IGCSE Maths: Hypothesis Testing Exam Focus | IGCSE 数学：假设检验考点精讲

📚 IGCSE Maths: Hypothesis Testing Exam Focus | IGCSE 数学：假设检验考点精讲

In IGCSE Mathematics, hypothesis testing is a powerful tool for making decisions based on data. It helps us assess whether observed outcomes provide enough evidence to support a particular claim about a population parameter. This topic is often examined through problems involving binomial probabilities, as students learn to formalise suspicions using null and alternative hypotheses, significance levels, and critical regions.

在IGCSE数学中，假设检验是一种基于数据做出决策的强大工具。它帮助我们评估观察到的结果是否提供了足够证据来支持关于总体参数的特定主张。这一考点常通过涉及二项式概率的问题来考查，学生需要学习用零假设与备择假设、显著性水平和临界区域来将猜想形式化。

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

Hypothesis testing is a statistical method used to decide whether there is enough evidence to reject a proposed statement about a population. In IGCSE, we often test claims about proportions, such as whether a coin is biased or whether a new production process has reduced the defect rate. The process begins by assuming that the current situation is true, then calculating how likely the observed sample results would be under that assumption.

假设检验是一种统计方法，用于判断是否有足够证据拒绝关于总体的某个提议陈述。在IGCSE中，我们经常检验关于比例的声称，比如一枚硬币是否被做了手脚，或者新生产工艺是否降低了次品率。整个过程从假设当前状况成立开始，然后计算在该假设下观察到样本结果的可能性有多大。

If the probability of obtaining results at least as extreme as ours is very small, we have strong evidence against the initial assumption and may reject it. This small probability is called the significance level. In IGCSE exams, typical significance levels are 5% (0.05) or 1% (0.01).

如果得到至少与我们观察到的结果一样极端的概率非常小，我们就有了反对初始假设的有力证据，可以拒绝它。这个很小的概率称为显著性水平。在IGCSE考试中，常见的显著性水平是5%（0.05）或1%（0.01）。

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

The null hypothesis, denoted H₀, is a statement of no change, no difference, or no effect. It is assumed to be true until evidence suggests otherwise. For example, if we suspect a coin is biased towards heads, we might write H₀: p = 0.5, where p is the probability of getting heads.

零假设，记作H₀，是一个表示无变化、无差异或无效果的陈述。它被假定为真，直到有证据表明并非如此。例如，如果我们怀疑一枚硬币偏向于正面，我们可以写H₀：p = 0.5，其中p是得到正面的概率。

The alternative hypothesis, denoted H₁ or Hₐ, represents what we are trying to prove. It states that there is a change, a difference, or an effect. For the coin example, the alternative could be one-sided: H₁: p > 0.5 (coin biased towards heads) or two-sided: H₁: p ≠ 0.5 (coin is not fair). Choosing the correct H₁ is crucial for the test.

备择假设，记作H₁或Hₐ，代表我们试图证明的内容。它表明存在变化、差异或效应。对于硬币的例子，备择假设可以是单侧的：H₁：p > 0.5（硬币偏向正面），或者双侧的：H₁：p ≠ 0.5（硬币不公平）。选择正确的H₁对于检验至关重要。

3. Significance Level | 显著性水平

The significance level, often written as α (alpha), is the threshold probability we set for deciding whether a result is statistically significant. It represents the maximum probability of making a Type I error — rejecting a true null hypothesis. In IGCSE, we usually work with α = 0.05 (5%) or α = 0.01 (1%).

显著性水平，常写作α（阿尔法），是我们设定的用于判断结果是否具有统计显著性的概率阈值。它表示犯第一类错误——拒绝一个真实的零假设——的最大概率。在IGCSE中，我们通常使用α = 0.05（5%）或α = 0.01（1%）。

If the computed probability (the p-value) from our sample is less than α, we reject H₀. If it is greater than or equal to α, we do not reject H₀. For example, at the 5% level, if a result has only a 3% chance of occurring by random variation, we consider it significant.

如果从样本计算出的概率（p值）小于α，我们拒绝H₀。如果它大于或等于α，我们不拒绝H₀。例如，在5%的水平下，如果某个结果仅有3%的可能性是随机变化造成的，我们就认为它是显著的。

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 of change, such as H₁: p > 0.5 or H₁: p < 0.3. The entire significance level α is placed in one tail of the distribution. This makes it easier to reject H₀ for results in the specified direction.

当备择假设指明了变化方向时，例如H₁：p > 0.5或H₁：p < 0.3，我们使用单尾检验。整个显著性水平α被放在分布的单个尾部。这使我们更容易拒绝指定方向上的H₀。

A two-tailed test is used when H₁ simply states that the parameter is different from the hypothesised value (e.g. H₁: p ≠ 0.5). In this case, α is split equally between the two tails, typically α/2 in each tail. This means we require more extreme results in either direction to reject H₀.

当备择假设仅仅声明参数不同于假设值（例如H₁：p ≠ 0.5）时，使用双尾检验。此时，α被平均分配到两个尾部，通常每个尾部为α/2。这意味着我们需要在任一方向上得到更极端的结果才能拒绝H₀。

In IGCSE exam questions, carefully read the wording: ‘more than’, ‘increased’, ‘higher’ suggest a one-tailed test; ‘different from’, ‘changed’ suggest a two-tailed test.

在IGCSE考试题目中，仔细审题：“多于”、“增加”、“更高”暗示单尾检验；“不同于”、“发生了变化”暗示双尾检验。

5. Critical Regions and Critical Values | 临界区域与临界值

The critical region (or rejection region) is the set of values of the test statistic for which we reject H₀. The boundary values that separate the critical region from the acceptance region are called critical values. For a binomial test with n trials and H₀: p = p₀, we find the smallest number of successes that would lead to rejection, given α.

临界区域（或拒绝域）是检验统计量的取值集合，当统计量落入该区域时我们拒绝H₀。将临界区域与接受区域分开的边界值称为临界值。对于试验次数为n且H₀：p = p₀的二项检验，我们找到在给定α下会导致拒绝的最小的成功次数。

For a one-tailed test H₁: p > p₀, we find the smallest integer c such that P(X ≥ c) ≤ α. Then the critical region is X ≥ c. For H₁: p < p₀, we find the largest integer c such that P(X ≤ c) ≤ α, and the critical region is X ≤ c. In a two-tailed test, we halve α and find critical values at both ends.

对于单尾检验H₁：p > p₀，我们找到满足P(X ≥ c) ≤ α的最小整数c，此时临界区域为X ≥ c。对于H₁：p < p₀，我们找到满足P(X ≤ c) ≤ α的最大整数c，临界区域为X ≤ c。在双尾检验中，我们将α减半并在两端寻找临界值。

6. The p-Value Approach | p值方法

The p-value is the probability of obtaining a test result at least as extreme as the one observed, assuming H₀ is true. A small p-value (typically less than α) indicates that such an extreme result is unlikely under H₀, so we reject H₀. A large p-value suggests the observed result is consistent with H₀, so we do not reject it.

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检验结果的概率。很小的p值（通常小于α）表明在H₀下出现这种极端结果不太可能，因此我们拒绝H₀。很大的p值则表明观测结果与H₀一致，因此我们不拒绝它。

In binomial hypothesis testing, the p-value can be calculated directly using the binomial distribution. For example, if we observed 12 heads in 20 tosses and H₁: p > 0.5, the p-value is P(X ≥ 12) where X ~ B(20, 0.5). If this p-value is less than α, we reject H₀.

在二项假设检验中，可以直接利用二项分布计算p值。例如，如果我们在20次抛掷中观察到12次正面且H₁：p > 0.5，则p值为P(X ≥ 12)，其中X ~ B(20, 0.5)。如果这个p值小于α，我们就拒绝H₀。

7. Hypothesis Testing with Binomial Distribution | 二项分布的假设检验

IGCSE hypothesis testing questions almost always involve a binomial random variable. We define X as the number of successes in n independent trials, with probability p of success on each trial. Under H₀, p is set to a specific value. We then determine whether the observed number of successes falls inside the critical region.

IGCSE的假设检验问题几乎总是涉及二项随机变量。我们将X定义为n次独立试验中成功的次数，每次试验成功的概率为p。在H₀下，p被设定为一个特定值。然后我们判断观测到的成功次数是否落入临界区域。

To perform the test, we can use either the critical region method or the p-value method. Tables of cumulative binomial probabilities or a calculator’s binomCdf function are often used to compute P(X ≤ k) or P(X ≥ k). The formula P(X ≥ k) = 1 – P(X ≤ k – 1) is essential.

进行检验时，我们可以使用临界区域法或p值法。通常使用累积二项概率表或计算器的binomCdf功能来计算P(X ≤ k)或P(X ≥ k)。公式P(X ≥ k) = 1 – P(X ≤ k – 1)是必不可少的。

8. Worked Example: One-tailed Test | 实例分析：单尾检验

A manufacturer claims that at most 5% of their light bulbs are defective. A customer suspects the defect rate is higher and tests a random sample of 30 bulbs, finding 4 defective ones. Test at the 5% significance level whether the customer’s suspicion is supported.

某制造商声称其灯泡的次品率最多为5%。一位顾客怀疑次品率更高，并随机抽样了30个灯泡，发现4个是次品。在5%的显著性水平下检验该顾客的怀疑是否有依据。

Let X be the number of defective bulbs. H₀: p = 0.05, H₁: p > 0.05 (one-tailed). Under H₀, X ~ B(30, 0.05). We need to find the p-value: P(X ≥ 4) = 1 – P(X ≤ 3). Using binomial tables or calculation, P(X ≤ 3) ≈ 0.9392.

设X为次品灯泡的数量。H₀：p = 0.05，H₁：p > 0.05（单尾）。在H₀下，X ~ B(30, 0.05)。我们需要求p值：P(X ≥ 4) = 1 – P(X ≤ 3)。使用二项分布表或计算，P(X ≤ 3) ≈ 0.9392。

p-value = 1 – 0.9392 = 0.0608

Since p-value = 0.0608 > 0.05, we do not reject H₀. There is insufficient evidence at the 5% level to support the claim that the defect rate is higher than 5%.

由于p值 = 0.0608 > 0.05，我们不拒绝H₀。在5%的水平下，没有足够证据支持次品率高于5%的说法。

Alternatively, using the critical value method: we need the smallest c with P(X ≥ c) ≤ 0.05. P(X ≥ 5) = 1 – P(X ≤ 4) ≈ 1 – 0.9818 = 0.0182 ≤ 0.05, so critical region is X ≥ 5. Since 4 is not in the critical region, we do not reject H₀.

或者使用临界值法：我们需要满足P(X ≥ c) ≤ 0.05的最小c。P(X ≥ 5) = 1 – P(X ≤ 4) ≈ 1 – 0.9818 = 0.0182 ≤ 0.05，因此临界区域为X ≥ 5。因为4不在临界区域内，我们不拒绝H₀。

9. Worked Example: Two-tailed Test | 实例分析：双尾检验

A student claims that a coin is biased. In 50 tosses, the coin lands on heads 31 times. Test at the 5% significance level whether the coin is fair.

一位学生声称一枚硬币有偏。在50次抛掷中，硬币正面朝上31次。在5%的显著性水平下检验该硬币是否公平。

Let X be the number of heads. H₀: p = 0.5, H₁: p ≠ 0.5 (two-tailed). Under H₀, X ~ B(50, 0.5). For a two-tailed test at α = 0.05, we split α into 0.025 in each tail. We need P(X ≤ c₁) ≤ 0.025 and P(X ≥ c₂) ≤ 0.025.

设X为正面朝上的次数。H₀：p = 0.5，H₁：p ≠ 0.5（双尾）。在H₀下，X ~ B(50, 0.5)。对于α = 0.05的双尾检验，我们把α分为每尾0.025。我们需要P(X ≤ c₁) ≤ 0.025和P(X ≥ c₂) ≤ 0.025。

Using symmetry of B(50, 0.5), find the lower critical value c₁ such that P(X ≤ c₁) is as large as possible while ≤ 0.025. With cumulative tables: P(X ≤ 17) ≈ 0.0164 (≤ 0.025), P(X ≤ 18) ≈ 0.0325 (> 0.025). So c₁ = 17. By symmetry, the upper critical value c₂ = 50 – 17 = 33. Critical region: X ≤ 17 or X ≥ 33.

利用B(50, 0.5)的对称性，找到下临界值c₁，使得P(X ≤ c₁)尽可能大同时≤ 0.025。查累积表：P(X ≤ 17) ≈ 0.0164 (≤ 0.025)，P(X ≤ 18) ≈ 0.0325 (> 0.025)。所以c₁ = 17。由对称性，上临界值c₂ = 50 – 17 = 33。临界区域：X ≤ 17 或 X ≥ 33。

The observed value 31 is not in the critical region. Therefore, we do not reject H₀. There is insufficient evidence at the 5% level to conclude the coin is biased.

观测值31不在临界区域内。因此，我们不拒绝H₀。在5%的水平下，没有足够证据断定硬币有偏。

10. Interpreting Conclusions | 结论解释

After conducting a hypothesis test, you must write a clear conclusion in the context of the problem. Do not simply say ‘reject H₀’ or ‘do not reject H₀’. Instead, state whether there is sufficient evidence to support the alternative claim. Use phrases like ‘there is sufficient evidence at the α level to suggest that…’ or ‘the evidence does not support the claim that…’.

进行假设检验后，你必须结合问题背景写出清晰的结论。不要简单地说“拒绝H₀”或“不拒绝H₀”。相反，要说明是否有足够证据支持备择主张。使用诸如“在α水平下有足够证据表明……”或“证据不支持……的说法”之类的表述。

Remember that not rejecting H₀ does not prove H₀ is true; it only means the sample data are not inconsistent with H₀. A higher sample size or a different significance level could lead to a different conclusion. Also, be careful to match the conclusion to the alternative hypothesis: if H₁ was one-tailed, the conclusion must reflect the direction.

请记住，不拒绝H₀并不证明H₀为真；它只意味着样本数据与H₀没有不一致。更大的样本量或不同的显著性水平可能导致不同的结论。此外，要确保结论与备择假设匹配：如果H₁是单尾的，结论必须反映该方向。

11. Common Mistakes and Tips | 常见错误与技巧

One common mistake is forgetting to adjust the significance level for a two-tailed test. In such cases, always halve α when finding critical values or comparing p-values. Another pitfall is using the wrong inequality: for H₁: p > p₀, we need P(X ≥ observed); for H₁: p < p₀, we need P(X ≤ observed).

一个常见错误是忘记在双尾检验中调整显著性水平。在这种情况下，寻找临界值或比较p值时始终要将α减半。另一个陷阱是使用错误的不等式：对于H₁：p > p₀，我们需要P(X ≥ 观测值)；对于H₁：p < p₀，我们需要P(X ≤ 观测值)。

Students sometimes confuse the p-value with the significance level. The p-value is computed from the data, while α is chosen before the test. Only reject H₀ if p-value < α. Also, never state 'accept H₀' — the correct phrasing is 'do not reject H₀' or 'retain H₀'.

学生有时会混淆p值与显著性水平。p值是根据数据计算得出的，而α是在检验前选择的。只有在p值 < α时才拒绝H₀。此外，永远不要说“接受H₀”——正确的表述是“不拒绝H₀”或“保留H₀”。

When using binomial tables, check whether they give cumulative probabilities P(X ≤ k) or individual probabilities P(X = k). For P(X ≥ k), use the complement rule precisely: P(X ≥ k) = 1 – P(X ≤ k – 1).

使用二项分布表时，要检查表格给出的是累积概率P(X ≤ k)还是单点概率P(X = k)。对于P(X ≥ k)，精确使用补集规则：P(X ≥ k) = 1 – P(X ≤ k – 1)。

12. Exam Practice Summary | 考试练习总结

In IGCSE exams, hypothesis testing questions usually carry high marks because they require a structured approach. Always follow these steps: (1) Define the random variable and state the hypotheses clearly. (2) Write down the distribution under H₀. (3) Determine whether the test is one-tailed or two-tailed and identify the significance level. (4) Calculate either the p-value or the critical region. (5) Compare the test statistic with the critical value or the p-value with α. (6) Write a contextual conclusion.

在IGCSE考试中，假设检验题目通常分值较高，因为它们需要有条理的方法。始终遵循以下步骤：(1) 定义随机变量并清晰陈述假设。(2) 写出H₀下的分布。(3) 判断检验是单尾还是双尾，并确定显著性水平。(4) 计算p值或确定临界区域。(5) 将检验统计量与临界值比较，或将p值与α比较。(6) 写出结合背景的结论。

Practice with past paper questions is essential to become comfortable with interpreting word problems and using binomial probability tables efficiently. Remember that the logic of hypothesis testing remains the same whether you are testing a coin, a die, or a manufacturing process.

通过历年真题进行练习至关重要，这能帮助你熟练解读文字题并高效使用二项概率表。记住，无论检验的是硬币、骰子还是生产工艺，假设检验的逻辑始终保持不变。

Published by TutorHao | Mathematics Revision Series | aleveler.com

更多咨询请联系16621398022（同微信）

IGCSE Maths: Hypothesis Testing Exam Focus | IGCSE 数学：假设检验 考点精讲