📚 Sample Mean and Variance, and Their Estimation | 样本均值与方差及其估计
In statistics, a fundamental task is to estimate unknown population parameters using sample data. The sample mean and sample variance are the most common estimators for the population mean μ and population variance σ². Understanding their properties—especially bias, sampling distributions, and the reasoning behind Bessel's correction—is essential for making reliable inferences in IB Mathematics.
在统计学中，一项基本任务是用样本数据估计未知的总体参数。样本均值和样本方差是最常用的对于总体均值μ和总体方差σ²的估计量。理解它们的性质——尤其是偏差、抽样分布以及贝塞尔校正的原理——对于在IB数学中进行可靠的推断至关重要。
1. Population Parameters: μ and σ² | 总体参数：μ 与 σ²
For a population of size N tion variance σ² measures the spread of the data around μ, given by σ² = (Σᵢ (xᵢ − μ)²)/N.
对于大小为 N（有限或概念上的无限）的总体，总体均值 μ 是所有观测值的平均值。在有限总体中，它定义为 μ = (Σᵢ xᵢ)/N。总体方差 σ² 衡量数据围绕 μ 的离散程度，由 σ² = (Σᵢ (xᵢ − μ)²)/N 给出。
These parameters are usually unknown and fixed. We draw a random sample of size n to estimate them. The goal is to construct estimators that are unbiased, consistent, and efficient, enabling valid conclusions about the population.
这些参数通常是未知且固定的。我们抽取一个大小为 n 的随机样本来估计它们。目标是构建无偏、一致且高效的估计量，从而能够对总体得出有效的结论。
2. Sample Mean as an Estimator | 样本均值作为估计量
For a sample (x₁, x₂, …, xₙ), the sample mean is the arithmetic average x̄ timator of the population mean μ. Because the sample mean depends on the particular sample drawn, it is a random variable with its own distribution.
对于一个样本 (x₁, x₂, …, xₙ)，样本均值为算术平均值 x̄ = (Σᵢ xᵢ)/n。它是对总体均值 μ 的自然点估计量。由于样本均值依赖于所抽取的特定样本，它是一个随机变量，拥有自己的分布。
The expected value of the sample mean is E(x̄ n unbiased estimator of μ. No matter what the population distribution looks like, on average, the sample mean hits the true mean.
样本均值的期望值为 E(x̄) = μ，这证明了样本均值是 μ 的无偏估计量。无论总体分布形态如何，平均而言，样本均值都能命中真实均值。
3. Variance and Standard Error of the Sample Mean | 样本均值的方差与标准误差
nce σ² is known and observations are independent, the variance of the sample mean is Var(x̄) = σ²/n. Consequently, the standard deviation of x̄, known as the standard error of the mean, is SE(x̄) = σ/√n.
如果总体方差 σ² 已知且观测值独立，则样本均值的方差为 Var(x̄) = σ²/n。因此，x̄ 的标准差——称为均值的标准误差——为 SE(x̄) = σ/√n。
This relationship shows that as the sample size n increases, the precision of x̄ improves: the standard error decreases by a factor of 1/√n. Quadrupling the sample size halves the standard error.
这一关系表明，随着样本容量 n 增加，x̄ 的精确度提高：标准误差以 1/√n 的比例减小。样本量增至四倍，标准误差减半。
4. Sample Variance: Definition and Bessel's Correction | 样本方差：定义与贝塞尔校正
population variance σ², we usually rely on the sample variance defined as s² = (Σᵢ (xᵢ − x̄)²)/(n−1). The divisor n−1 instead of n is known as Bessel's correction. Without this correction, the estimator calculated with n would be biased, tending to underestimate the true variance.
当估计总体方差 σ² 时，我们通常使用样本方差，定义为 s² = (Σᵢ (xᵢ − x̄)²)/(n−1)。除数用 n−1 而非 n 被称为贝塞尔校正。若不进行此校正，使用 n 计算的估计量将是有偏的，往往低估真实方差。
making s² an unbiased estimator of the population variance. The sample standard deviation s is the square root of s², but note that s itself is not unbiased for σ; however, it is a consistent estimator.
关键性质在于 E(s²) = σ²，使得 s² 成为总体方差的无偏估计量。样本标准差 s 是 s² 的平方根，但请注意 s 本身对于 σ 并非无偏；但它是一致的估计量。
5. Why n−1? Degrees of Freedom | 为什么是 n−1？自由度
The reason for the n−1 divisor lies in the concept of degrees of freedom. When we compute deviations (xᵢ − x̄), we are using the sample mean x̄, which is itself calculated from the data. Because Σ (xᵢ − x̄) = 0, only n−1 of these deviations are free to vary; the last one is determined by the others.
使用 n−1 除数的原因在于自由度的概念。当我们计算离差 (xᵢ − x̄) 时，使用了从数据中算出的样本均值 x̄。由于 Σ (xᵢ − x̄) = 0，这些离差中只有 n−1 个可以自由变动；最后一个被其余所决定。
Mathematically, one can show that Σ (xᵢ − μ)² = Σ (xᵢ − x̄)² + n(x̄ − μ)². Taking expectations and using E[(x̄ − μ)²] = σ²/n gives E[Σ (xᵢ − x̄)²] = (n−1)σ², confirming that dividing by n−1 yields an unbiased estimate.
从数学上可以证明 Σ (xᵢ − μ)² = Σ (xᵢ − x̄)² + n(x̄ − μ)²。取期望并利用 E[(x̄ − μ)²] = σ²/n 可得到 E[Σ (xᵢ − x̄)²] = (n−1)σ²，从而确认除以 n−1 能产生无偏估计。
6. Unbiasedness and Consistency | 无偏性与一致性
An estimator is called unbiased if its expected value equals the true parameter. Both x̄ and s² enjoy this property: E(x̄ res that, over many repeated samples, we neither systematically overestimate nor underestimate the parameter.
如果一个估计量的期望值等于真实参数，则称其为无偏的。x̄ 和 s² 都具有这一性质：E(x̄) = μ 且 E(s²) = σ²。无偏性保证了在大量重复抽样中，我们既不会系统性地高估也不会低估参数。
Furthermore, as the sample size grows, the variances of both estimators shrink to zero, meaning they converge in probability to the true parameters. This property is called consistency and is essential for any reliable estimator.
此外，随着样本量的增加，两个估计量的方差都趋近于零，这意味着它们依概率收敛到真实参数。这一性质称为一致性，是任何可靠估计量的必要条件。
7. Sampling Distribution of the Mean (Normal Population) | 均值的抽样分布（正态总体）
If the population is normally distributed with mean μ and variance σ², then for any sample size n, the sample mean follows an exact normal distribution: x̄ ~ N(μ, σ²/n). This allows us to make precise probability statements about x̄ using standard normal tables.
如果总体服从均值为 μ、方差为 σ² 的正态分布，那么对于任意样本容量 n，样本均值都精确服从正态分布：x̄ ~ N(μ, σ²/n)。这使得我们可以使用标准正态表对 x̄ 做出精确的概率陈述。
When σ is unknown, we replace it with the sample standard deviation s. The resulting standardized statistic t = (x̄ − μ)/(s/√n) follows a t-distribution with n−1 degrees of freedom. This forms the basis for t-tests and t-based confidence intervals.
当 σ 未知时，我们用样本标准差 s 替代它。得到的标准化统计量 t = (x̄ − μ)/(s/√n) 服从自由度为 n−1 的 t 分布。这构成了 t 检验和基于 t 分布的置信区间的基础。
8. Central Limit Theorem (CLT) | 中心极限定理
A remarkably powerful result is the Central Limit Theorem: regardless of the population distribution (as long as it has a finite variance), the distribution of the sample mean x̄ approaches a normal distribution as the sample size n increases. For practical purposes, n ≥ 30 is often considered sufficient.
一个非常强大的结论是中心极限定理：无论总体分布如何（只要方差有限），随着样本量 n 的增加，样本均值 x̄ 的分布趋近于正态分布。在实际应用中，n ≥ 30 通常被认为是足够的。
Specifically, for large n, x̄ is approximately N(μ, σ²/n). This enables the use of normal-based inference for means even when the population is skewed, uniform, or otherwise non-normal. The same property justifies the large-sample confidence intervals and hypothesis tests encountered in IB examinations.
具体而言，对于较大的 n，x̄ 近似服从 N(μ, σ²/n)。这使得即使总体是偏态的、均匀的或其他非正态的，也可以对均值使用基于正态的推断。同一性质也支持IB考试中遇到的大样本置信区间和假设检验。
9. Confidence Interval for μ When σ Is Known | σ 已知时 μ 的置信区间
If the population standard deviation σ is known (or the sample is very large), a 95% confidence interval for the population mean μ is given by x̄ ± z* × (σ/√n), where the critical value z* ≈ 1.96. This interval is constructed such that in 95% of all possible samples, it will contain μ.
如果总体标准差 σ 已知（或样本非常大），总体均值 μ 的 95% 置信区间由 x̄ ± z* × (σ/√n) 给出，其中临界值 z* ≈ 1.96。该区间的构造使得在所有可能的样本中，有 95% 的区间会包含 μ。
More generally, for a confidence level (1−α) × 100%, we use z* = zα/2 from the standard normal distribution. Wider intervals correspond to higher confidence. As n grows, the margin of error z*σ/√n shrinks, yielding more precise estimates.
更一般地，对于置信水平 (1−α) × 100%，我们使用标准正态分布中的 z* = zα/2。更宽的区间对应更高的置信水平。随着 n 增大，误差范围 z*σ/√n 缩小，从而产生更精确的估计。
10. Confidence Interval for μ When σ Is Unknown (t-Interval) | σ 未知时 μ 的置信区间（t 区间）
In most real scenarios, σ is unknown and must be estimated by s. The appropriate interval is the t-interval: x̄ ± tₙ₋₁,α/2 × (s/√n), where tₙ₋₁,α/2 is the critical value from the t-distribution with n−1 degrees of freedom.
在大多数实际情形中，σ 是未知的，必须由 s 估计。合适的区间是 t 区间：x̄ ± tₙ₋₁,α/2 × (s/√n)，其中 tₙ₋₁,α/2 是自由度为 n−1 的 t 分布的临界值。
The t-distribution has heavier tails than the normal distribution, reflecting the extra uncertainty from estimating σ. As n increases, the t-distribution approaches the standard normal, and the t-interval becomes very close to the z-interval. For small samples, however, using the t-interval is essential to maintain the stated confidence level.
t 分布比正态分布具有更厚的尾部，这反映了因估计 σ 而引入的额外不确定性。随着 n 增加，t 分布趋近于标准正态分布，t 区间变得与 z 区间非常接近。然而，对于小样本，使用 t 区间对于维持所声称的置信水平至关重要。
11. Summary and Common Pitfalls | 总结与常见误区
The sample mean x̄ is an unbiased and consistent estimator of μ. The sample variance s² with divisor n−1 is unbiased for σ². The standard error of x̄ decreases as 1/√n, and the Central Limit Theorem ensures approximate normality for large samples, enabling powerful inference tools.
样本均值 x̄ 是 μ 的无偏且一致的估计量。除数为 n−1 的样本方差 s² 是对 σ² 的无偏估计。x̄ 的标准误差以 1/√n 的比例递减，中心极限定理确保了大样本的近似正态性，从而提供了强大的推断工具。
Common mistakes include: using n instead of n−1 when estimating σ²; applying z-based intervals when σ is unknown and the sample is small; and forgetting that s is not an unbiased estimator of σ, though s² is unbiased. Mastery of these concepts is fundamental for success in IB Mathematics statistical analysis.
常见错误包括：估计 σ² 时使用 n 而非 n−1；在 σ 未知且样本较小时使用基于 z 的区间；以及忘记 s 不是 σ 的无偏估计量（尽管 s² 是无偏的）。掌握这些概念是在IB数学统计分析中取得成功的基础。
Published by TutorHao | IB 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