📚 Correlation: Correlation Coefficient and Scatter Plot Interpretation | 相关性：相关系数与散点图解读
Correlation is a statistical tool that measures the strength and direction of a linear relationship between two quantitative variables. In A-Level mathematics and statistics, understanding correlation is essential for interpreting real-world data and for building a foundation for regression analysis.
相关性是统计学中用于衡量两个定量变量之间线性关系的强度和方向的工具。在 A-Level 数学与统计中，理解相关性是解读实际数据以及为回归分析打下基础的关键。
1. What Is Correlation? | 什么是相关性？
Correlation describes whether two variables move together. If one variable tends to increase when the other increases, the correlation is positive. If one variable tends to decrease when the other increases, the correlation is negative. If there is no consistent pattern, the correlation is close to zero.
相关性描述的是两个变量是否一同变化。如果一个变量随着另一个变量的增加而增加，则相关性为正；如果一个变量随着另一个变量的增加而减少，则相关性为负；如果不存在一致的模式，则相关性接近于零。
It is important to note that correlation is only concerned with linear association. Two variables may be strongly related in a curved way, yet the correlation coefficient may still be close to zero.
需要注意的是，相关性仅关注线性关联。两个变量可能以曲线方式密切相关，但相关系数仍可能接近于零。
2. Scatter Plots: The Visual Foundation | 散点图：可视化的基础
A scatter plot is a graph in which each point represents one observation, plotted with one variable on the horizontal axis (x) and the other on the vertical axis (y). Before calculating any numerical coefficient, always draw a scatter plot.
散点图是一种图表，其中每个点代表一个观测值，一个变量标在横轴 (x) 上，另一个变量标在纵轴 (y) 上。在计算任何数值系数之前，务必先绘制散点图。
When reading a scatter plot, look for three features: direction, form, and strength. Direction tells you whether the association is positive or negative. Form tells you whether the relationship is linear or curved. Strength tells you how closely the points cluster around a straight line.
阅读散点图时，应关注三个特征：方向、形态和强度。方向告诉我们关联是正还是负；形态告诉我们关系是线性还是弯曲；强度告诉我们点围绕一条直线的聚集程度。
3. Positive, Negative, and Zero Correlation | 正相关、负相关与零相关
In a positive correlation, the points on a scatter plot generally rise from left to right. For example, height and shoe size in a group of teenagers usually show a positive correlation.
在正相关中，散点图上的点通常从左到右上升。例如，一组青少年的身高与鞋码通常呈现出正相关。
In a negative correlation, the points generally fall from left to right. For example, the speed of a car and the time taken to travel a fixed distance are negatively correlated.
在负相关中，点通常从左到右下降。例如，汽车的速度与行驶固定距离所需的时间呈负相关。
When the points show no obvious upward or downward trend, the correlation is said to be zero or very weak. For instance, a person’s shoe size and their exam score in mathematics are likely to have near-zero correlation.
当点没有明显的上升或下降趋势时，相关性被称为零相关或极弱相关。例如，一个人的鞋码与其数学考试成绩之间很可能接近于零相关。
4. The Pearson Correlation Coefficient | 皮尔逊相关系数
The most common measure of linear correlation is the Pearson product-moment correlation coefficient, denoted by r. It is defined using the sums of squares and the sum of cross-products.
最常用的线性相关度量是皮尔逊积矩相关系数，记作 r。它通过平方和与交叉乘积和来定义。
r = Sxy / √(Sxx × Syy)
where
其中
- Sxx = Σ(xᵢ − x̄)²
- Syy = Σ(yᵢ − ȳ)²
- Sxy = Σ(xᵢ − x̄)(yᵢ − ȳ)
Here x̄ is the mean of the x-values and ȳ is the mean of the y-values. Alternatively, Sxy, Sxx, and Syy can be calculated from summary statistics using the formulas:
这里 x̄ 是 x 值的均值，ȳ 是 y 值的均值。或者，Sxy、Sxx 和 Syy 可以通过以下汇总统计量公式计算：
Sxx = Σxᵢ² − (Σxᵢ)²/n
Syy = Σyᵢ² − (Σyᵢ)²/n
Sxy = Σxᵢyᵢ − (Σxᵢ)(Σyᵢ)/n
These forms are especially useful in examinations because they avoid calculating every deviation from the mean separately.
这些形式在考试中尤其有用，因为它们避免了一次一次地计算每个观测值与均值的偏差。
5. Key Properties of r | 相关系数 r 的重要性质
The value of r is always between −1 and 1 inclusive. This is guaranteed by the Cauchy-Schwarz inequality.
r 的值始终在 −1 和 1 之间（包含端点）。这是由柯西-施瓦茨不等式保证的。
- r = 1 means a perfect positive linear relationship.
- r = −1 means a perfect negative linear relationship.
- r = 0 means no linear relationship.
- r = 1 表示完全正线性关系。
- r = −1 表示完全负线性关系。
- r = 0 表示不存在线性关系。
r is unitless, which means it does not depend on the units of measurement. If you change centimetres to metres, or pounds to kilograms, the value of r remains unchanged.
r 是无量纲的，这意味着它不依赖于测量单位。如果把厘米换成米，或把磅换成千克，r 的值不会改变。
r is symmetric: the correlation between x and y is the same as the correlation between y and x. Also, r is unaffected by adding a constant to all values of x or y, or by multiplying all values by a positive constant.
r 具有对称性：x 与 y 的相关系数等于 y 与 x 的相关系数。此外，给所有 x 或 y 值加上一个常数，或乘以一个正的常数，r 不会改变。
6. Interpreting the Value of r | 解读 r 的数值
Although there are no universal cut-offs, the following guide is commonly used for a sample of moderate size:
虽然没有统一的标准，但以下指南常用于中等规模的样本：
| Value of r | Strength of Linear Relationship | 线性关系强度 |
| |r| = 0.9 to 1.0 | Very strong | 非常强 |
| |r| = 0.7 to 0.9 | Strong | 强 |
| |r| = 0.4 to 0.7 | Moderate | 中等 |
| |r| = 0 to 0.4 | Weak or none | 弱或无 |
A high value of r means that the points lie close to a straight line, but it does not mean that one variable causes the other to change.
r 的值高意味着点紧密地围绕一条直线分布，但并不意味着一个变量导致另一个变量变化。
7. The Coefficient of Determination r² | 决定系数 r²
The square of the correlation coefficient, r², is called the coefficient of determination. It represents the proportion of the total variation in y that is explained by the linear relationship with x.
相关系数的平方 r² 被称为决定系数。它表示 y 的总变异中能被与 x 的线性关系所解释的比例。
For example, if r = 0.8, then r² = 0.64, so 64% of the variation in y is explained by the linear relationship with x. The remaining 36% is due to other factors or randomness.
例如，如果 r = 0.8，那么 r² = 0.64，即 y 的变异中有 64% 可由与 x 的线性关系解释，其余 36% 归因于其他因素或随机性。
When a question asks “what percentage of the variation is explained by the regression line?”, you should calculate r², not r.
当题目问“回归线解释了多少比例的变异”时，应计算 r²，而不是 r。
8. Spearman’s Rank Correlation Coefficient | 斯皮尔曼等级相关系数
Spearman’s rank correlation coefficient, usually denoted by rₛ, is a non-parametric measure of correlation. It is calculated by ranking the data and then applying a formula to the ranks.
斯皮尔曼等级相关系数，通常记作 rₛ，是一种非参数的相关性度量。它通过将数据排序，然后对排名应用公式来计算。
rₛ = 1 − (6Σdᵢ²) / (n(n² − 1))
where dᵢ is the difference between the ranks of the i-th observation, and n is the number of observations. This formula is valid when there are no tied ranks.
其中 dᵢ 是第 i 个观测值的两个排名之差，n 是观测值个数。在没有并列排名时该公式有效。
Spearman’s coefficient measures monotonic relationships, not just linear ones. It is often used when the data are ordinal or when the relationship is curved but consistently increasing or decreasing.
斯皮尔曼系数衡量的是单调关系，而不仅仅是线性关系。当数据是顺序数据，或关系是曲线但持续上升或下降时，通常使用它。
9. Worked Example: Calculating r | 例题：计算 r
Consider five students with the following numbers of hours studied (x) and test scores (y):
考虑五名学生，学习小时数 (x) 和考试成绩 (y) 如下：
| x (hours) | 2 | 3 | 5 | 6 | 9 |
| y (score) | 50 | 55 | 70 | 75 | 90 |
First calculate the required sums: Σx = 25, Σy = 340, Σx² = 155, Σy² = 24450, Σxy = 1795, and n = 5.
首先计算所需的求和：Σx = 25，Σy = 340，Σx² = 155，Σy² = 24450，Σxy = 1795，n = 5。
Sxx = 155 − 25²/5 = 155 − 125 = 30
Syy = 24450 − 340²/5 = 24450 − 23120 = 1330
Sxy = 1795 − (25 × 340)/5 = 1795 − 1700 = 95
r = 95 / √(30 × 1330) = 95 / √39900 ≈ 95 / 199.75 ≈ 0.476
The correlation is moderate and positive. The scatter plot would show a generally rising trend, but the points would not lie very close to a straight line.
相关性为中等正相关。散点图会呈现总体上升的趋势，但点不会非常接近一条直线。
10. Outliers and Their Influence | 异常值及其影响
An outlier is an observation that lies far away from the general pattern of the data. A single outlier can have a dramatic effect on the value of r, especially in a small sample.
异常值是指远离数据总体模式的观测值。一个单独的异常值可能对 r 的值产生巨大影响，尤其是在小样本中。
An outlier that follows the overall trend may increase r, while an outlier that goes against the trend may reduce r. Always examine the scatter plot before trusting r.
一个与总体趋势一致的异常值可能会增大 r，而与趋势相反的异常值可能会减小 r。在相信 r 之前，务必先检查散点图。
For example, if most points show a strong positive linear trend but one point has a very large x and a very small y, the calculated r may become much weaker or even negative.
例如，如果大多数点呈现强正线性趋势，但有一个点的 x 非常大而 y 非常小，那么计算出的 r 可能会变得弱得多，甚至变为负值。
11. Correlation Is Not Causation | 相关性不等于因果性
One of the most common errors in statistics is to conclude that because two variables are correlated, one causes the other. This conclusion is not justified by correlation alone.
统计学中最常见的错误之一，是因为两个变量相关就得出一个导致另一个的结论。仅凭相关不能证明因果。
There may be a hidden variable, sometimes called a confounding variable, that affects both x and y. For example, ice cream sales and drowning rates are positively correlated, but ice cream does not cause drowning. The confounding variable is hot weather.
可能存在一个隐藏变量，有时称为混杂变量，它同时影响 x 和 y。例如，冰淇淋销量与溺水率呈正相关，但冰淇淋并不会导致溺水。混杂变量是炎热天气。
Correlation can also be coincidental or due to chance. In examinations, when asked whether a high correlation proves causation, the correct answer is always no.
相关性也可能是巧合或偶然导致的。在考试中，当被问及高相关是否证明因果关系时，正确答案始终是否定的。
12. Exam Tips and Common Pitfalls | 考试技巧与常见误区
Always sketch or describe the scatter plot before calculating r. Examiners often award marks for recognising the direction and strength from a graph.
在计算 r 之前，务必先绘制或描述散点图。考官通常会根据你从图形中识别方向和强度来给分。
When using the formula, be careful with negative values of Sxy. A negative Sxy leads to a negative r. Also remember that r is always between −1 and 1. If your calculation gives a value outside this range, check your arithmetic.
使用公式时，注意 Sxy 可能为负值。负的 Sxy 会导致负的 r。还要记住 r 始终在 −1 和 1 之间。如果计算结果超出这个范围，请检查计算过程。
Do not confuse r with the regression coefficient. The regression slope b can be any real number, but the correlation coefficient r is always between −1 and 1.
不要把 r 与回归系数混淆。回归斜率 b 可以是任何实数，但相关系数 r 始终在 −1 和 1 之间。
Finally, interpret the value of r in context. Saying “r = 0.8” is not enough; you should state that there is a strong positive linear relationship between the two variables in the context of the question.
最后，要结合情境解释 r 的值。只说“r = 0.8”不够；你应当说明这两个变量在题目情境中存在强正线性关系。
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