Congress of Data: Representation in Statistics | 数据之会：统计中的表示方法

📚 Congress of Data: Representation in Statistics | 数据之会：统计中的表示方法

The term “Congress” evokes the idea of bringing representatives together. In A-Level statistics, we gather raw data — our “delegates” — and must choose appropriate methods to represent them. Representation in data is not merely about drawing graphs; it is about revealing patterns, trends, and distributions that would otherwise remain hidden in raw numbers.

“Congress”一词使人联想到将代表们汇聚一堂。在A-Level统计学中，我们收集原始数据——即我们的”代表”——然后必须选择合适的方法来表示它们。数据的表示不仅仅是画图，而是揭示隐藏在原始数字背后的模式、趋势和分布规律。

---

1. Types of Data: Knowing Your Delegates | 数据类型：了解你的代表

Before representing data, you must classify the data type. Categorical (qualitative) data are non-numerical, such as colours or school subjects. Numerical (quantitative) data are measured or counted and further divided into discrete data (integer counts, e.g., number of students) and continuous data (measured on a scale, e.g., height or time).

在表示数据之前，必须对数据类型进行分类。分类（定性）数据是非数值型的，例如颜色或学科。数值（定量）数据通过测量或计数获得，进一步分为离散数据（整数计数，如学生人数）和连续数据（在标尺上测量，如身高或时间）。

Continuous data can take any value within a range, and this distinction determines which representation method is appropriate. For instance, a stem-and-leaf diagram works well for discrete data, while a histogram is suited to continuous data.

连续数据可以取某一范围内的任意值，这一区别决定了哪种表示方法适用。例如，茎叶图适合离散数据，而直方图适用于连续数据。

---

2. Frequency Tables and Grouped Data | 频数表与分组数据

A frequency table collects raw data into classes. For grouped continuous data, we record class intervals (e.g., 10 ≤ x < 20), class midpoints, and frequencies. The midpoint is calculated as (lower bound + upper bound) ÷ 2.

频数表将原始数据汇总到各个组别中。对于分组的连续数据，我们记录组区间（如10 ≤ x < 20）、组中点和频数。中点计算公式为（下界 + 上界）÷ 2。

The class width (or class interval width) is the difference between the upper and lower boundaries of a class. For example, in the interval 10 ≤ x < 20, the class width is 10. These values become critical when constructing histograms and calculating estimates for the mean.

组距是一个组的上边界与下边界之差。例如，区间10 ≤ x < 20的组距为10。这些数值在构建直方图和计算均值估计值时至关重要。

When data are grouped, individual values are lost; we estimate the mean using midpoints:

当数据分组后，个体数值会丢失；我们使用中点来估计均值：

x̄ = Σ(fᵢxᵢ) ÷ Σfᵢ

where fᵢ is the frequency of the i-th class and xᵢ is its midpoint.

其中 fᵢ 是第i组的频数，xᵢ 是其组中点。

---

3. Histograms: The Voice of Continuous Data | 直方图：连续数据的发声

A histogram is used for continuous data, especially when class intervals have unequal widths. Unlike a bar chart, there are no gaps between bars. The key formula is:

直方图用于连续数据，尤其当组区间宽度不相等时。与条形图不同，直方图的条之间没有间隙。关键公式为：

Frequency Density = Frequency ÷ Class Width

频数密度 = 频数 ÷ 组距

The vertical axis of a histogram is frequency density, not raw frequency. The area of each bar represents the frequency:

直方图的纵轴是频数密度，而不是原始频数。每个条形的面积代表频数：

Frequency = Frequency Density × Class Width

频数 = 频数密度 × 组距

Height (cm)	Frequency	Class Width	Frequency Density
140 ≤ h < 150	20	10	20 ÷ 10 = 2
150 ≤ h < 155	15	5	15 ÷ 5 = 3
155 ≤ h < 170	30	15	30 ÷ 15 = 2

When reading a histogram, remember: the total area under the histogram equals the total frequency, Σf.

在读直方图时，记住：直方图下的总面积等于总频数Σf。

---

4. Box Plots: A Quintet of Summary Statistics | 箱线图：五个汇总统计量的合奏

A box plot (box-and-whisker diagram) summarises a data set using five key values: the minimum, lower quartile (Q₁), median (Q₂), upper quartile (Q₃), and maximum. This five-number summary provides a compact visual representation of the data’s spread and centre.

箱线图（箱须图）使用五个关键值来概括数据集：最小值、下四分位数（Q₁）、中位数（Q₂）、上四分位数（Q₃）和最大值。这个五数概括提供了数据分布和中心趋势的紧凑可视化表示。

The interquartile range (IQR) is calculated as:

四分位距（IQR）的计算公式为：

IQR = Q₃ − Q₁

The IQR measures the middle 50% of the data and is a robust measure of spread, unaffected by extreme outliers.

IQR衡量数据的中间50%，是不受极端离群值影响的稳健离散度量。

To determine quartiles from a data set, arrange values in ascending order. If n is the number of observations, Q₁ is the value at the (n+1)/4-th position, and Q₃ is at the 3(n+1)/4-th position — using interpolation when necessary.

要确定数据集的四分位数，先将数据按升序排列。若n为观测值个数，Q₁位于第(n+1)/4个位置，Q₃位于第3(n+1)/4个位置——必要时使用插值法。

---

5. Stem-and-Leaf Diagrams: A Congress in Two Columns | 茎叶图：两列中的代表大会

A stem-and-leaf diagram organises data by splitting each value into a “stem” (leading digit(s)) and a “leaf” (final digit). It preserves original data values while showing the full distribution shape.

茎叶图将每个数值拆分为”茎”（前导数字）和”叶”（最后一位数字）来组织数据。它在展示整体分布形状的同时保留了原始数据值。

For example, test scores 42, 37, 51, 48, 44, 39 can be displayed as:

例如，考试成绩42、37、51、48、44、39可表示为：

3 | 7 9
4 | 2 4 8
5 | 1

Key: 4 | 2 means 42. To compare two data sets, use a back-to-back stem-and-leaf diagram. Always include a key, and always order leaves from smallest to largest.

图例：4 | 2表示42。要比较两个数据集，可以使用背靠背茎叶图。务必包含图例，并始终将叶子从小到大排列。

---

6. Cumulative Frequency Graphs: The March of Accumulation | 累积频数图：累积的行军

A cumulative frequency graph plots cumulative frequency against the upper class boundary. The points are joined with a smooth curve (or straight lines for a cumulative frequency polygon).

累积频数图以累积频数对组上边界作图。各点用平滑曲线连接（或用直线连接形成累积频数折线图）。

From this graph, you can estimate the median, quartiles, and percentiles. The median corresponds to the value at cumulative frequency n/2, Q₁ at n/4, and Q₃ at 3n/4.

从该图中，可以估计中位数、四分位数和百分位数。中位数对应累积频数为n/2处的值，Q₁对应n/4处，Q₃对应3n/4处。

To estimate the number of values below a given threshold x, read the cumulative frequency directly from the y-axis at x. This is particularly useful when answering questions about “how many students scored below 60 marks” or similar.

要估计低于某阈值x的数值个数，直接在x处从y轴读取累积频数。这在回答”有多少学生得分低于60分”之类的问题时特别有用。

---

7. Scatter Diagrams: Representing Bivariate Data | 散点图：表示二元数据

When two variables are measured for the same individual, we use a scatter diagram to represent the relationship. Each point represents an (x, y) pair. The pattern of points reveals the correlation:

当对同一个体测量两个变量时，我们使用散点图来表示两者之间的关系。每个点代表一个(x, y)坐标对。点的分布模式揭示了相关性：

• Positive correlation: as x increases, y tends to increase (points slope upward).

  正相关：随着x增大，y趋向增大（点向上倾斜）。

• Negative correlation: as x increases, y tends to decrease (points slope downward).

  负相关：随着x增大，y趋向减小（点向下倾斜）。

• No correlation: no apparent trend in the points.

  无相关：点中无明显趋势。

The strength of the linear relationship is quantified by Pearson’s product-moment correlation coefficient, r, which lies in the range −1 ≤ r ≤ 1. Values close to 1 indicate strong positive correlation; values close to −1 indicate strong negative correlation; r ≈ 0 indicates no linear correlation.

线性关系的强度由皮尔逊积矩相关系数r来量化，其取值范围为−1 ≤ r ≤ 1。接近1的值表示强正相关；接近−1的值表示强负相关；r ≈ 0表示无线性相关。

r = Sₓᵧ ÷ √(SₓₓSᵧᵧ)

where Sₓᵧ = Σxy − (ΣxΣy)/n, Sₓₓ = Σx² − (Σx)²/n, and Sᵧᵧ = Σy² − (Σy)²/n.

其中 Sₓᵧ = Σxy − (ΣxΣy)/n，Sₓₓ = Σx² − (Σx)²/n，Sᵧᵧ = Σy² − (Σy)²/n。

---

8. Outliers: The Rebels in the Congress | 离群值：大会中的异见者

An outlier is a data point that is unusually large or small compared with the rest of the data. In Edexcel A-Level, the standard definition for an outlier is:

离群值是相比数据主体异常大或异常小的数据点。在Edexcel A-Level中，离群值的标准定义为：

Outlier < Q₁ − 1.5 × IQR or Outlier > Q₃ + 1.5 × IQR

离群值 < Q₁ − 1.5 × IQR 或 离群值 > Q₃ + 1.5 × IQR

When drawing box plots, outliers are plotted as individual points beyond the whiskers. The whiskers then extend to the smallest and largest values that are not outliers. Remember to comment on the effect of outliers on the mean (pulled toward the outlier) versus the median (largely resistant).

绘制箱线图时，离群值作为独立点画在须的外侧。须则延伸至非离群的最小值和最大值。记得讨论离群值对均值（被拉向离群值方向）和中位数（基本不受影响）的不同影响。

---

9. Choosing the Appropriate Representation | 选择恰当的表示方法

Edexcel examiners often ask why a particular representation is preferred. Here is a practical guide:

Edexcel考官经常问为什么选某种表示方法。这里提供一份实用指南：

Representation | 表示方法	Best Used For | 最佳用途
Histogram | 直方图	Continuous data, unequal class widths | 连续数据、不等组距
Box plot | 箱线图	Comparing distributions, showing outliers | 比较分布、显示离群值
Stem-and-leaf | 茎叶图	Small data sets, preserving raw values | 小数据集、保留原始值
Cumulative frequency | 累积频数图	Estimating median, quartiles, percentiles | 估计中位数、四分位数、百分位数
Scatter diagram | 散点图	Showing correlation between two variables | 展示两个变量的相关性

A histogram is preferred over a bar chart when data are continuous. A box plot is preferred when comparing two or more distributions because it allows side-by-side comparison of quartiles and outliers.

当数据为连续型时，直方图优于条形图。当比较两个或多个分布时，箱线图更优，因为它可以并排比较四分位数和离群值。

---

10. Common Pitfalls and Exam Strategies | 常见误区与应试策略

In the Edexcel exam, students lose marks on data representation questions due to several recurring errors. First, forgetting to use frequency density on the y-axis of a histogram is the single most common mistake. Second, when drawing a box plot, students often misidentify the quartile positions, especially when n is even or odd.

在Edexcel考试中，学生在数据表示题目上失分主要源于几个反复出现的错误。首先，忘记在直方图的y轴使用频数密度是最常见的错误。其次，绘制箱线图时，学生经常错误定位四分位数的位置，尤其是当n为偶数或奇数时。

When plotting a cumulative frequency curve, always plot the cumulative frequency against the upper class boundary, not the midpoint. Also, remember that the curve must always be non-decreasing. For scatter diagrams, if you are asked to interpret r, use phrases such as “a strong positive linear correlation between x and y” — a full sentence with context.

绘制累积频数曲线时，始终以组上边界（而非中点）对应累积频数作图。另外，记住曲线必须始终非递减。对于散点图，如果要求解释r，应使用完整的句子并联系上下文，例如”x与y之间存在强正线性相关”。

Finally, always show your working when calculating statistics from grouped data. Use the formula sheet provided by Edexcel, and where possible, check that your median estimate from a cumulative frequency graph is consistent with the box plot you draw.

最后，在分组数据计算统计量时，务必展示你的计算过程。使用Edexcel提供的公式表，并尽可能检查从累积频数图中估计的中位数与你绘制的箱线图一致。

---

11. Worked Example: A Mini Congress in Action | 例题精讲：一次小型数据大会

Problem: The masses (in grams) of 30 bags of rice are grouped as follows: 950 ≤ m < 970 (f = 4), 970 ≤ m < 990 (f = 8), 990 ≤ m < 1000 (f = 12), 1000 ≤ m < 1030 (f = 6). Construct a histogram and estimate the mean mass.

问题：30袋大米的重量（克）分组如下：950 ≤ m < 970（频数=4），970 ≤ m < 990（频数=8），990 ≤ m < 1000（频数=12），1000 ≤ m < 1030（频数=6）。构建直方图并估计平均重量。

Solution: First, compute frequency densities:

解答：首先计算频数密度：

Class | 组别	Width | 组距	Frequency | 频数	Freq Density | 频数密度	Midpoint | 中点
950–970	20	4	0.2	960
970–990	20	8	0.4	980
990–1000	10	12	1.2	995
1000–1030	30	6	0.2	1015

Estimate of the mean:

均值的估计：

x̄ = (4×960 + 8×980 + 12×995 + 6×1015) ÷ 30

= (3840 + 7840 + 11940 + 6090) ÷ 30 = 29710 ÷ 30 ≈ 990.3 g

For the histogram, each bar’s height equals its frequency density, and its width spans the class interval. Always label axes as “Frequency Density” and the relevant quantity (e.g., “Mass (g)”).

对于直方图，每个条的高度等于其频数密度，宽度跨越组区间。始终将坐标轴标注为”频数密度”和相应量（如”重量(g)”）。

---

12. Final Review: Bringing It All Together | 总结回顾：融会贯通

Representation in statistics is a congress of ideas: each method has its role, its strengths, and its appropriate context. A well-chosen representation not only answers the exam question but also communicates the data story clearly.

统计中的表示是思想的汇聚：每种方法都有其角色、优势和适用的情境。精心选择的表示方法不仅能解答考试题目，还能清晰地传达数据背后的故事。

Key takeaways for Edexcel A-Level Mathematics:

Edexcel A-Level数学的关键要点：

• Always classify data as categorical, discrete, or continuous before choosing a representation.

  在选择表示方法之前，始终先将数据分类为定性的、离散的或连续的。

• Use frequency density for histograms when class widths are unequal.

  当组距不相等时，直方图使用频数密度。

• Box plots and cumulative frequency graphs are powerful tools for quartiles, medians, and outliers.

  箱线图和累积频数图是处理四分位数、中位数和离群值的强大工具。

• Scatter diagrams plus the correlation coefficient r reveal bivariate relationships.

  散点图结合相关系数r揭示二元变量之间的关系。

• Show clear working and label every graph fully — marks are awarded for presentation.

  展示清晰的计算过程并完整标注每个图形——绘图规范可以得分。

Mastering data representation is not merely an exam skill; it is the foundation of statistical literacy. When data come together in congress, the right representation gives every number a voice.

掌握数据表示不仅是一项考试技能，更是统计素养的基础。当数据汇聚一堂时，恰当的表示让每个数字都有发声的机会。

Published by TutorHao | Mathematics Revision Series | aleveler.com

更多咨询请联系16621398022（同微信）