Two students walk out of the A-Level Maths exam. One is beaming — the 12-mark binomial hypothesis testing question was a breeze. The other looks defeated — they confused the null hypothesis with the alternative and lost crucial marks. What was the difference? The first student understood not just the formulas, but the logic behind them. If you’re preparing for Edexcel, AQA, OCR, or CIE A-Level Mathematics, this guide will take you from confusion to confidence in binomial distributions and hypothesis testing.
两个学生走出 A-Level 数学考场。一个笑容满面——那道 12 分的二项分布假设检验题轻松搞定。另一个面如死灰——他把零假设和备择假设搞反了，丢了关键分。区别在哪里？第一个学生不仅懂公式，更懂公式背后的逻辑。如果你正在备战 Edexcel、AQA、OCR 或 CIE A-Level 数学，本指南将带你从困惑走向自信，彻底掌握二项分布与假设检验。
1. What Is a Binomial Distribution? / 什么是二项分布？
A binomial distribution models the number of successes in a fixed number of independent trials, where each trial has exactly two possible outcomes: success or failure. Think of flipping a coin 10 times and counting the heads, or checking 20 products off an assembly line and counting the defective ones. If each trial has the same probability of success p, and the trials are independent, you’re in binomial territory.
二项分布描述的是在固定次数的独立试验中，成功次数的概率分布。每次试验只有两种可能结果：成功或失败。想象抛硬币 10 次并数正面朝上的次数，或者检查流水线上的 20 件产品并统计次品数量。如果每次试验的成功概率 p 相同，且各次试验相互独立，那你就进入了二项分布的世界。
In A-Level exam notation, we write: X ~ B(n, p), where n is the number of trials and p is the probability of success in each trial. The random variable X represents the number of successes.
在 A-Level 考试符号中，我们写作：X ~ B(n, p)，其中 n 是试验次数，p 是每次试验的成功概率。随机变量 X 表示成功的次数。
For a variable to be binomially distributed, it must satisfy four conditions — and examiners love to test these:
要满足二项分布，必须满足四个条件——考官特别喜欢考这些：
- Fixed number of trials (n) — you know exactly how many trials there are before you start / 固定试验次数 (n)——开始之前你就知道有多少次试验
- Two outcomes per trial — success or failure, nothing in between / 每次试验两种结果——成功或失败，没有中间状态
- Constant probability (p) — the probability of success doesn’t change from trial to trial / 恒定概率 (p)——每次试验的成功概率不变
- Independent trials — one trial’s outcome doesn’t affect another / 独立试验——一次试验的结果不影响其他试验
2. The Binomial Probability Formula / 二项分布概率公式
This is the most important formula in the entire topic. Commit it to memory and understand how every part works:
这是整个主题中最重要的公式。请牢记于心，并理解每一部分的作用：
Let’s break this down piece by piece:
让我们逐一拆解：
or
: the number of ways to choose r successes from n trials. Your calculator has a dedicated nCr button — use it! / 组合数：从 n 次试验中选出 r 次成功的方式数。计算器上有专门的 nCr 按键——用它！
: probability of getting r successes in a row /
: probability of getting (n-r) failures /
Worked Example 1 / 例题 1
A fair die is rolled 8 times. Find the probability of getting exactly 3 sixes.
一个公平的骰子掷 8 次。求恰好掷出 3 次六点的概率。
Here: n = 8, r = 3, p = 1/6, (1-p) = 5/6
So there’s about a 10.4% chance of rolling exactly 3 sixes in 8 rolls. Not rare, but not common either!
所以掷 8 次骰子，恰好出现 3 次六点的概率约为 10.4%。不罕见，但也不常见！
Using the Formula for Range Probabilities / 使用公式计算区间概率
Examiners frequently ask for P(X ≤ 3), P(X > 5), or P(2 ≤ X ≤ 6). The key insight: the binomial distribution is discrete, so P(X ≤ 3) = P(X = 0) + P(X = 1) + P(X = 2) + P(X = 3). You add up individual probabilities. Your calculator’s binomial CD (cumulative distribution) function does this instantly — learn to use it!
考官经常要求计算 P(X ≤ 3)、P(X > 5) 或 P(2 ≤ X ≤ 6)。关键洞察：二项分布是离散的，所以 P(X ≤ 3) = P(X = 0) + P(X = 1) + P(X = 2) + P(X = 3)。你把各个概率加起来。计算器上的二项分布 CD（累积分布）功能可以瞬间完成——学会使用它！
3. Mean, Variance, and Shape / 均值、方差与分布形态
Every binomial distribution has two key summary statistics that appear repeatedly in exam questions:
每个二项分布都有两个关键的汇总统计量，在考试题目中反复出现：
| Statistic / 统计量 | Formula / 公式 | Intuition / 直觉理解 |
|---|---|---|
| Mean / 均值 |  |
Average number of successes you’d expect / 你预期获得的平均成功次数 |
| Variance / 方差 |  |
Measures how spread out the distribution is / 衡量分布的离散程度 |
| Standard Deviation / 标准差 |  |
Typical deviation from the mean / 典型的偏离均值程度 |
Critical insight about shape: When p = 0.5, the binomial distribution is perfectly symmetric. When p < 0.5, it skews right (tail extends to higher values). When p > 0.5, it skews left. As n increases, the distribution becomes more symmetric and looks increasingly like a normal distribution — hence the Normal approximation for large n (when np > 5 and n(1-p) > 5).
关于形态的关键洞察：当 p = 0.5 时，二项分布完全对称。当 p < 0.5 时，分布右偏（尾部延伸到较高值）。当 p > 0.5 时，分布左偏。随着 n 增大，分布变得更对称，越来越像正态分布——这就是大 n 情况下的正态近似（当 np > 5 且 n(1-p) > 5 时适用）。
Worked Example 2 / 例题 2
A biased coin has P(heads) = 0.3. It is tossed 50 times. Find the mean and variance of the number of heads.
一枚偏倚硬币，P(正面) = 0.3。抛掷 50 次。求正面朝上次数的均值和方差。
Mean = np = 50 × 0.3 = 15 heads
Variance = np(1-p) = 50 × 0.3 × 0.7 = 10.5
Standard deviation = √10.5 ≈ 3.24
We’d expect around 15 heads, give or take about 3. The distribution is right-skewed (p < 0.5), with the right tail potentially reaching toward 25-30 heads.
我们预期大约 15 次正面，误差约 3 次。分布为右偏（p < 0.5），右尾可能延伸到 25-30 次正面。
4. Introduction to Hypothesis Testing / 假设检验简介
Now we reach the topic that separates A* students from A students: hypothesis testing. This is where the binomial distribution becomes a powerful tool for making decisions based on data. At its core, hypothesis testing asks: “Does the evidence support my claim, or could this just be random chance?”
现在我们来到了区分 A* 学生和 A 学生的主题：假设检验。在这里，二项分布成为基于数据做出决策的强大工具。假设检验的核心问题是：“证据支持我的主张，还是这仅仅是随机偶然？”
The Five-Step Framework / 五步框架
Every A-Level hypothesis test follows the same structure. Master this framework, and you master the topic:
每个 A-Level 假设检验都遵循相同的结构。掌握这个框架，你就掌握了这个主题：
| Step | Description / 描述 | Key Words / 关键词 |
|---|---|---|
| 1 | Define hypotheses / 定义假设 | H₀: null hypothesis (status quo) / H₁: alternative hypothesis (what you suspect) |
| 2 | State significance level / 陈述显著性水平 | Usually α = 0.05 (5%) or 0.01 (1%), given in the question |
| 3 | Define test statistic and distribution / 定义检验统计量和分布 | X ~ B(n, p) — specify n and the p under H₀ |
| 4 | Calculate p-value or critical region / 计算 p 值或临界域 | Use calculator or tables; compare against α |
| 5 | Draw conclusion in context / 在上下文中得出结论 | “Reject H₀” or “Do not reject H₀” — always in words, always with context |
H₀ and H₁: The Most Common Source of Confusion / 最常见的混淆来源
Null hypothesis (H₀): This is the “nothing has changed” position. It assumes the claimed probability equals some specific value. Think of it as the “boring” hypothesis that nothing interesting is happening. / 零假设 (H₀)：这是”什么都没变”的立场。它假设声称的概率等于某个特定值。可以把它看作”无聊”的假设——没发生什么有意思的事。
Alternative hypothesis (H₁): This is what you’re trying to prove. It states that the probability has changed (two-tailed test) or moved in a specific direction (one-tailed test). / 备择假设 (H₁)：这是你试图证明的。它声明概率已改变（双尾检验）或朝特定方向移动（单尾检验）。
Critical rule from examiners’ reports: H₀ always contains an equals sign (=). H₁ never does. If you write H₀: p ≥ 0.5, you’re wrong — it should be H₀: p = 0.5 (with H₁: p < 0.5 for a lower-tail test).
来自考官报告的关键规则：H₀ 总是包含等号（=）。H₁ 从不包含。如果你写 H₀: p ≥ 0.5，那是错的——应该写 H₀: p = 0.5（下尾检验对应 H₁: p < 0.5）。
5. One-Tailed vs Two-Tailed Tests / 单尾与双尾检验
The direction of the test changes EVERYTHING — the critical region, the p-value calculation, and the conclusion. Here’s how to tell which one to use:
检验的方向改变一切——临界域、p 值计算和结论都不同。以下是如何判断使用哪种：
| Clue in Question / 题目中的线索 | Test Type / 检验类型 | H₁ / 备择假设 |
|---|---|---|
| “Has the probability increased?” / “概率是否增加了？” | Upper-tail / 上尾 |  |
| “Has the probability decreased?” / “概率是否减少了？” | Lower-tail / 下尾 | $latex \displaystyle H_1: p < k$ |
| “Has the probability changed?” / “概率是否改变了？” | Two-tailed / 双尾 |  |
Two-tailed test rule: When H₁ is p ≠ k, you split the significance level between both tails. For a 5% significance level, each tail gets 2.5%. So you reject H₀ if the test statistic falls in the lower 2.5% or upper 2.5% of the distribution.
双尾检验规则：当 H₁ 为 p ≠ k 时，你将显著性水平平分到两个尾部。对于 5% 显著性水平，每个尾部各占 2.5%。所以如果检验统计量落在分布的下 2.5% 或上 2.5% 区域，你就拒绝 H₀。
6. Finding Critical Values / 寻找临界值
The critical value is the boundary that separates the rejection region from the acceptance region. There are two equivalent approaches:
临界值是分隔拒绝域和接受域的边界。有两种等效的方法：
Method 1 — Critical Region Approach: Find the value(s) of X where P(X ≥ k) ≤ α/2 (upper tail) or P(X ≤ k) ≤ α/2 (lower tail). If your observed test statistic falls in this region, reject H₀.
方法一——临界域法：找到使得 P(X ≥ k) ≤ α/2（上尾）或 P(X ≤ k) ≤ α/2（下尾）的 X 值。如果你观察到的检验统计量落在这个区域，拒绝 H₀。
Method 2 — p-Value Approach: Calculate the probability of observing a result at least as extreme as yours, assuming H₀ is true. If p-value < α, reject H₀. This is increasingly preferred by exam boards.
方法二——p 值法：计算在 H₀ 为真的前提下，观察到至少与你得到的结果一样极端的值的概率。如果 p 值 < α，拒绝 H₀。各考试局越来越倾向于这种方法。
Calculator Tips / 计算器技巧
For Casio FX-991EX or CG50: Use Menu → Statistics → DIST → BINOMIAL → Bcd for cumulative probabilities. For finding critical values, use InvB (inverse binomial). For TI-Nspire: Use Menu → Statistics → Distributions → Binomial Cdf.
对于 Casio FX-991EX 或 CG50：使用 Menu → 统计 → 分布 → 二项分布 → Bcd 计算累积概率。要寻找临界值，使用 InvB（逆二项分布）。对于 TI-Nspire：使用 Menu → Statistics → Distributions → Binomial Cdf。
7. Full Worked Example — Hypothesis Test / 完整例题——假设检验
A pharmaceutical company claims that a new drug is effective for 70% of patients. A doctor suspects the drug is less effective than claimed and tests it on 20 patients, finding that only 10 show improvement. Test at the 5% significance level whether this evidence suggests the drug is less effective than claimed.
一家制药公司声称一种新药对 70% 的患者有效。一位医生怀疑该药的实际效果不如声称的那么好，在 20 名患者上测试，发现只有 10 名显示出改善。以 5% 的显著性水平检验，这个证据是否表明该药的实际有效率低于声称值。
Step 1 — Hypotheses / 假设：
H₀: p = 0.7 (the drug is effective 70% of the time / 药物有效率为 70%)
H₁: p < 0.7 (the drug is effective less than 70% of the time / 药物有效率低于 70%)
Step 2 — Significance level / 显著性水平： α = 0.05
Step 3 — Distribution under H₀ / H₀ 下的分布： X ~ B(20, 0.7)
Step 4 — Find critical region or p-value / 寻找临界域或 p 值：
We need P(X ≤ 10) assuming p = 0.7. Using the calculator:
Using cumulative binomial tables or calculator: P(X ≤ 10) ≈ 0.0480
Step 5 — Conclusion / 结论：
Since p-value = 0.0480 < 0.05, we reject H₀. There is sufficient evidence at the 5% significance level to suggest that the drug is effective for less than 70% of patients. The doctor’s suspicion is supported by the data.
因为 p 值 = 0.0480 < 0.05，我们拒绝 H₀。在 5% 显著性水平上有充分证据表明，该药对不到 70% 的患者有效。医生的怀疑得到数据支持。
Alternative approach using critical region: Find c such that P(X ≤ c) ≤ 0.05. From tables, P(X ≤ 9) ≈ 0.0171 and P(X ≤ 10) ≈ 0.0480. The critical region for a 5% lower-tail test is X ≤ 10 (since 0.0480 ≤ 0.05). Since observed X = 10 falls in the critical region, reject H₀.
使用临界域的替代方法：找到使得 P(X ≤ c) ≤ 0.05 的 c。查表得 P(X ≤ 9) ≈ 0.0171，P(X ≤ 10) ≈ 0.0480。5% 下尾检验的临界域是 X ≤ 10（因为 0.0480 ≤ 0.05）。由于观察值 X = 10 落在临界域内，拒绝 H₀。
8. Common Exam Pitfalls and How to Avoid Them / 常见考试陷阱及应对策略
Having marked thousands of A-Level scripts, examiners consistently flag the same mistakes. Here are the top five and how to dodge them:
批阅了数千份 A-Level 试卷后，考官们反复指出相同的错误。以下是前五名及应对方法：
Pitfall 1: Confusing H₀ and H₁ / 陷阱一：搞混 H₀ 和 H₁
What students do: Write H₀: p > 0.5 or H₁: p = 0.5. Both are wrong.
The fix: H₀ always has “=”. H₁ has “<", ">“, or “≠”. The null hypothesis is the one you’re trying to disprove — it’s the skeptical position. / 修正方法：H₀ 总是带 “=”。H₁ 带 “<"、">” 或 “≠”。零假设是你要试图推翻的——它是怀疑者的立场。
Pitfall 2: Wrong Tail / 陷阱二：选错尾部
What students do: Use a two-tailed test when the question says “increased,” or use an upper-tail test when the data shows a decrease.
The fix: Read the wording carefully. “Increased” = upper-tail. “Decreased” = lower-tail. “Changed” or “different” = two-tailed. / 修正方法：仔细读题。”增加”=上尾。”减少”=下尾。”改变”或”不同”=双尾。
Pitfall 3: Forgetting to Double the p-Value / 陷阱三：忘记将 p 值加倍
What students do: In a two-tailed test, they calculate P(X ≥ observed) or P(X ≤ observed) and compare directly to α.
The fix: For two-tailed tests with symmetric calculations, p-value = 2 × P(X ≥ observed) or 2 × P(X ≤ observed), whichever tail you observed in. Compare this doubled value to α. / 修正方法：对于对称计算的双尾检验，p 值 = 2 × P(X ≥ 观察值) 或 2 × P(X ≤ 观察值)，取决于你观察到的尾部。将加倍后的值与 α 比较。
Pitfall 4: Using the Wrong n or p / 陷阱四：用了错误的 n 或 p
What students do: Use the sample proportion in the binomial distribution instead of the claimed value from H₀.
The fix: The binomial distribution is ALWAYS set up using the p from H₀, not the sample estimate. X ~ B(n, p_under_H0). Always. / 修正方法：二项分布始终使用 H₀ 中的 p 来设定，而不是样本估计值。X ~ B(n, H₀_下的_p)。始终如此。
Pitfall 5: Weak Conclusion / 陷阱五：结论不充分
What students do: Write “Reject H₀” with no context, no mention of significance level, no real-world interpretation.
The fix: Use this template: “Since [p-value] < [α] OR [test statistic] is in the critical region, we reject H₀. There is sufficient evidence at the [α]% significance level to suggest that [real-world claim]." / 修正方法：使用这个模板：”由于 [p 值] < [α] 或 [检验统计量] 在临界域内，我们拒绝 H₀。在 [α]% 显著性水平上有充分证据表明 [现实主张]。"
9. Type I and Type II Errors / 第一类错误和第二类错误
No hypothesis test is perfect. Understanding errors takes your answer from A-grade to A*-grade, especially on longer exam questions:
没有哪个假设检验是完美的。理解错误类型会让你的答案从 A 级提升到 A* 级，尤其在较长的考题中：
| Error Type / 错误类型 | Definition / 定义 | Probability / 概率 | Real-World Analogy / 现实类比 |
|---|---|---|---|
| Type I / 第一类 | Rejecting H₀ when it’s actually true / H₀ 为真时拒绝它 |  (significance level) |
False alarm — convicting an innocent person / 虚惊——给无辜者定罪 |
| Type II / 第二类 | Not rejecting H₀ when it’s actually false / H₀ 为假时未拒绝它 |  (depends on true p) |
Missed detection — letting a guilty person go free / 漏检——放走犯罪者 |
Exam tip: If a question asks “explain what a Type I error means in this context,” don’t just repeat the definition. Say: “A Type I error would occur if the company concludes the drug is less effective than 70% when in reality it IS 70% effective — they might withdraw a perfectly good drug from the market.” Context is everything.
考试技巧：如果题目问”在这个背景下解释第一类错误的含义”，不要只是重复定义。要说：”如果公司得出结论认为药物有效率低于 70%，而实际上它确实有 70% 的有效率，那就发生了第一类错误——公司可能会将一个完全有效的药撤出市场。”背景就是一切。
10. The Binomial Distribution in the Bigger Picture / 二项分布在更大图景中的位置
Binomial distribution is not just a standalone topic — it connects to almost every other part of A-Level Statistics:
二项分布不仅是一个独立主题——它几乎与 A-Level 统计学的每个其他部分都有关联：
- Normal Approximation: When n is large (np ≥ 5 and nq ≥ 5), Binomial ~ Normal. Apply continuity correction. This appears in Paper 3 for all major exam boards. / 正态近似：当 n 较大时（np ≥ 5 且 nq ≥ 5），二项分布近似正态分布。应用连续性校正。这出现在所有主要考试局的 Paper 3 中。
- Poisson Approximation: When n is large and p is small (typically n ≥ 50, p ≤ 0.1), Binomial ~ Poisson(λ = np). / 泊松近似：当 n 大且 p 小时（通常 n ≥ 50，p ≤ 0.1），二项分布近似泊松分布 λ = np。
- Chi-Squared Tests: The binomial provides the theoretical foundation for goodness-of-fit tests — the expected frequencies under H₀ come from binomial probabilities. / 卡方检验：二项分布为拟合优度检验提供了理论基础——H₀ 下的期望频率来自二项概率。
- Sampling Distributions: The sample proportion p̂ follows an approximately normal distribution whose variance is derived from the binomial variance:
/ 抽样分布：样本比例 p̂ 近似遵循正态分布，其方差来自二项方差：
11. Exam Strategy and Time Management / 考试策略与时间管理
Binomial and hypothesis testing questions typically appear as 8-15 mark questions in A-Level Pure/Statistics papers. Here’s how to approach them efficiently:
二项分布和假设检验题目通常在 A-Level 纯数/统计试卷中以 8-15 分的题目出现。以下是高效应对的方法：
| Time / 时间 | Marks / 分值 | What to Do / 做什么 |
|---|---|---|
| 2 min | 2-3 marks | State H₀, H₁, and define X ~ B(n, p) / 陈述 H₀、H₁，定义 X ~ B(n, p) |
| 3-4 min | 4-5 marks | Calculate probabilities, find critical region or p-value / 计算概率，寻找临界域或 p 值 |
| 2 min | 2-3 marks | Write conclusion in context, discuss errors if asked / 在上下文中写出结论，如被要求则讨论错误 |
| 1 min | — | Check: Are the hypotheses correct? Did I use p from H₀? Is my conclusion in context? / 检查：假设是否正确？我是否使用了 H₀ 中的 p？结论是否在上下文中？ |
Golden rule: Marks are awarded for METHOD, not just the final answer. Even if your numerical answer is wrong, you can score most of the marks by showing correct hypotheses, correct distribution, and a clear step-by-step approach. Never leave a hypothesis testing question blank!
黄金法则：分数取决于方法，而不仅仅是最终答案。即使数值答案错误，通过展示正确的假设、正确的分布和清晰的逐步方法，你也能获得大部分分数。永远不要留空假设检验题！
12. Practice Questions / 练习题
Try these before your exam. Answers are worth working out yourself — that’s where the learning happens:
考试前试试这些。答案值得你自己算出来——学习就发生在那里：
Q1: A spinner has a 25% chance of landing on red. In 15 spins, find:
(a) The probability of exactly 5 reds
(b) The probability of at least 3 reds
(c) The expected number of reds and its standard deviation
问题 1：一个转盘有 25% 的机会停在红色区域。旋转 15 次，求：
(a) 恰好 5 次红色的概率
(b) 至少 3 次红色的概率
(c) 红色的期望次数及其标准差
Q2: A factory claims that at most 10% of its products are defective. A quality inspector tests 30 products and finds 5 defectives. Test at the 5% significance level whether the factory’s claim is valid. Also explain what a Type I error means in this context. (12 marks)
问题 2：一家工厂声称其产品次品率不超过 10%。质检员测试了 30 件产品，发现 5 件次品。以 5% 的显著性水平检验工厂的声称是否有效。同时解释在这种背景下第一类错误的含义。（12 分）
13. Summary and Key Takeaways / 总结与关键要点
Let’s distill everything into seven essential takeaways that will serve you in the exam hall:
让我们将所有内容浓缩为七个能帮你在考场中受益的关键要点：
- Check the four binomial conditions first — many questions start with “explain why this situation can be modelled by a binomial distribution” / 首先检查四个二项条件——很多题目以”解释为什么这种情况可以用二项分布建模”开头
- H₀ always has “=”, and you test using p from H₀, not the sample proportion / H₀ 总是带 “=”，使用 H₀ 中的 p 进行检验，不是样本比例
- One-tailed vs two-tailed depends on the wording of H₁, not on what the data shows / 单尾还是双尾取决于 H₁ 的措辞，而不是数据显示的内容
- For two-tailed tests, double the one-tailed p-value before comparing to α / 双尾检验中，将单尾 p 值加倍后再与 α 比较
- Always conclude in context — “reject H₀” alone gets zero marks for interpretation / 始终在上下文中下结论——仅仅写”拒绝 H₀”在解释分上得零分
- Type I error = false positive (rejecting true H₀), Type II error = false negative (not rejecting false H₀) / 第一类错误=假阳性（拒绝为真的 H₀），第二类错误=假阴性（未拒绝为假的 H₀）
- Show all working! Even with a calculator, write down the formula and the key steps — examiners award method marks generously / 展示所有过程！即使有计算器，也要写下公式和关键步骤——考官在方法分上给分慷慨
🎓 需要 A-Level 数学辅导？
无论你是在备战 Edexcel、AQA、OCR 还是 CIE A-Level 数学考试，我们经验丰富的导师都可以帮助你掌握二项分布、假设检验以及所有核心统计和纯数主题。
📞 16621398022 同微信
关注微信公众号 tutorhao 获取更多免费 A-Level 学习资源和备考指南。
访问 www.tutorhao.com/alevel/ 浏览完整课程目录。
🎓 Need A-Level Maths Tutoring?
Whether you’re preparing for Edexcel, AQA, OCR, or CIE A-Level Mathematics, our experienced tutors can help you master binomial distributions, hypothesis testing, and all core Statistics and Pure Mathematics topics.
📞 Contact: 16621398022 (WeChat)
Follow our WeChat Official Account tutorhao for more free A-Level study resources and exam guides.
Visit www.tutorhao.com/alevel/ to browse our full course catalog.
Discover more from tutorhao
Subscribe to get the latest posts sent to your email.
Categories: ALEVEL