A-Level Mathematics: Outlier Identification and Treatment Methods | A-Level 数学：异常值识别与处理方法

📚 A-Level Mathematics: Outlier Identification and Treatment Methods | A-Level 数学：异常值识别与处理方法

In statistics, an outlier is a data point that lies an unusual distance from other values in a random sample from a population. Identifying outliers is a core skill in A-Level Mathematics and Statistics, because a single extreme value can distort summary statistics, alter the shape of a distribution, and even change the conclusion of a hypothesis test.

在统计学中，异常值是指与随机样本中其他数值距离异常远的数据点。识别异常值是 A-Level 数学与统计学的核心技能，因为一个极端值就可能扭曲汇总统计量、改变分布形态、甚至改变假设检验的结论。

---

1. What Is an Outlier? | 什么是异常值？

An outlier is an observation that appears to deviate markedly from other members of the sample in which it occurs. In A-Level statistics, we usually classify outliers using formal numerical rules rather than by eye.

异常值是看起来与所在样本中其他观测值显著偏离的观测值。在 A-Level 统计学中，我们通常使用正式的数值规则而非目测来分类异常值。

Two formal methods are required in most examination boards:

大多数考试局要求掌握两种正式方法：

• the 1.5 × IQR rule, based on quartiles; | 基于四分位数的 1.5 × IQR 规则；
• the mean ± 2 or 3 standard deviation rule, based on the normal distribution. | 基于正态分布的均值 ± 2 或 3 个标准差规则。

---

2. The 1.5 × IQR Rule | 1.5 × IQR 规则

This is the most common method in A-Level examinations. First, find the lower quartile Q₁ and the upper quartile Q₃, then calculate the interquartile range:

这是 A-Level 考试中最常用的方法。首先求出下四分位数 Q₁ 与上四分位数 Q₃，然后计算四分位距：

IQR = Q₃ − Q₁

An observation x is considered an outlier if it lies outside the fences:

若观测值 x 位于下列界限之外，则被视为异常值：

Lower fence: Q₁ − 1.5 × IQR
Upper fence: Q₃ + 1.5 × IQR

Any value below the lower fence or above the upper fence is flagged as an outlier. Note that the rule uses 1.5, not 2 or 3; this is a convention specifically used for exploratory data analysis.

任何低于下界或高于上界的数值都被标记为异常值。注意规则使用 1.5 而非 2 或 3；这是探索性数据分析中专门采用的约定。

---

3. The Mean ± k × Standard Deviation Rule | 均值 ± k × 标准差规则

For data that are approximately normally distributed, we can use the mean μ and standard deviation σ. An observation is often treated as an outlier if it is more than 2 or 3 standard deviations away from the mean:

对于近似正态分布的数据，我们可以使用均值 μ 和标准差 σ。如果观测值偏离均值超过 2 或 3 个标准差，通常被视为异常值：

|x − μ| > 2σ or |x − μ| > 3σ

In a normal distribution, about 95% of data lie within 2 standard deviations of the mean, and about 99.7% lie within 3 standard deviations. A value beyond 3σ is very unusual and often merits investigation.

在正态分布中，约 95% 的数据落在均值 2 个标准差以内，约 99.7% 落在 3 个标准差以内。超过 3σ 的值非常罕见，通常值得进一步调查。

This method is especially useful when the data set is large and the distribution is symmetric. For skewed data, the IQR rule is generally preferred.

当数据集较大且分布对称时，这种方法尤其有用。对于偏斜数据，通常更推荐使用 IQR 规则。

---

4. Box Plots and Outlier Visualisation | 箱线图与异常值可视化

A box plot (box-and-whisker plot) is the standard graphical tool for showing outliers. The box spans from Q₁ to Q₃, with a line at the median. The whiskers extend to the smallest and largest values that are NOT outliers — that is, the most extreme values still within the fences.

箱线图（箱须图）是展示异常值的标准图形工具。箱体从 Q₁ 延伸到 Q₃，中间线位于中位数。须线延伸到不是异常值的最小值与最大值——即仍在界限内的最极端值。

Outliers are plotted individually as dots or crosses beyond the whiskers.

异常值以点或叉的形式单独绘制在须线之外。

• Mild outlier: beyond 1.5 × IQR but within 3 × IQR. | 温和异常值：超过 1.5 × IQR 但在 3 × IQR 之内。
• Extreme outlier: beyond 3 × IQR. | 极端异常值：超过 3 × IQR。

When drawing a box plot for an outlier-containing data set, the whiskers stop at the last non-outlier value, and the outliers are shown separately. This prevents the whiskers from being misleadingly long.

在绘制含异常值数据集的箱线图时，须线止于最后一个非异常值，异常值单独显示。这样可以避免须线被误导性地拉长。

---

5. Context: When Is an Outlier Not an Outlier? | 背景：异常值何时不算异常？

Statistical rules identify potential outliers, but context decides whether a flagged value is truly anomalous. A reading of 40 °C in a human temperature study is an outlier; the same reading in a chemical reaction experiment may be a perfectly valid observation.

统计规则识别潜在的异常值，但背景决定被标记的值是否真正异常。在人体体温研究中，40 °C 的读数是异常值；而在化学反应实验中，相同的读数可能完全是有效的观测。

There are three main sources of outliers:

异常值主要有三个来源：

• Measurement or recording errors — should usually be corrected or removed. | 测量或记录错误——通常应纠正或删除。
• Genuine extreme observations — must be retained unless there is strong evidence of error. | 真实的极端观测——除非有强有力的错误证据，否则必须保留。
• Data from a different population — may need to be analysed separately. | 来自不同总体的数据——可能需要进行单独分析。

In your exam, always justify your decision: state whether you are removing the value, and give a reason based on the context of the question.

在考试中，务必为你的决定说明理由：说明你是否删除该值，并基于题目背景给出理由。

---

6. Handling Outliers: Removal vs Retention | 处理异常值：删除与保留

Once an outlier is identified, the statistician must decide what to do with it. The two basic options are removal and retention, and there is also a third option: retain but report separately.

一旦识别出异常值，统计学家必须决定如何处理。两个基本选项是删除与保留，还有第三种选项：保留但单独报告。

Option | 选项	When to use | 适用情形
Remove | 删除	Known error; data entry mistake; impossible value (e.g. negative height). | 已知错误；录入失误；不可能的值（如负身高）。
Retain | 保留	Genuine extreme value; small sample size; no evidence of error. | 真实的极端值；样本量小；没有错误证据。
Retain and report separately | 保留并单独报告	Large sample; outlier is influential; want robust estimates for the majority. | 大样本；异常值影响大；希望为大多数数据获得稳健估计。

Removing an outlier without justification is dangerous: it can introduce bias and make results look better than they really are. Examiners reward clear reasoning about whether a data point should be excluded.

在没有正当理由的情况下删除异常值是危险的：它可能引入偏差，使结果看起来比实际更好。考官会赞赏对数据点是否应被排除的清晰推理。

---

7. Effect of Outliers on Measures of Central Tendency | 异常值对集中趋势量数的影响

Different averages respond differently to outliers. The mean is pulled toward the outlier because it uses every data value. The median is far more resistant because it depends only on the middle value(s).

不同的平均数

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

更多咨询请联系16621398022（同微信）