📚 Understanding Outliers in A-Level Statistics | 理解A-Level统计学中的离群值
In A-Level mathematics, an outlier is a data point that differs significantly from other observations. Recognising and handling outliers is a key skill in statistical analysis.
在A-Level数学中，离群值是指与其他观测值差异显著的数据点。识别并处理离群值是统计分析中的关键技能。
1. What is an Outlier? | 什么是离群值？
An outlier is a value that lies far outside the overall pattern of a data set. It may be unusually small or unusually large compared to the rest of the data.
离群值是指远离数据集整体模式的值。与其他数据相比，它可能异常小或异常大。
For example, in the data set {2, 3, 4, 5, 6, 40}, the value 40 is a possible outlier because it is far greater than the other values.
例如，在数据集{2, 3, 4, 5, 6, 40}中，40可能是离群值，因为它远大于其他数值。
Outliers can arise from measurement errors, data entry mistakes, or genuine rare events. Understanding their origin helps decide how to treat them.
离群值可能来自测量误差、数据录入错误或真实的稀有事件。了解其来源有助于决定如何处理它们。
2. The 1.5 × IQR Rule | 1.5倍四分位距规则
The most common method in Edexcel A-Level is the interquartile range (IQR) rule. First, find the lower quartile (Q₁) and upper quartile (Q₃). The IQR is Q₃ − Q₁.
Edexcel A-Level中最常用的方法是四分位距（IQR）规则。首先找到下四分位数Q₁和上四分位数Q₃，IQR等于Q₃ − Q₁。
Any data value below Q₁ − 1.5 × IQR or above Q₃ + 1.5 × IQR is considered an outlier.
任何小于Q₁ − 1.5 × IQR或大于Q₃ + 1.5 × IQR的数据值都被视为离群值。
Lower fence = Q₁ − 1.5 × IQR
Upper fence = Q₃ + 1.5 × IQR
3. Worked Example of the IQR Rule | IQR规则的应用示例
Consider the ordered data set: 3, 5, 7, 8, 9, 11, 15, 21, 25.
考虑有序数据集：3, 5, 7, 8, 9, 11, 15, 21, 25。
Here, the median is 9. Q₁ is the median of the lower half (3, 5, 7, 8), so Q₁ = (5+7)/2 = 6. Q₃ is the median of the upper half (11, 15, 21, 25), so Q₃ = (15+21)/2 = 18.
这里，中位数为9。Q₁是下半部分(3, 5, 7, 8)的中位数，所以Q₁ = (5+7)/2 = 6。Q₃是上半部分(11, 15, 21, 25)的中位数，所以Q₃ = (15+21)/2 = 18。
IQR = 18 − 6 = 12. Lower fence = 6 − 1.5 × 12 = −12. Upper fence = 18 + 1.5 × 12 = 36.
IQR = 18 − 6 = 12。下限 = 6 − 1.5 × 12 = −12。上限 = 18 + 1.5 × 12 = 36。
Since all values lie between −12 and 36, there are no outliers in this data set.
由于所有值都在−12和36之间，该数据集中没有离群值。
4. Using Standard Deviation to Detect Outliers | 用标准差检测离群值
Another method assumes a normal distribution. Values more than 2 or 3 standard deviations from the mean are often flagged as outliers.
另一种方法假设数据服从正态分布。距离均值超过2或3个标准差的值通常被标记为离群值。
For a sample with mean x̄ and standard deviation s, a value x is an outlier if |x − x̄| > 2s (or sometimes 3s).
对于均值为x̄、标准差为s的样本，如果|x − x̄| > 2s（有时用3s），则x是离群值。
However, this method is sensitive to the outlier itself, because the mean and standard deviation are affected by extreme values.
然而，这种方法对离群值本身很敏感，因为均值和标准差会受到极端值的影响。
5. The Effect of Outliers on the Mean and Median | 离群值对均值和中位数的影响
An outlier can greatly distort the mean, pulling it toward the extreme value. The median, being the middle value, is more robust and less affected.
离群值会严重扭曲均值，使其向极端值方向偏移。而中位数作为中间值，更具稳健性，受影响较小。
For example, in {2, 3, 4, 5, 6, 40}, the mean is (2+3+4+5+6+40)/6 = 10, but the median is (4+5)/2 = 4.5.
例如，在{2, 3, 4, 5, 6, 40}中，均值是(2+3+4+5+6+40)/6 = 10，但中位数是(4+5)/2 = 4.5。
Thus, when outliers are present, the median is often a better measure of central tendency than the mean.
因此，当存在离群值时，中位数往往比均值更能代表集中趋势。
6. Effect on Variance and Standard Deviation | 对方差和标准差的影响
Outliers inflate the variance and standard deviation because they increase the total squared distance from the mean.
离群值会增大方差和标准差，因为它们增加了与均值的总平方距离。
This makes the spread of the data appear larger than it really is, which can lead to misleading conclusions if not handled carefully.
这会使数据的离散程度看起来比实际更大，如果不谨慎处理，可能导致误导性结论。
Robust measures such as the IQR and median absolute deviation are less sensitive to outliers and are often preferred in exploratory analysis.
像IQR和中位绝对偏差等稳健度量对离群值不太敏感，在探索性分析中常常更受青睐。
7. Box Plots and Outliers | 箱线图与离群值
A box plot visually displays outliers as individual points beyond the whiskers. The whiskers extend to the smallest and largest values that are not outliers.
箱线图将离群值显示为须线之外的单独点。须线延伸到不是离群值的最小值和最大值。
In Edexcel questions, you may be asked to draw a box plot and identify outliers using the fences.
在Edexcel的考题中，你可能被要求绘制箱线图并使用上下限识别离群值。
Values beyond the fences are plotted as dots or asterisks, making the distribution of outliers clear.
超出上下限的值以点或星号绘制，使离群值的分布一目了然。
8. When to Remove an Outlier | 何时移除离群值
An outlier should not be removed without justification. If it is due to a recording error, it may be corrected or deleted. If it represents a real event, it should be kept.
没有充分理由不应移除离群值。如果是记录错误，可以纠正或删除；如果代表真实事件，则应保留。
Always report whether outliers were removed and explain the reason in your analysis.
在分析中，务必说明是否移除了离群值，并解释原因。
Removing a genuine outlier can bias results, while keeping an erroneous one can distort conclusions.
移除真实的离群值可能导致偏差，而保留错误的离群值则可能扭曲结论。
9. Outliers in Linear Regression | 线性回归中的离群值
In regression analysis, an outlier can have a strong influence on the least squares regression line, especially if it has an extreme x-value.
在回归分析中，离群值对最小二乘回归线可能有很强的影响，尤其是当它拥有极端的x值时。
Such a point is called a high-leverage point. It can pull the regression line toward itself and reduce the overall fit.
这样的点称为高杠杆点。它可能将回归线拉向自身并降低整体拟合度。
Detecting outliers in regression often involves examining residuals — large residual values indicate potential outliers.
在回归中检测离群值通常需要检查残差——较大的残差值表示潜在的离群值。
10. Exam Tips for Edexcel A-Level | Edexcel A-Level考试提示
Always define your quartiles correctly using the method taught in your specification. Write down the fence formulas clearly before substituting numbers.
务必使用考纲规定的方法正确计算四分位数。在代入数值前，清晰写出上下限公式。
When asked to comment on outliers, refer to both the IQR rule and the context of the data. Justify whether the outlier should be included.
当被要求评论离群值时，要同时引用IQR规则和数据背景，并说明是否应包含该离群值。
Practice drawing box plots and calculating quartiles from stem-and-leaf diagrams and cumulative frequency graphs, as these are common question types.
练习从茎叶图和累积频率图绘制箱线图并计算四分位数，这是常见题型。
Finally, remember that an outlier is not always a mistake — it may be the most interesting observation in the study.
最后，记住离群值不一定是错误——它可能是研究中最重要的观测值。
Published by TutorHao | Mathematics Revision Series | aleveler.com
更多咨询请联系16621398022（同微信）
屏轩国际教育cambridge primary/secondary checkpoint, cat4, ukiset,ukcat,igcse,alevel,PAT,STEP,MAT, ibdp,ap,ssat,sat,sat2课程辅导，国外大学本科硕士研究生博士课程论文辅导Cancel reply