁: p < 0.80 (left‑tailed test)
α = 0.05, n = 15.
Under H₀, X ~ B(15, 0.80) where X = number of satisfied customers.
Observed X = 10. The p‑value = P(X ≤ 10). From binomial tables, P(X ≤ 10) = 0.1642 (approximately).
0.1642 > 0.05, so we do not reject H₀.
Conclusion: There is insufficient evidence at the 5% significance level to reject the company's claim. The data do not provide enough proof that the proportion satisfied is less than 80%.
0.1642 > 0.05，因此我们不拒绝 H₀。
结论：在 5% 显著性水平下，没有足够证据拒绝公司的声明。数据未能提供足够证明说明满意比例低于 80%。
11. Interpreting Conclusions in Context | 结合情境解读结论
A hypothesis test never 'proves' that H₀ is true or false; it only tells us whether the sample data are unlikely under H₀. If you reject H₀, you say there is evidence for H₁. If you do not reject H₀, you say there is not enough evidence for H₁ – this does not confirm H₀ is correct, only that it cannot be ruled out. Always write your conclusion in plain English, referring back to the question.
假设检验从不 “证明” H₀ 为真或为假；它只告诉我们样本数据在 H₀ 下是否不太可能出现。如果你拒绝 H₀，则说有证据支持 H₁。如果你不拒绝 H₀，则说没有足够的证据支持 H₁——这并不证实 H₀ 正确，只意味着不能将其排除。一定要用平实的语言结合题意写出结论。
Phrases like 'there is sufficient evidence to suggest...' or 'the result is significant at the 5% level' are very useful.
“有充分证据表明……” 或 “结果在 5% 水平下是显著的” 这类措辞非常有用。
12. Common Mistakes to Avoid | 常见错误提醒
Mistake 1: Writing H₁ with an equals sign (H₁: p = 0.6). The alternative hypothesis must not contain '='.
错误 1：在 H₁ 中使用等号（H₁: p = 0.6）。备择假设中绝不能包含 “=”。
Mistake 2: Using the wrong tail. For 'at least' claims, the test is usually left‑tailed; for 'no more than', it is right‑tailed. Always identify the direction from the context, not just the words in isolation.
错误 2：用错尾部。对于 “at least” 类声明，检验通常是左尾的；对于 “no more than” 类声明，检验是右尾的。务必结合语境判断方向，而不能只看孤立词汇。
Mistake 3: Forgetting to state the distribution X ~ B(n, p₀) under H₀. This statement often earns a method mark.
错误 3：忘记陈述 H₀ 下的分布 X ~ B(n, p₀)。这个陈述通常能挣得方法分。
Mistake 4: Confusing the observed value with the sample size. The test statistic is the count of successes, not n itself.
错误 4：混淆观察值与样本大小。检验统计量是成功的次数，而不是样本大小 n 本身。
Mistake 5: Misinterpreting 'do not reject H₀' as 'H₀ is true'. Always say 'insufficient evidence' rather than 'prove'.
错误 5：将 “不拒绝 H₀” 误解为 “H₀ 为真”。永远要说 “证据不足” 而不是 “证明”。
Mistake 6: In a two‑tailed test, failing to double the probability when calculating the p‑value.
错误 6：在双尾检验中，计算 p 值时忘记将概率加倍。
Avoiding these errors will make your solutions clear, accurate, and examiner‑friendly.
避免这些错误能使你的解答清晰、准确并受阅卷人欢迎。
Published by TutorHao | Mathematics Revision Series | aleveler.com
更多咨询请联系16621398022（同微信）
或截图后用「扫一扫 → 相册」选择本图