The Central Limit Theorem | 中心极限定理

📚 The Central Limit Theorem | 中心极限定理

In A-Level statistics, one of the most powerful ideas is that averages tend to follow a normal distribution, even when the original data do not. This result is called the Central Limit Theorem (CLT). It underpins confidence intervals, hypothesis tests, and many real-world decisions made from sample data.

在 A-Level 统计学中，一个最强大的思想是：即使原始数据不服从正态分布，其平均值往往也近似服从正态分布。这个结论称为中心极限定理（CLT）。它是置信区间、假设检验以及许多基于样本数据的实际决策的理论基础。

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 sufficiently large random samples from any population with a finite mean μ and finite variance σ², then the distribution of the sample means will be approximately normal. This is true regardless of the shape of the original population distribution.

中心极限定理指出，如果从任何具有有限均值 μ 和有限方差 σ² 的总体中抽取足够大的随机样本，那么样本均值的分布将近似服从正态分布。无论原始总体分布的形状如何，这个结论都成立。

X̄ ≈ N(μ, σ²/n)

X̄ 近似服从 N(μ, σ²/n)

This means that the sample mean, X̄, has a normal distribution with the same centre as the population mean μ, but with a much smaller spread, especially for large samples.

这意味着样本均值 X̄ 服从正态分布，其中心与总体均值 μ 相同，但离散程度小得多，尤其是当样本量较大时。

2. Formal Statement of the CLT | 中心极限定理的正式表述

Let X₁, X₂, …, Xₙ be independent and identically distributed random variables with mean μ and variance σ². Define the sample mean X̄ = (X₁ + X₂ + … + Xₙ)/n. Then, as n increases, the standardised variable converges in distribution to the standard normal distribution N(0, 1).

设 X₁, X₂, …, Xₙ 是独立同分布的随机变量，均值为 μ，方差为 σ²。定义样本均值 X̄ = (X₁ + X₂ + … + Xₙ)/n。随着 n 增大，标准化变量的分布收敛于标准正态分布 N(0, 1)。

Z = (X̄ – μ) / (σ/√n) → N(0, 1) approximately

Z = (X̄ – μ) / (σ/√n) 近似服从 N(0, 1)

The symbol → denotes convergence in distribution. In practice, we use this result as an approximation for finite but large n.

符号 → 表示依分布收敛。在实际应用中，对于有限但较大的 n，我们将这一结果作为近似使用。

3. Key Conditions for the CLT | 中心极限定理的关键条件

The CLT requires random sampling, independence of observations, and a sufficiently large sample size. The population mean and variance must be finite. As a rule of thumb, n ≥ 30 is often considered large enough, but if the population is very skewed, a larger sample may be needed.

中心极限定理要求随机抽样、观测值独立，以及样本量足够大。总体均值和方差必须是有限的。经验法则通常认为 n ≥ 30 已经足够大，但如果总体分布严重偏斜，可能需要更大的样本。

Random sampling: every individual has an equal chance of selection.
随机抽样：每个个体被选中的机会相等。
Independence: one observation does not affect another.
独立性：一个观测值不会影响另一个观测值。
Finite variance: extreme outliers should not dominate without limit.
有限方差：极端离群值不应无限制地占据主导。

If these conditions are met, the approximation improves as n increases, but it is never perfect for small samples from highly non-normal populations.

如果满足这些条件，随着 n 增大，近似程度会提高；但对于来自高度非正态总体的小样本，近似永远不会完美。

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

The sampling distribution of X̄ has mean μ and variance σ²/n. Its standard deviation, σ/√n, is called the standard error. This distribution becomes more concentrated around μ as n increases, reflecting that larger samples give more precise estimates.

样本均值 X̄ 的抽样分布具有均值 μ 和方差 σ²/n。它的标准差 σ/√n 称为标准误。随着 n 增大，这个分布更加集中在 μ 附近，这反映了大样本能给出更精确的估计。

For a normal population, X̄ is exactly normal for any sample size. For a non-normal population, X̄ is only approximately normal when n is large. The table below summarises the difference.

对于正态总体，无论样本量大小，X̄ 都精确服从正态分布。对于非正态总体，仅当 n 较大时 X̄ 才近似服从正态分布。下表总结了这种差异。

Population distribution / 总体分布	Sample size / 样本量	Distribution of X̄ / X̄ 的分布
Normal / 正态	Any n / 任意 n	Exactly normal / 精确正态
Non-normal / 非正态	Small n / 小样本	Not necessarily normal / 不一定正态
Non-normal / 非正态	Large n / 大样本	Approximately normal by CLT / 由 CLT 近似正态

5. The Standard Error of the Mean | 均值的标准误

The standard error SE = σ/√n measures the variability of sample means from sample to sample. It decreases as the sample size increases, but at a rate proportional to the square root of n. To halve the standard error, you need four times the sample size.

标准误 SE = σ/√n 衡量的是不同样本之间样本均值的波动程度。它随着样本量的增大而减小，但减小速率与 n 的平方根成正比。要使标准误减半，需要将样本量增加到原来的四倍。

SE = σ / √n

标准误 = σ / √n

If the population standard deviation σ is unknown, the sample standard deviation s is used instead. This gives an estimated standard error s/√n, which is common in practical work.

如果总体标准差 σ 未知，则使用样本标准差 s 代替。这样就得到了估计的标准误 s/√n，在实际工作中非常常见。

6. Why the CLT Matters in Statistics | 中心极限定理在统计学中的重要性

The CLT justifies using normal-based methods for means even when the population is not normal. This allows statisticians to construct confidence intervals and perform hypothesis tests using the normal distribution, provided the sample size is large enough.

中心极限定理为在总体非正态时对均值使用基于正态分布的方法提供了理论依据。这使得统计学家能够在样本量足够大的情况下，使用正态分布来构建置信区间并进行假设检验。

Without the CLT, we would need to know the exact shape of the population distribution before making inferences about its mean. The CLT gives a universal route to probability calculations involving X̄.

如果没有中心极限定理，在对总体均值进行推断之前，我们需要知道总体分布的确切形状。中心极限定理为涉及 X̄ 的概率计算提供了一条通用的途径。

7. Worked Example: Applying the CLT to a Non-Normal Population | 应用示例：非正态总体的中心极限定理

Suppose the waiting time at a bus stop has a highly skewed distribution with mean 12 minutes and standard deviation 8 minutes. A random sample of 36 waiting times is taken. By the CLT, the sample mean is approximately normal with mean 12 and standard error 8/√36 = 8/6 ≈ 1.33 minutes. Therefore, X̄ ~ N(12, 1.33²).

假设公交车站的等待时间服从一个严重偏斜的分布，均值为 12 分钟，标准差为 8 分钟。随机抽取 36 个等待时间。根据中心极限定理，样本均值近似服从正态分布，均值为 12，标准误为 8/√36 = 8/6 ≈ 1.33 分钟。因此，X̄ ~ N(12, 1.33²)。

To find the probability that the sample mean is less than 10 minutes, we calculate the z-score:

为了求样本均值小于 10 分钟的概率，我们计算 z 分数：

z = (10 – 12) / 1.33 ≈ -1.50

Using standard normal tables, P(Z < -1.50) ≈ 0.0668. So there is about a 6.7% chance that the average waiting time in a sample of 36 is less than 10 minutes.

查标准正态分布表，P(Z < -1.50) ≈ 0.0668。因此，在 36 个样本中平均等待时间小于 10 分钟的概率约为 6.7%。

8. Common Misconceptions and Pitfalls | 常见误区与陷阱

A common mistake is to think the CLT says the sample data themselves become normally distributed as n increases. In fact, the CLT is about the distribution of the sample mean, not the raw data. Another misconception is that n ≥ 30 always works; for heavily skewed or heavy-tailed populations, a larger n may be necessary.

一个常见错误是认为中心极限定理说的是随着 n 增大，样本数据本身会变成正态分布。事实上，中心极限定理是关于样本均值的分布，而不是原始数据。另一个误区是认为 n ≥ 30 总是足够；对于严重偏斜或重尾的总体，可能需要更大的 n。

Also, the CLT does not apply to the sample variance or other statistics in the same way. It specifically concerns the mean of independent observations.

此外，中心极限定理并不以同样的方式适用于样本方差或其他统计量。它专门针对独立观测值的均值。

9. The CLT and Confidence Intervals | 中心极限定理与置信区间

When the sample size is large, an approximate 95% confidence interval for the population mean is given by X̄ ± 1.96(σ/√n). If σ is unknown, the sample standard deviation s can be used as an estimate, giving X̄ ± 1.96(s/√n). This interval is valid because of the CLT.

当样本量较大时，总体均值的一个近似 95% 置信区间为 X̄ ± 1.96(σ/√n)。如果 σ 未知，可以用样本标准差 s 作为估计，得到 X̄ ± 1.96(s/√n)。由于中心极限定理，这个区间是有效的。

95% CI for μ: X̄ ± 1.96 × (σ / √n)

μ 的 95% 置信区间：X̄ ± 1.96 × (σ / √n)

The value 1.96 is the z-score that cuts off 2.5% in each tail of the standard normal distribution. For other confidence levels, different z-scores are used, such as 1.645 for 90% and 2.576 for 99%.

数值 1.96 是标准正态分布中每侧尾部截去 2.5% 的 z 分数。对于其他置信水平，使用不同的 z 分数，例如 90% 用 1.645，99% 用 2.576。

10. The CLT in Hypothesis Testing | 中心极限定理在假设检验中的应用

In a hypothesis test for a population mean, the test statistic z = (X̄ – μ₀)/(σ/√n) is compared with critical values from the standard normal distribution. The CLT ensures that this z-statistic is approximately standard normal for large samples, even when the population is not normal.

在对总体均值进行假设检验时，检验统计量 z = (X̄ – μ₀)/(σ/√n) 与标准正态分布的临界值进行比较。中心极限定理保证了在大样本下，即使总体不服从正态分布，这个 z 统计量也近似服从标准正态分布。

z = (X̄ – μ₀) / (σ/√n)

If the calculated z falls in the critical region, we reject the null hypothesis H₀. Otherwise, we do not reject H₀. The CLT makes this procedure robust to non-normal population shapes.

如果计算出的 z 值落入拒绝域，我们就拒绝原假设 H₀。否则，我们不拒绝 H₀。中心极限定理使这一过程对非正态总体形状具有稳健性。

11. Quick Revision Summary | 快速复习总结

Remember: the CLT states that for large n, X̄ is approximately N(μ, σ²/n). The standard error is σ/√n. The approximation works regardless of the population shape, provided the sample is random and n is sufficiently large. Use the CLT to justify normal approximations for means in confidence intervals and hypothesis

Published by TutorHao | A-Level Computer Science Revision Series | aleveler.com

更多咨询请联系16621398022（同微信）

The Central Limit Theorem | 中心极限定理