📚 IB Mathematics: The Meaning and Calculation of Pearson's Correlation Coefficient | IB 数学：皮尔逊相关系数的含义与计算
In IB Mathematics (Analysis and Approaches or Applications and Interpretation), one of the most important statistical tools is Pearson's product-moment correlation coefficient, denoted by r. It measures the direction and strength of a linear relationship between two quantitative variables.
在IB数学（分析与方法或应用与解释）中，最重要的统计工具之一就是皮尔逊积矩相关系数，记作r。它衡量两个定量变量之间线性关系的方向与强度。
1. What is Correlation? | 什么是相关关系？
Correlation describes whether two variables move together. When one variable tends to increase as the other increases, we call this a positive association; when one increases while the other decreases, the association is negative.
相关关系描述两个变量是否一起变化。当一个变量随着另一个变量增加而增加时，我们称之为正相关；当一个变量增加而另一个减少时，则为负相关。
Pearson's r is a single number that summarises how closely the points on a scatter diagram follow a straight-line pattern.
皮尔逊r是一个单一数值，用于概括散点图上的点遵循直线模式的程度。
2. The Scatter Diagram | 散点图
Before computing r, always plot the data in a scatter diagram. The x-variable is plotted horizontally and the y-variable vertically. A scatter plot reveals whether a linear model is sensible and also identifies outliers.
在计算r之前，始终先将数据绘制成散点图。x变量画在水平方向，y变量画在垂直方向。散点图能揭示线性模型是否合理，也能识别异常值。
If the points form a narrow elliptical cloud sloping upward, r will be close to +1. If the cloud slopes downward, r will be close to −1. If the points form a random blob, r will be near 0.
如果点形成一个向上倾斜的窄椭圆云状，r将接近+1；如果云状向下倾斜，r将接近−1；如果点呈随机团状，r将接近0。
3. Covariance | 协方差
Pearson's r is based on covariance, which measures the joint variability of two variables. For sample data, the covariance is defined as:
皮尔逊r建立在协方差的基础上，协方差衡量两个变量的联合变异性。对于样本数据，协方差定义为：
sₓᵧ = Σ((xᵢ − x̄)(yᵢ − ȳ)) / (n − 1)
Here x̄ and ȳ are the sample means, and n is the number of pairs. A positive covariance indicates that x and y tend to deviate from their means in the same direction.
其中x̄和ȳ是样本均值，n是数据对的数量。正的协方差表示x和y倾向于偏离各自均值时方向相同。
However, covariance has units, so its magnitude is hard to interpret. Dividing by the product of the standard deviations removes units and creates a dimensionless coefficient: this is Pearson's r.
然而，协方差带有单位，因此其大小难以解释。将协方差除以标准差的乘积可以消除单位，得到一个无量纲的系数：这就是皮尔逊r。
4. Defining Pearson's r | 皮尔逊相关系数的定义
For a sample of n paired observations (x₁, y₁), (x₂, y₂), ..., (xₙ, yₙ), Pearson's correlation coefficient is:
对于n对观测值(x₁, y₁), (x₂, y₂), ..., (xₙ, yₙ)，皮尔逊相关系数定义为：
r = Σ((xᵢ − x̄)(yᵢ − ȳ)) / √(Σ(xᵢ − x̄)² Σ(yᵢ − ȳ)²)
This formula is symmetric: r is the same whether x is regressed on y or y is regressed on x. It only measures linear association, not a functional relationship in general.
这个公式是对称的：无论用x对y回归还是用y对x回归，r都相同。它只衡量线性关联，而非一般意义上的函数关系。
5. Properties of r | 相关系数的性质
The most important properties of Pearson's r are listed below:
皮尔逊r最重要的性质如下：
r always satisfies −1 ≤ r ≤ 1.
r始终满足−1 ≤ r ≤ 1。
erfect positive linear relationship: all points lie exactly on a line with positive slope.
r = 1表示完全正线性关系：所有点恰好落在一条正斜率的直线上。
erfect negative linear relationship: all points lie exactly on a line with negative slope.
r = −1表示完全负线性关系：所有点恰好落在一条负斜率的直线上。
r = 0 means no linear relationship, although a non-linear relationship may still exist.
r = 0表示没有线性关系，但仍可能存在非线性关系。
r is independent of the units of measurement; changing centimetres to metres does not change r.
r与测量单位无关；将厘米换成米不会改变r。
r is independent of the choice of which variable is x and which is y.
r不依赖于哪个变量选作x、哪个选作y。
6. Interpreting Strength and Direction | 强度与方向的解释
Although the exact interpretation depends on context, many IB textbooks use approximate bands for the strength of correlation:
虽然精确的解释取决于具体情境，许多IB教材使用近似的区间来衡量相关性强度：
| |r| value | Strength of linear correlation | 中文强度描述 |
| 0.00 – 0.19 | Very weak | 极弱 |
| 0.20 – 0.39 | Weak | 弱 |
| 0.40 – 0.59 | Moderate | 中等 |
| 0.60 – 0.79 | Strong | 强 |
| 0.80 – 1.00 | Very strong | 极强 |
The sign of r indicates direction: positive r means as one variable increases, the other tends to increase; negative r means the opposite. Note that the magnitude |r| measures strength, while the sign measures direction.
r的符号表示方向：正r意味着一个变量增加时另一个也倾向于增加；负r则相反。注意|r|的大小衡量强度，而符号表示方向。
7. Calculating r from Data | 由数据计算r
For IB exam questions, you may be asked to calculate r using summary statistics. A convenient computational formula is:
在IB考试题目中，你可能会需要利用汇总统计量来计算r。一个方便的计算公式是：
r = [Σxᵢyᵢ − (Σxᵢ)(Σyᵢ)/n] / √([Σxᵢ² − (Σxᵢ)²/n][Σyᵢ² − (Σyᵢ)²/n])
This form avoids calculating the means first and is often faster. You will need the sums Σx, Σy, Σxy, Σx², and Σy².
这个形式避免了先计算均值，通常更快。你需要求出总和Σx、Σy、Σxy、Σx²和Σy²。
8. Worked Example | 计算示例
Consider five students with hours studied (x) and test score (y):
考虑五名学生的学习小时数(x)与考试成绩(y)：
| x | 2 | 4 | 6 | 8 | 10 |
| y | 55 | 65 | 75 | 80 | 95 |
First compute the required sums:
首先计算所需的总和：
Σx = 30, Σy = 370, Σxy = 2×55 + 4×65 + 6×75 + 8×80 + 10×95 = 2480
Σx² = 4 + 16 + 36 + 64 + 100 = 220, Σy² = 3025 + 4225 + 5625 + 6400 + 9025 = 28300
Now substitute into the formula:
现在代入公式：
r = [2480 − (30)(370)/5] / √[220 − 900/5][28300 − 136900/5] = (2480 − 2220) / √((220 − 180)(28300 − 27380))
r = 260 / √40 × 920 = 260 / √36800 ≈ 260 / 191.83 ≈ 1.355
191.83, giving 260 / 191.83 ≈ 1.355. Since this is invalid, we must review the data. Wait — riginal sums are inconsistent with a correlation ≤ 1. Let us instead use a valid data set.
这个结果超过1，显然是不可能的。数值大于1说明计算有误；请检查运算。事实上重新计算：220 − 900/5 = 220 − 180 = 40，而28300 − 136900/5 = 28300 − 27380 = 920，因此40 × 920 = 36800，√36800 ≈ 191.83，于是260 / 191.83 ≈ 1.355。既然结果无效，我们必须重新审视数据。等等——260² = 67600，而36800小于67600，所以根号内的乘积太小；这说明原来给出的总和与相关系数≤1不相容。让我们改用一组有效的数据。
Let us choose a set where the linear relationship is not so perfect that arithmetic mistakes confuse the example. Consider:
让我们选择一组线性关系不至于太完美、以免运算错误造成困惑的数据。考虑：
| x | 1 | 2 | 3 | 4 | 5 |
| y | 2 | 4 | 5 | 4 | 6 |
Now compute the sums:
现在计算各总和：
Σx = 15, Σy = 21, Σxy = 2 + 8 + 15 + 16 + 30 = 71
Σx² = 1 + 4 + 9 + 16 + 25 = 55, Σy² = 4 + 16 + 25 + 16 + 36 = 97
Substitute into the formula:
代入公式：
r = [71 − (15)(21)/5] / √[55 − 225/5][97 − 441/5] = (71 − 63) / √((55 − 45)(97 − 88.2))
r = 8 / √10 × 8.8 = 8 / √88 ≈ 8 / 9.381 ≈ 0.853
The correlation is approximately 0.853, indicating a strong positive linear relationship between x and y.
相关系数约为0.853，表明x与y之间存在强的正线性关系。
9. The Coefficient of Determination r² | 决定系数r²
In IB courses, you will often encounter r², called the coefficient of determination. It represents the proportion of the total variation in one variable that is explained by its linear relationship with the other variable.
在IB课程中，你常会见到r²，称为决定系数。它表示一个变量的总变异中，可以由它与另一个变量的线性关系来解释的比例。
For r = 0.853, r² ≈ 0.728. This means about 72.8% the variation in y can be explained by the linear relationship with x; the remaining 27.2% is due to other factors or random variation.
对于r = 0.853，r² ≈ 0.728。这意味着y的变异中约72.8%可由与x的线性关系解释；其余27.2%由其他因素或随机变异造成。
10. Limitations and Pitfalls | 局限性与常见误区
Pearson' s. A perfect non-linear curve can give r = 0 even when the variables are strongly related.
皮尔逊r只能衡量线性关系。一条完美的非线性曲线即使变量间关系很强，也可能给出r = 0。
Outliers can dramatically change r. A single extreme point can inflate or deflate the correlation. Always inspect the scatter diagram before trusting r.
异常值会显著改变r。一个极端点就可能使相关系数变大或变小。在相信r之前，始终要检查散点图。
Moreover, r does not describe the slope of the relationship. A correlation of 0.9 may have a very small slope if the variables have different scales. The slope is a separate concept in linear regression.
此外，r不能描述关系的斜率。如果两个变量的尺度不同，0.9的相关系数可能对应很小的斜率。斜率是线性回归中另一个独立的概念。
Finally, combining two distinct groups in one data set can produce a misleading correlation. This is known as Simpson's paradox or the ecological fallacy.
最后，将两个不同的群体合并到同一数据集中会产生误导性的相关。这被称为辛普森悖论或生态学谬误。
11. Correlation vs Causation | 相关 ≠ 因果
A fundamental rule in statistics is that correlation does not imply causation. A high r may be due to chance, to reverse causation, or to a lurking variable that influences both x and y.
统计学中的基本法则是：相关并不意味着因果。较高的r可能来自偶然性、反向因果，或同时影响x和y的潜在混淆变量。
For example, ice cream sales and drowning deaths are highly correlated because both rise in summer; the weather is the lurking variable. The same logic applies to many IB exam contexts.
例如，冰淇淋销量与溺水死亡人数高度相关，因为两者都在夏季上升；天气是潜在的混淆变量。同样的逻辑适用于许多IB考试情境。
12. Using Technology | 使用技术工具
In IB examinations, a GDC (graphical display calculator) is often used to find r. Enter the data into lists, then use the two-variable statistics or linear regression function to read r.
在IB考试中，通常使用图形计算器(GDC)求r。将数据输入列表，然后使用双变量统计或线性回归功能读取r。
When using your calculator, make sure you choose the correct regression mode. In the linear regression output, r is displayed alongside r². If you do not see r, you may need to turn on the correlation coefficient display in the calculator settings.
使用计算器时，务必选择正确的回归模式。在线性回归输出中，r会与r²一起显示。如果没有看到r，你可能需要在计算器设置中开启相关系数显示功能。
13. Summary | 总结
Pearson's r is a dimensionless measure of linear correlation that always lies between −1 and 1. It is calculated by dividing covariance by the product of standard deviations, and it is symmetric, unit-free, and sensitive to outliers.
皮尔逊r是线性相关的无量纲度量，始终位于−1与1之间。它的计算方法是协方差除以标准差的乘积；它是对称的、无单位的，并且对异常值敏感。
Always visualise data first, calculate r carefully, interpret |r| as strength and the sign as direction, and never confuse correlation with causation. With these principles, you can confidently solve IB statistics questions on correlation.
务必先可视化数据，再仔细计算r，将|r|解释为强度、将符号解释为方向，并且绝不把相关混淆为因果。掌握这些原则，你就能自信地解决IB统计中关于相关的题目。
Published by TutorHao | 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