Central Limit Theorem and Its Applications | 中心极限定理及其应用

📚 Central Limit Theorem and Its Applications | 中心极限定理及其应用

The Central Limit Theorem (CLT) is one of the most powerful and elegant results in probability and statistics. It explains why the normal distribution appears so frequently in nature and provides the theoretical foundation for statistical inference. In IB Mathematics, the CLT is a core concept in both Analysis and Approaches (AA) and Applications and Interpretation (AI), and it appears in exam questions on sampling distributions, confidence intervals, and hypothesis testing.

中心极限定理（CLT）是概率论与统计学中最强大、最优雅的定理之一。它解释了正态分布为何在自然界中如此频繁地出现，并为统计推断提供了理论基础。在IB数学中，CLT在分析与方法（AA）和应用与解释（AI）中都是核心概念，并出现在抽样分布、置信区间和假设检验的考题之中。

1. What Is the Central Limit Theorem? | 什么是中心极限定理？

The Central Limit Theorem states that if you take a sufficiently large random sample from any population with mean μ and finite variance σ², the sampling distribution of the sample mean (X̄) will be approximately normally distributed, regardless of the shape of the original population distribution.

中心极限定理指出：如果从均值为 μ、方差 σ² 有限的任意总体中抽取足够大的随机样本，那么样本均值 X̄ 的抽样分布将近似服从正态分布，无论原始总体分布的形状如何。

More precisely, for a sample of size n, the standardized sample mean follows:

更精确地说，对于容量为 n 的样本，标准化后的样本均值服从：

Z = (X̄ − μ) / (σ / √n) → N(0, 1) as n → ∞

In plain language: the sample mean X̄ is approximately normally distributed with mean μ and standard deviation σ/√n.

用通俗的话说：样本均值 X̄ 近似服从均值为 μ、标准差为 σ/√n 的正态分布。

2. Conditions for the CLT | 中心极限定理的适用条件

While the CLT is remarkably robust, certain conditions must be satisfied for it to apply correctly in practice.

尽管中心极限定理极其稳健，但在实践中正确应用仍需满足某些条件。

Random sampling: The sample must be collected randomly from the population. A biased or convenience sample invalidates the CLT.

随机抽样：样本必须从总体中随机选取。有偏抽样或便利抽样会使CLT失效。
Independence: Each observation must be independent of the others. In practice, this requires that the sample size n be less than 10% of the population when sampling without replacement.

独立性：每个观测值之间必须相互独立。实际操作中，当无放回抽样时，样本容量 n 应小于总体规模的10%。
Sample size: A common rule of thumb is n ≥ 30 for the CLT to provide a good approximation. However, if the population distribution is heavily skewed or has extreme outliers, a larger sample may be needed.

样本容量：经验法则为 n ≥ 30 时CLT能提供良好近似。然而，如果总体分布严重偏斜或存在极端异常值，则可能需要更大的样本。
Finite variance: The population must have a finite variance. This condition is almost always met in IB-level problems.

方差有限：总体必须具有有限方差。在IB级别的问题中，这一条件几乎总是满足的。

Additionally, when the population is already normally distributed, X̄ is exactly normally distributed for any sample size n, not just approximately.

此外，当总体本身服从正态分布时，X̄ 对任意样本容量 n 都精确服从正态分布，而不仅仅是近似。

3. The Sampling Distribution of the Sample Mean | 样本均值的抽样分布

A key consequence of the CLT is that we can fully describe the sampling distribution of X̄ using only three facts.

CLT的一个关键推论是，我们仅需三个事实就能完整描述 X̄ 的抽样分布。

Property	Value
Mean of X̄	E(X̄) = μ
Variance of X̄	Var(X̄) = σ²/n
Standard deviation of X̄ (standard error)	σ/√n

The standard deviation of the sample mean, σ/√n, is called the standard error of the mean. Notice that as n increases, the standard error decreases — the sample mean becomes a more precise estimator of μ.

样本均值的标准差 σ/√n 被称为均值的标准误。注意随着 n 增大，标准误减小——样本均值作为 μ 的估计量变得更加精确。

For example, if the population standard deviation is 10 and you take a sample of size 25, the standard error is 10/√25 = 2. If you increase the sample to 100, the standard error drops to 10/√100 = 1. This square-root relationship explains why doubling the sample size only reduces the standard error by a factor of √2 ≈ 1.41, not by half.

例如，若总体标准差为10，取容量25的样本，则标准误为 10/√25 = 2。若将样本增大到100，标准误降至 10/√100 = 1。这种平方根关系解释了为何将样本量加倍仅能使标准误缩小 √2 ≈ 1.41 倍，而非一半。

4. Standardising the Sample Mean | 样本均值的标准化

To use the CLT in probability calculations, we convert the sample mean to a z-score using the standard error.

为了在概率计算中使用CLT，我们通过标准误将样本均值转化为 z 分数。

Z = (X̄ − μ) / (σ / √n)

Once standardised, we can use the standard normal distribution table or a GDC (graphical display calculator) to find probabilities.

标准化后，我们就可以使用标准正态分布表或图形计算器（GDC）求概率。

Worked example: A machine fills bags of rice with mean weight 1000 g and standard deviation 20 g. A random sample of 50 bags is selected. Find the probability that the sample mean weight exceeds 1005 g.

例题：一台机器灌装大米，袋均重量为1000克，标准差为20克。随机抽取50袋，求样本均值超过1005克的概率。

Solution: Here μ = 1000, σ = 20, n = 50. The standard error is 20/√50 ≈ 2.828. We want:

解：这里 μ = 1000，σ = 20，n = 50。标准误为 20/√50 ≈ 2.828。我们需要求：

P(X̄ > 1005) = P(Z > (1005 − 1000) / 2.828) = P(Z > 1.77)

From the standard normal table, P(Z > 1.77) = 1 − 0.9616 = 0.0384. So the probability is approximately 3.84%.

查标准正态分布表，P(Z > 1.77) = 1 − 0.9616 = 0.0384。因此概率约为3.84%。

5. The Continuity Correction | 连续性修正

When the CLT is applied to data from a discrete distribution — such as counts obtained from the binomial or Poisson distribution — a continuity correction improves the accuracy of the normal approximation.

当CLT应用于离散分布的数据时——例如来自二项分布或泊松分布的计数——连续性修正可以提高正态近似的精确度。

The idea is that a discrete value x should be treated as covering the interval from x − 0.5 to x + 0.5. For example:

其思想是，离散取值 x 应被视为覆盖从 x − 0.5 到 x + 0.5 的区间。例如：

P(X = k) is approximated by P(k − 0.5 < X < k + 0.5), using the normal distribution.

P(X = k) 由正态分布近似为 P(k − 0.5 < X < k + 0.5)。
P(X ≤ k) becomes P(X < k + 0.5).

P(X ≤ k) 变为 P(X < k + 0.5)
P(X ≥ k) becomes P(X > k − 0.5).

P(X ≥ k) 变为 P(X > k − 0.5)

In IB examinations, the continuity correction is required when approximating a binomial distribution with a normal distribution, which is acceptable when n is large, p is not too close to 0 or 1, and np ≥ 5 and n(1 − p) ≥ 5.

在IB考试中，当用正态分布近似二项分布时需要连续性修正。这种近似在 n 较大、p 不太接近0或1且 np ≥ 5、n(1 − p) ≥ 5 时是可接受的。

Worked example: A coin is tossed 80 times. Use a normal approximation with continuity correction to estimate the probability of getting between 35 and 45 heads inclusive.

例题：一枚硬币抛掷80次。使用带连续性修正的正态近似，估计正面朝上的次数在35到45之间（含端点）的概率。

Solution: For X ~ B(80, 0.5), we have μ = np = 40 and σ = √(np(1−p)) = √(80 × 0.5 × 0.5) = √20 ≈ 4.472.

解：对 X ~ B(80, 0.5)，我们有 μ = np = 40，σ = √(np(1−p)) = √(80 × 0.5 × 0.5) = √20 ≈ 4.472。

P(35 ≤ X ≤ 45) ≈ P(34.5 < X < 45.5)

= P((34.5 − 40)/4.472 < Z < (45.5 − 40)/4.472) = P(−1.23 < Z < 1.23)

From the table, P(−1.23 < Z < 1.23) = 2(0.8907 − 0.5) = 0.7814, or approximately 78.1%.

查表得 P(−1.23 < Z < 1.23) = 2(0.8907 − 0.5) = 0.7814，约为78.1%。

6. Applying the CLT to the Sum of Observations | 中心极限定理在观测值求和中的应用

The CLT also applies to the sum of sample observations, T = X₁ + X₂ + … + Xₙ. In this case:

CLT同样适用于样本观测值的总和 T = X₁ + X₂ + … + Xₙ。在这种情况下：

T ≈ N(nμ, nσ²)

That is, the sum follows an approximately normal distribution with mean nμ and standard deviation σ√n.

即总和近似服从均值为 nμ、标准差为 σ√n 的正态分布。

Worked example: The heights of adult males are normally distributed with mean 175 cm and standard deviation 8 cm. A random sample of 20 adult males is selected. Find the probability that the total height of the sample exceeds 3560 cm.

例题：成年男性身高服从均值为175厘米、标准差为8厘米的正态分布。随机抽取20名成年男性，求样本总身高超过3560厘米的概率。

Solution: Here n = 20, so T ≈ N(20 × 175, 20 × 8²) = N(3500, 1280). The standard deviation of T is √1280 ≈ 35.78.

解：这里 n = 20，所以 T ≈ N(20 × 175, 20 × 8²) = N(3500, 1280)。T 的标准差为 √1280 ≈ 35.78。

P(T > 3560) = P(Z > (3560 − 3500)/35.78) = P(Z > 1.68) ≈ 0.0465

A common mistake is to confuse the standard deviation of the total (σ√n) with the standard error of the mean (σ/√n). Always ask yourself whether the question refers to an average or a total.

一个常见错误是混淆总体的标准差 σ√n 与均值标准误 σ/√n。解题时务必问自己：题目问的是平均值还是总和？

7. Confidence Intervals for the Population Mean | 总体均值的置信区间

The CLT provides the foundation for constructing confidence intervals. A confidence interval gives a range of plausible values for the population mean μ based on sample data.

CLT为构造置信区间提供了理论基础。置信区间基于样本数据给出总体均值 μ 的一个合理取值范围。

For a population with known variance σ², a (1 − α)100% confidence interval for μ is:

对于方差 σ² 已知的总体，μ 的 (1 − α)100% 置信区间为：

X̄ ± z* × (σ / √n)

where z* is the critical value from the standard normal distribution. Common values are z* = 1.645 for 90% confidence, z* = 1.960 for 95%, and z* = 2.576 for 99%.

其中 z* 是来自标准正态分布的临界值。常见取值：90%置信水平时 z* = 1.645，95%时 z* = 1.960，99%时 z* = 2.576。

The width of the confidence interval depends on three factors: the confidence level, the population standard deviation, and the sample size. Increasing n reduces the width, which is why larger samples give more precise estimates.

置信区间的宽度取决于三个因素：置信水平、总体标准差和样本量。增大 n 会减小宽度，这就是为什么更大的样本能给出更精确的估计。

Worked example: The weights of 64 randomly selected oranges from a large orchard have a sample mean of 180 g. Given that the population standard deviation is 25 g, find a 95% confidence interval for the population mean weight.

例题：从一个大果园随机选取64个橙子，样本均重为180克。已知总体标准差为25克，求总体均重的95%置信区间。

Solution: The standard error is 25/√64 = 3.125. The 95% confidence interval is:

解：标准误为 25/√64 = 3.125。95%置信区间为：

180 ± 1.96 × 3.125 = (173.875, 186.125)

We can say with 95% confidence that the population mean weight lies between 173.9 g and 186.1 g.

我们可以说，在95%的置信水平下，总体均重位于173.9克到186.1克之间。

8. Hypothesis Testing Based on the CLT | 基于中心极限定理的假设检验

The CLT also underpins hypothesis testing for a population mean. The z-test for a single mean uses the same standardised statistic:

CLT同样是总体均值假设检验的基础。单均值 z 检验使用相同的标准化统计量：

z = (X̄ − μ₀) / (σ / √n)

where μ₀ is the hypothesised population mean under the null hypothesis H₀.

其中 μ₀ 是原假设 H₀ 下的总体均值假设值。

In IB Mathematics AI, students are expected to set up null and alternative hypotheses, calculate the test statistic, compute the p-value using a GDC or formula, and draw a conclusion by comparing the p-value to the significance level α.

在IB数学AI中，学生需要设立原假设与备择假设、计算检验统计量、使用GDC或公式计算 p 值，并通过比较 p 值与显著性水平 α 得出结论。

Worked example: A school claims that its students spend an average of 2 hours per day on homework. A random sample of 100 students shows a mean of 2.15 hours. The population standard deviation is known to be 0.8 hours. Test at the 5% significance level whether the true mean is greater than 2 hours.

例题：某学校声称其学生每天平均花费2小时做作业。随机抽取100名学生，样本均值为2.15小时。已知总体标准差为0.8小时。在5%显著性水平下检验真实均值是否大于2小时。

H₀: μ = 2 and H₁: μ > 2 (one-tailed test).

H₀：μ = 2，H₁：μ > 2（单尾检验）。
The test statistic is z = (2.15 − 2) / (0.8/√100) = 0.15 / 0.08 = 1.875.

检验统计量为 z = (2.15 − 2) / (0.8/√100) = 0.15 / 0.08 = 1.875。
The critical value for a one-tailed test at α = 0.05 is z* = 1.645. Since 1.875 > 1.645, we reject H₀.

α = 0.05 单尾检验的临界值为 z* = 1.645。因为 1.875 > 1.645，我们拒绝 H₀。

There is sufficient evidence to support the claim that the true mean homework time exceeds 2 hours per day.

有充分证据支持真实均值每日作业时间超过2小时的说法。

9. Common Pitfalls in IB Exams | IB考试中的常见误区

Students frequently lose marks on CLT questions due to a few recurring errors. Being aware of these pitfalls will help you avoid them.

学生在CLT题目中常因几个反复出现的错误而失分。意识到这些陷阱将帮助你避免它们。

Using σ instead of σ/√n: When the question asks about a sample mean, you must use the standard error σ/√n, not the population standard deviation σ. This is the single most common error.

使用 σ 而非 σ/√n：当题目问的是样本均值时，必须使用标准误 σ/√n，而不是总体标准差 σ。这是最常见的错误。
Forgetting to check conditions: The CLT requires n ≥ 30 (or a normal population). If the question provides a small sample from a skew分布, the normal approximation is not valid.

忘记检查条件：CLT要求 n ≥ 30（或总体为正态分布）。如果题目给出的是来自偏斜分布的小样本，正态近似不成立。
Omitting the continuity correction: When approximating a binomial distribution, always apply ±0.5 corrections.

省略连续性修正：用正态分布近似二项分布时，总是应用 ±0.5 修正。
Confusing sample mean and sample total: The sample total T has standard deviation σ√n, while the sample mean X̄ has standard deviation σ/√n.

混淆样本均值与样本总和：样本总和 T 的标准差为 σ√n，而样本均值 X̄ 的标准差为 σ/√n。
Wrong direction in probability: For P(X̄ > a), standardise as (a − μ)/(σ/√n); for P(X̄ < a), use the same z-score but find the area to the left. Many students mix up greater-than and less-than probabilities.

概率方向错误：对于 P(X̄ > a)，标准化为 (a − μ)/(σ/√n)；对于 P(X̄ < a)，使用相同 z 分数但求左侧面积。许多学生混淆大于和小于的概率。

10. When the Sample Size Is Small | 当样本量较小时

If n is small (typically n < 30) and the population distribution is not normal, the CLT cannot be used. In this case, other techniques are required.

如果 n 较小（通常 n < 30）且总体分布不是正态的，则不能使用CLT。此时需要使用其他方法。

However, if the population itself is normally distributed, the sample mean X̄ is exactly normally distributed with mean μ and variance σ²/n for any n, no matter how small. In IB standard level (SL) papers, this situation is common.

然而，如果总体本身服从正态分布，那么无论 n 多小，样本均值 X̄ 都精确服从均值为 μ、方差为 σ²/n 的正态分布。在IB标准水平（SL）考试中，这种情况很常见。

When σ is unknown, the t-distribution is used instead of the normal distribution. IB Mathematics HL students are expected to know that the t-distribution approaches the normal distribution as the degrees of freedom increase. For a random sample X₁, …, Xₙ from a normal distribution, the statistic:

当 σ 未知时，应使用 t 分布而非正态分布。IB数学HL学生需要知道，随着自由度增加，t 分布趋近于正态分布。对于来自正态分布的随机样本 X₁, …, Xₙ，统计量：

t = (X̄ − μ) / (s / √n)

follows a t-distribution with n − 1 degrees of freedom, where s is the sample standard deviation.

服从自由度为 n − 1 的 t 分布，其中 s 是样本标准差。

11. Real-World Applications of the CLT | 中心极限定理的实际应用

Beyond the classroom, the CLT plays a crucial role in many real-world contexts.

在课堂之外，CLT在众多现实场景中发挥着至关重要的作用。

Election polling: Opinion polls use CLT to estimate the proportion of voters supporting a candidate and to calculate the margin of error.

选举民调：民意调查利用CLT估计支持某候选人的选民比例，并计算误差范围。
Quality control: Manufacturers use control charts based on the CLT to monitor production processes. Every few minutes, they sample a few items and plot the sample mean; if the mean falls outside control limits, the process may be out of control.

质量控制：制造商使用基于CLT的控制图监控生产流程。每隔几分钟抽取少量产品并绘制样本均值；若均值落在控制限之外，则可能表示过程失控。
Insurance and finance: Actuaries use the CLT to model the total claims from many independent policies, allowing them to set premium levels that keep the probability of insolvency very low.

保险与金融：精算师使用CLT对众多独立保单的总索赔进行建模，从而设定保费水平，使破产概率保持在极低水平。
Genetics: The CLT explains why many biological measurements (height, blood pressure, etc.) follow approximately normal distributions, even if the underlying genetic and environmental factors are not themselves normal.

遗传学：CLT解释了为何许多生物测量值（身高、血压等）近似服从正态分布，即使底层遗传和环境因素本身不服从正态分布。

In each of these applications, the key insight is the same: averages from sufficiently large samples tell us something reliable about the population, no matter how unusual the underlying distribution may be.

在这些应用中，核心思想都是相同的：足够大样本的平均值能告诉我们关于总体的可靠信息，无论背后的分布多么不寻常。

12. Summary and Exam Strategies | 总结与应试策略

The Central Limit Theorem is a bridge between probability and statistics. It says that if we take a large enough random sample, the sample mean is approximately normal — even if the population is far from normal. This single result justifies the use of normal distribution calculations in countless practical settings.

中心极限定理是连接概率与统计的桥梁。它表明，只要抽取足够大的随机样本，样本均值就近似服从正态分布——即使总体远远偏离正态。这一结果使得正态分布计算在无数实际场景中得以应用。

Key formulas to remember:

需要牢记的关键公式：

X̄ ≈ N(μ, σ²/n) and Z = (X̄ − μ) / (σ/√n)

Total: T ≈ N(nμ, nσ²)

Confidence interval: X̄ ± z*(σ/√n)

For IB exams, follow this systematic approach to CLT problems:

对于IB考试，遵循以下系统化的步骤解决CLT问题：

Identify: Is the question about a single value, a sample mean, or a sample total? This determines which formula you use.

识别：题目问的是单个值、样本均值还是样本总和？这决定了使用哪个公式。
Check: Is n large enough? Is the population normal? Are the observations independent?

检查：n 是否足够大？总体是否正态？观测是否独立？
Standardise: Convert the relevant statistic to a z-score on the standard normal distribution.

标准化：将相关统计量转化为标准正态分布上的 z 分数。
Calculate: Use proper notation and show each step. In IB, method marks are often awarded even if the final answer is incorrect.

计算：使用规范的符号，逐步展示。在IB考试中，即使最终答案有误，步骤分通常仍然有效。
Interpret: Always link your answer back to the context of the question in the conclusion.

解释：始终在结论中将答案与题目情境联系起来。

The CLT is not merely a theoretical result — it is the practical engine behind many of the statistics you will encounter in IB Mathematics and beyond. Master it, and you will find the statistics component of your examinations considerably more approachable.

CLT不仅是理论成果，更是你在IB数学及更高级统计中所遇到的许多方法的实际操作引擎。掌握它，你会发现考试中的统计部分变得和蔼可亲得多。

Published by TutorHao | Mathematics Revision Series | aleveler.com

更多咨询请联系16621398022（同微信）

Central Limit Theorem and Its Applications | 中心极限定理及其应用