📚 Pre-U CIE Statistics: Common Misconceptions and Correction Methods | Pre-U CIE 统计：常见误区与纠正方法
In Cambridge Pre-U Mathematics (9798) and Further Mathematics (9795), statistics questions are marked with method marks that depend on correct statistical reasoning, not just correct arithmetic. Many students lose those marks through a small set of persistent misconceptions: using the wrong divisor, confusing the direction of a conditional probability, adding standard deviations instead of variances, or applying a continuity correction on the wrong side. This article collects eleven of the most damaging misconceptions in the Pre-U statistics components and pairs each one with an explicit correction method, worked formulas and an exam-ready checklist.
在剑桥 Pre-U 数学（9798）与进阶数学（9795）中，统计题的评分十分看重统计推理的方法分，而不只是算对数字。许多学生的失分来自少数几个顽固的误区：用错了除数、搞混条件概率的方向、把标准差直接相加而不是把方差相加，或者在错误的边界上做连续性修正。本文汇总了 Pre-U 统计部分中最容易造成失分的十一个误区，并为每一个误区配上明确的纠正方法、公式和考场可直接使用的检查清单。
1. Confusing Population and Sample Variance | 混淆总体方差与样本方差
A frequent error is using the divisor n when computing the variance of a data set that is actually a sample. Doing so produces the population variance, that is, the second moment about the mean, rather than the unbiased sample variance. Because the divisor n is smaller than n − 1, the result is systematically too small, and the underestimation becomes serious when the sample size is small.
一个常见错误是在计算其实属于样本的数据集的方差时使用除数 n。这样得到的是总体方差，也就是关于均值的二阶矩，而不是无偏的样本方差。由于除数 n 比 n − 1 小，结果会系统性地偏小，而且当样本容量较小时，这种低估会变得相当严重。
The correction method is simple but must be applied consciously at the start of every question. Ask whether you are describing a complete population or a subset drawn from a larger group. For a sample of size n, compute the unbiased estimate with the divisor n − 1; for a discrete probability distribution, use the population formula, because the distribution itself describes the whole population.
纠正方法很简单，但必须在每道题开始时自觉地执行。先判断你描述的是一个完整的总体，还是从更大群体中抽取的一个子集。对于容量为 n 的样本，用除数 n − 1 计算无偏估计；对于离散概率分布，则使用总体公式，因为该分布本身就描述了整个总体。
s² = Σ(xᵢ − x̄)² / (n − 1)
σ² = Σ(xᵢ − μ)² / N = Σxᵢ²P(X = xᵢ) − μ²
A quick comparison of contexts makes the choice unambiguous:
下面这张对照表可以让你的选择变得明确无误：
| Context | 情境 | Divisor | 除数 | Symbol | 符号 | Reason | 理由 |
|---|---|---|---|
| Sample of data | 数据样本 | n − 1 | s² | Unbiased estimator | 无偏估计量 |
| Whole population | 完整总体 | N | σ² | Descriptive measure | 描述性度量 |
| Probability distribution | 概率分布 | — | Var(X) | Model of whole population | 描述整个总体的模型 |
2. The σx and sx Calculator Trap | 计算器上 σx 与 sx 的陷阱
Almost every Pre-U candidate uses a graphical calculator, and almost every statistics paper contains at least one question where the wrong key is pressed. In the standard statistics mode of a graphical calculator, σx (or σx with the population label) returns the population standard deviation with divisor n, while sx (or σn − 1) returns the sample standard deviation with divisor n − 1. Quoting the wrong one loses the accuracy mark even when the method is perfect.
几乎所有 Pre-U 考生都会使用图形计算器，而几乎每份统计试卷中都至少有一道题会因为按错键而丢分。在图形计算器的标准统计模式中，σx（标为总体标准差）给出的是以 n 为除数的总体标准差，而 sx（或标为 σn − 1）给出的是以 n − 1 为除数的样本标准差。即使方法完全正确，报错哪一个也会丢掉精度分。
ant sx. If it says a discrete random variable X has the given probability distribution, you want the population value, which you should obtain from Σx²P(X = x) − μ² rather than from the statistics mode. Write the symbol down in your working first, then read the matching value off the screen.
纠正方法是在碰到计算器之前先确定要用哪个符号。如果题目说"随机抽取 20 个元件进行测试"，你要的是 sx；如果题目说"离散随机变量 X 具有如下概率分布"，你要的是总体值，而这个值应当由 Σx²P(X = x) − μ² 得到，而不是从统计模式里读。先在草稿上写下你要的符号，再从屏幕上读对应的数值。
- Sample data with a stated sample size | 给出了样本容量的数据 → use s and s² | 用 s 与 s²
- Probability distribution table | 概率分布表 → use μ, σ and σ² | 用 μ、σ 与 σ²
- Combined or pooled data sets | 合并数据集 → recompute from raw totals, never average the standard deviations | 用原始总量重算，绝不能把标准差取平均
3. Standard Deviation versus Standard Error | 标准差与标准误
The standard deviation describes the spread of individual observations, while the standard error describes the spread of the sample mean as an estimator. Students routinely write σ where the standard error σ / √n s for a mean. The two quantities coincide only when n = 1, which never happens in an exam.
标准差描述的是单个观测值的离散程度，而标准误描述的是样本均值这一估计量的离散程度。学生常常在该用标准误 σ / √n 的地方写成 σ，尤其是在均值的置信区间与假设检验中。这两个量只有在 n = 1 时才相等，而考试中永远不会出现这种情况。
The correction method is to name the quantity you are describing. If the sentence is about "the mean of a sample of 25", the standard error applies. If the sentence is about "the time taken by a randomly chosen student", the standard deviation applies. When the population standard deviation is unknown and the sample is large, replace σ by s in the standard error formula and state that you are using the sample estimate.
纠正方法是先明确你描述的对象。如果句子说的是"25 个样本的均值"，就应该用标准误；如果句子说的是"随机抽取的一名学生所用时间"，就应该用标准差。当总体标准差未知而样本较大时，把标准误公式中的 σ 换成 s，并说明你使用的是样本估计值。
SE(x̄) = σ / √n or SE(x̄) = s / √n
Var(mean of n independent observations) = σ² / n
4. Mutually Exclusive versus Independent | 互斥与独立
are not two names for the same idea; they are almost opposites in most contexts. Two events are mutually exclusive when they cannot occur together, so P(A ∩ ndependent when the occurrence of one does not change the probability of the other, so P(A ∩ ero probability, mutual exclusivity actually implies dependence, because knowing that A happened makes B impossible.
互斥与独立并不是同一个概念的两个名字，在大多数情境下它们几乎是相反的。两个事件互斥是
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