Comparing Data Sets Using Statistical Measures — 数据比较：使用统计量进行有效对比

📚 Comparing Data Sets Using Statistical Measures | 数据比较：使用统计量进行有效对比

在 A-Level 数学（Edexcel Statistics 部分）中，比较两组或多组数据是考试的核心题型之一。单独看一组数据的平均值或极差远远不够，真正有效的比较需要同时考虑中心趋势（center）与离散程度（spread），并且要根据数据的分布形态选择恰当的统计量。这篇文章将系统梳理：均值、中位数、众数、极差、四分位距、方差与标准差各自的适用场景，箱线图与偏态判断的方法，编码数据（coding）对统计量的影响，以及考试中常见的陷阱与答题技巧。

In A-Level Mathematics (Edexcel Statistics component), comparing two or more data sets is one of the core question types in the exam. Looking at a single average or range in isolation is never enough: an effective comparison must consider both the centre and the spread of the data, and you must choose the right statistic according to the shape of the distribution. This article systematically covers: when to use the mean, median, mode, range, interquartile range, variance and standard deviation; how to compare distributions with box plots and skewness; how coding (linear transformations) affects statistics; and the common exam pitfalls with answering techniques.

一、为什么不能只看平均值：中心趋势与离散度的双重视角 | Why Averages Alone Are Not Enough: The Dual Lens of Centre and Spread

假设两个班级的数学测验平均分都是 62 分，这是否意味着两个班的成绩表现完全相同？答案显然是否定的。甲班可能所有学生都集中在 60 到 64 分之间，而乙班可能一半学生考了 95 分、另一半只考了 30 分。平均值相同，但数据的”形状”截然不同。这正是统计学家反复强调的观点：一个统计量只能描述数据的一个侧面，全面比较至少需要两个维度，即中心趋势（数据集中在哪个位置）和离散程度（数据分散得有多开）。

Suppose two classes both have a mean score of 62 on a maths test. Does that mean the two classes performed identically? Clearly not. Class A might have every student clustered between 60 and 64, while Class B might have half the students scoring 95 and the other half scoring 30. The means are the same, yet the shapes of the two data sets are completely different. This is the point statisticians constantly emphasise: a single statistic describes only one facet of the data, and a full comparison needs at least two dimensions, namely the central tendency (where the data are located) and the spread (how widely the data are dispersed).

在 A-Level 考试中，比较题的标准答题结构通常包含三步：第一步，分别计算两组数据的中心趋势量；第二步，分别计算两组数据的离散度量；第三步，结合上下文解释这些数值意味着什么，例如”乙班平均分更高，说明整体水平更好；但乙班标准差更大，说明学生之间差异也更大”。只写数值不给解释，通常会丢掉一半以上的分数。

In the A-Level exam, the standard structure for a comparison question has three steps: first, calculate a measure of central tendency for each data set; second, calculate a measure of spread for each data set; third, interpret what the values mean in context, for example “Class B has a higher mean, so its overall level is better; but Class B also has a larger standard deviation, so its students differ from one another more.” Writing numbers without interpretation usually loses more than half the marks.

二、三种中心趋势量：均值、中位数与众数的选择原则 | Three Measures of Central Tendency: When to Use the Mean, Median and Mode

均值（mean）是全部数据相加后除以数据个数，数学上记为 x̄ = Σx / n。均值的最大优点是利用了所有数据的信息，计算精确；它的最大缺点是容易受极端值（outliers）影响。例如一组数据 2, 3, 4, 5, 96，均值为 22，这个数值显然不能代表大多数数据。中位数（median）是把数据从小到大排序后位于正中间的值，它只取决于排序位置，因此对极端值不敏感，适合偏态分布或含有离群值的数据。众数（mode）是出现频率最高的数值，适用于描述定性数据或离散数据中最常见的类别，但在连续数据中往往没有意义，因为每个值都可能只出现一次。

The mean is the sum of all data values divided by the number of values, written as x̄ = Σx / n. Its greatest advantage is that it uses information from every data point and is mathematically precise; its greatest weakness is that it is easily distorted by extreme values (outliers). For example, for the data 2, 3, 4, 5, 96, the mean is 22, a figure that clearly does not represent most of the data. The median is the middle value when the data are arranged in ascending order; it depends only on position in the ranking, so it is insensitive to extreme values and is therefore suitable for skewed distributions or data containing outliers. The mode is the value that occurs most frequently; it is useful for describing categorical data or the most common category in discrete data, but it is often meaningless for continuous data because every value may occur only once.

选择原则可以总结为：数据对称且无离群值时优先用均值，因为它信息量最大；数据偏斜或存在离群值时用中位数，因为它稳健；需要描述”最常见情况”时用众数，例如调查学生最常选的科目。考试中经常出现一道小题：给定一组数据，要求判断哪个中心趋势量最适合，并给出理由。回答时要同时说明”数据是否有离群值”以及”分布是否对称”。

The selection rule can be summarised as follows: use the mean when the data are symmetric and free of outliers, because it carries the most information; use the median when the data are skewed or contain outliers, because it is robust; use the mode when you need to describe the “most common” case, such as the subject most students choose. A common exam question asks you to decide which measure of central tendency is most appropriate for a given data set and to justify your choice. In your answer you must comment on both whether outliers are present and whether the distribution is symmetric.

三、离散度三件套：极差、四分位距与标准差的区别 | The Three Spread Measures: Range, Interquartile Range and Standard Deviation

极差（range）是最大值减最小值，计算最简单，但只用了两个数据点，极易受单个离群值影响。四分位距（IQR）是上四分位数 Q3 减去下四分位数 Q1，即中间 50% 数据的宽度，它剔除了两端的极端值，因此与中位数搭配使用非常稳健。标准差（standard deviation）是方差（variance）的平方根，它衡量每个数据偏离均值的平均程度，是所有离散度量中信息量最大的一个，但与均值一样容易受极端值影响。

The range is the maximum value minus the minimum value. It is the simplest to calculate but uses only two data points and is extremely sensitive to a single outlier. The interquartile range (IQR) is the upper quartile Q3 minus the lower quartile Q1, that is, the width of the middle 50% of the data; it discards the extreme values at both ends, so it pairs robustly with the median. The standard deviation is the square root of the variance; it measures the average distance of each data value from the mean. It carries the most information of all the spread measures, but like the mean, it is affected by extreme values.

记忆口诀：均值配标准差，中位数配四分位距。当你在比较题中使用了中位数，那么离散度就应该用 IQR；如果你使用了均值，那么离散度就应该用标准差。这种”配套使用”的原则在 Edexcel 评分方案中反复出现，混搭（例如用中位数配标准差）虽然不算错，但往往不是最合适的组合，解释起来也缺乏逻辑一致性。

A useful rule of thumb: the mean goes with the standard deviation, and the median goes with the interquartile range. When you use the median in a comparison question, you should report the IQR as the spread; when you use the mean, you should report the standard deviation. This pairing principle appears again and again in Edexcel mark schemes. Mixing them (for example, median with standard deviation) is not strictly wrong, but it is usually not the most appropriate combination and is harder to justify logically.

四、方差与标准差的计算：未分组数据与分组数据 | Variance and Standard Deviation: Ungrouped and Grouped Data

未分组数据的方差公式有两种等价写法：Var(X) = Σ(x – x̄)² / n 与 Var(X) = Σx² / n – x̄²。第二种写法（展开式）在计算时更实用，因为它只需要累加 x 与 x² 两列。标准差则是方差的算术平方根。注意 Edexcel 考试中，如果数据被视为”样本”（sample），分母用 n – 1；如果被视为”总体”（population），分母用 n。题目通常会用词语暗示：从一批产品中”抽取”的数据是样本，全部学生的成绩则是总体。

For ungrouped data the variance has two equivalent forms: Var(X) = Σ(x – x̄)² / n and Var(X) = Σx² / n – x̄². The second (expanded) form is more practical for calculation because you only need to accumulate two columns, x and x². The standard deviation is the positive square root of the variance. Note that in the Edexcel exam, if the data are treated as a sample, the denominator is n – 1; if they are treated as the whole population, the denominator is n. The question wording usually gives the clue: data “sampled” from a batch of products are a sample, whereas the scores of all students in a school are the population.

分组数据（grouped data）通常以频数表形式给出，例如成绩区间 50-59、60-69 等。此时我们不知道每个原始值，只能用各区间的组中值（midpoint）x 近似代替，方差公式变为 Var ≈ Σfx² / Σf – (Σfx / Σf)²。注意：分组数据算出的均值与标准差只是近似值，因为组内数据的实际分布未知。Edexcel 考试常考”从频数表求均值和标准差”的大题，步骤固定：先补全 x、fx、fx² 三列，再代入公式。

Grouped data are usually presented in a frequency table, for example score intervals 50-59, 60-69, and so on. Since the original values are unknown, each interval is represented by its midpoint x, and the variance becomes Var ≈ Σfx² / Σf – (Σfx / Σf)². The mean and standard deviation obtained from grouped data are approximations, because the actual distribution within each interval is unknown. Edexcel frequently sets multi-part questions on finding the mean and standard deviation from a frequency table; the procedure is fixed: complete the three columns x, fx and fx², then substitute into the formula.

数据形式	均值公式	方差公式
未分组	x̄ = Σx / n	Σx² / n – x̄²
分组（频数表）	x̄ = Σfx / Σf	Σfx² / Σf – x̄²

五、箱线图：一张图对比两组数据的分布 | Box Plots: Comparing Two Distributions in a Single Diagram

箱线图（box plot，又称箱须图 box-and-whisker diagram）用五个关键数概括一组数据：最小值、Q1、中位数、Q3、最大值。画箱线图时，先按从小到大排序数据并求出五个数，然后画一条数轴，标出五点的位置，用矩形连接 Q1 与 Q3，在中位数处画一条竖线，再用两条须（whisker）连接矩形两端到最小值和最大值。Edexcel 要求能够从原始数据或频数表画出箱线图，也要能从箱线图反推出五个关键数。

A box plot (also called a box-and-whisker diagram) summarises a data set with five key numbers: the minimum, Q1, the median, Q3 and the maximum. To draw one, first sort the data and find the five numbers, then draw a number line, mark the five positions, join Q1 and Q3 with a rectangle, draw a vertical line at the median, and extend two whiskers from the box to the minimum and maximum. Edexcel requires you to draw a box plot from raw data or a frequency table, and also to read the five key numbers back from a given box plot.

箱线图在比较题中的价值在于”并排对比”：把两组数据的箱线图画在同一数轴上，一眼就能看出谁的中间 50% 更集中、谁的中位数更高、谁的数据范围更宽、谁存在更长的尾巴（偏态）。考试典型问法：”比较这两个箱线图，说明哪个班级成绩更好。”标准答法：中位数更高的一组整体更强；箱体更窄的一组更稳定、学生水平更一致；须更长的一端提示存在极端值或偏态。

The value of box plots in comparison questions lies in side-by-side comparison: when two box plots are drawn on the same axis, you can immediately see whose middle 50% is more concentrated, whose median is higher, whose data range is wider, and whose tail is longer (skewness). A typical exam question asks: “Compare these two box plots and state which class performed better.” The standard answer: the group with the higher median is stronger overall; the group with the narrower box is more stable and consistent; a longer whisker suggests extreme values or skewness.

六、百分位数与四分位数：位置型统计量的比较作用 | Percentiles and Quartiles: Positional Measures in Comparison

四分位数把排序后的数据分成四等份：Q1 是第 25 百分位数，Q2 就是中位数（第 50 百分位数），Q3 是第 75 百分位数。Edexcel 中四分位数的计算有多种约定：当数据个数为奇数时，常用”去掉中位数后取两半各自的中位数”的方法；也有的题目直接用 (n+1)/4 的位置插值。考试以题目给出的方法为准，不必纠结约定差异，但自己计算时务必写清步骤。

Quartiles divide sorted data into four equal parts: Q1 is the 25th percentile, Q2 is the median (50th percentile), and Q3 is the 75th percentile. Edexcel uses several conventions for quartiles: when the number of data values is odd, a common method is to remove the median and take the median of each half; some questions instead interpolate at position (n+1)/4. In the exam, follow the method stated in the question; do not worry about convention differences, but always show your working clearly.

百分位数（percentile）在实际比较中非常有用，例如”某学生成绩位于第 90 百分位数”意味着他超过 90% 的考生。在比较两组数据时，百分位数可以回答均值无法回答的问题：最高端的差距有多大？最低端的差距有多大？例如两个班级中位数相同，但甲班第 90 百分位数明显更高，说明甲班的尖子生更强。考试常要求从累积频率图（cumulative frequency graph）读出中位数与四分位数，再据此比较。

Percentiles are very useful in real comparisons. For example, “a student’s score is at the 90th percentile” means he outperformed 90% of the candidates. When comparing two data sets, percentiles can answer questions the mean cannot: how large is the gap at the top end? How large is the gap at the bottom end? Two classes may have the same median, but if Class A has a clearly higher 90th percentile, its top students are stronger. The exam often asks you to read the median and quartiles from a cumulative frequency graph and then compare the two groups.

七、离群值的识别与处理：何时剔除数据点 | Identifying and Handling Outliers: When to Exclude Data Points

离群值（outlier）是与数据主体明显偏离的极端值。Edexcel 最常用的判定规则是 1.5 倍 IQR 规则：小于 Q1 – 1.5×IQR 或大于 Q3 + 1.5×IQR 的数据点视为离群值。另一条常见规则是 2 倍标准差规则：与均值的距离超过 2 个标准差的点视为离群值（不同考试局标准略有差异，以题目说明为准）。识别离群值是不少学生的失分点，因为需要先正确求出四分位数或标准差，再代入不等式判断。

An outlier is an extreme value that deviates markedly from the main body of the data. The most commonly used rule in Edexcel is the 1.5 × IQR rule: any value less than Q1 – 1.5 × IQR or greater than Q3 + 1.5 × IQR is treated as an outlier. Another common rule is the 2 standard deviations rule: a point more than two standard deviations from the mean is an outlier (standards vary slightly between boards; follow the wording of the question). Identifying outliers is a frequent source of lost marks, because you must first compute the quartiles or the standard deviation correctly and then substitute into the inequalities.

识别出离群值之后怎么办？这是比较题的高阶考点。若题目要求”考虑离群值的影响”，标准说法是：离群值会拉高（或拉低）均值与标准差，但对中位数和 IQR 影响很小，因此在比较时应说明”剔除离群值后，均值更接近大多数数据的水平”；若题目明确说”剔除离群值后重新计算”，则需要去掉该数据点并重算均值、标准差等。注意：箱线图中离群值通常单独用星号或小圆点标出，须只延伸到最后一个非离群值。

What should you do once an outlier is identified? This is an advanced point in comparison questions. If the question asks you to “consider the effect of the outlier”, the standard statement is: the outlier pulls the mean and standard deviation up (or down), but has little effect on the median and IQR, so in the comparison you should note that “after removing the outlier, the mean is closer to the level of the majority of the data”. If the question explicitly says “remove the outlier and recalculate”, you must drop that data point and recompute the mean, standard deviation and so on. Note that in box plots outliers are usually marked separately with an asterisk or a dot, and the whisker extends only to the last non-outlier value.

八、对称与偏态：从分布形状判断该信哪个统计量 | Symmetric and Skewed Distributions: Which Statistic to Trust

分布的形状决定统计量的可信度。对称分布（symmetric distribution）中，均值、中位数、众数三者几乎重合，此时均值是最优的中心趋势量。正偏分布（positively skewed，右偏）中，长尾巴拖向右侧，此时均值被少数大值拉高，均值大于中位数大于众数，应该用中位数代表”典型水平”。负偏分布（negatively skewed，左偏）则相反，均值小于中位数，常见于”考试分数普遍偏高、少数人很低”的情形。

The shape of a distribution determines which statistic you can trust. In a symmetric distribution, the mean, median and mode nearly coincide, and the mean is the best measure of central tendency. In a positively skewed distribution, the long tail extends to the right; the mean is pulled up by a few large values, so mean > median > mode, and you should use the median to represent the “typical” level. A negatively skewed distribution is the opposite: the mean is less than the median, which is common when “most scores are high and a few are very low”.

Edexcel 要求会用两种方法判断偏态方向。方法一：比较均值与中位数的大小（均值大于中位数则正偏）。方法二：皮尔逊偏度系数 Skew = 3(均值 – 中位数) / 标准差，系数为正则正偏，为负则负偏，绝对值越大偏斜越严重。箱线图也能直观判断：正偏时中位数靠近箱体左侧、右侧须更长；负偏时相反。判断偏态后，比较题的解释就要相应调整：正偏数据说”中位数更能代表典型水平，因为少数高分拉高了均值”。

Edexcel requires you to determine the direction of skewness in two ways. Method one: compare the mean and the median (if the mean is greater than the median, the distribution is positively skewed). Method two: Pearson’s coefficient of skewness, Skew = 3(mean – median) / standard deviation; a positive coefficient means positive skew, a negative coefficient means negative skew, and the larger the absolute value, the more severe the skew. Box plots also show skew visually: positive skew places the median near the left of the box with a longer right whisker; negative skew is the reverse. Once you identify the skew, adjust your comparison language accordingly: for positively skewed data, say “the median better represents the typical level, because a few high scores inflate the mean”.

九、编码数据：线性变换如何改变统计量 | Coding Data: How Linear Transformations Change the Statistics

编码（coding）是 Edexcel 统计部分的必考技巧。当原始数据 x 较大或较繁琐时，可以令 y = (x – a) / b（常用如 y = (x – 100) / 10），先计算 y 的均值与方差，再反推 x 的统计量。核心结论：均值满足线性关系，即 x̄ = a + b·ȳ；方差满足 Var(X) = b²·Var(Y)；标准差满足 σx = b·σy（注意 b 取正值）。中位数、四分位数等位置型统计量也按均值的同样方式变换：Qx = a + b·Qy。

Coding is a compulsory technique in the Edexcel statistics component. When the original data x are large or awkward, you can define y = (x – a) / b (commonly y = (x – 100) / 10), compute the mean and variance of y first, then convert back to the statistics of x. The core results are: the mean follows the linear relation x̄ = a + b·ȳ; the variance transforms as Var(X) = b²·Var(Y); and the standard deviation transforms as σx = b·σy (taking b positive). Positional measures such as the median and quartiles transform in the same way as the mean: Qx = a + b·Qy.

编码技巧的考试价值：第一，大幅简化手算，例如把 195, 205, 210 这类数据编码成 y = (x – 200) / 5 后变成 -1, 1, 2，计算量骤减；第二，检验理解深度，题目常反着问：”已知编码后的均值和方差，求原始数据的均值和方差”，此时只要代入上述反变换公式即可。常见错误是把方差也按 b 的一次方变换，忘记方差要乘 b²。记住口诀：平移不影响离散度，缩放才影响，且方差按比例平方缩放。

The exam value of coding is twofold. First, it dramatically simplifies hand calculation: data such as 195, 205, 210 become -1, 1, 2 under y = (x – 200) / 5, cutting the arithmetic sharply. Second, it tests depth of understanding: questions often ask in reverse, “given the mean and variance of the coded data, find the mean and variance of the original data”, which only requires substituting into the inverse transformation. A common error is transforming the variance with b to the first power, forgetting that the variance scales by b². Remember the rule of thumb: translation does not affect spread, only scaling does, and variance scales by the square of the scale factor.

十、完整例题：比较两个班级的成绩 | Worked Example: Comparing the Scores of Two Classes

例题：甲班 10 名学生测验成绩为 45, 52, 58, 60, 62, 64, 66, 68, 70, 75；乙班 10 名学生成绩为 30, 55, 58, 60, 62, 64, 66, 68, 72, 95。要求：(a) 求两班各自的均值、中位数、标准差；(b) 比较两班成绩并说明理由。先看甲班：数据已排序，中位数为 (62+64)/2 = 63；均值为 620/10 = 62；方差用展开式 Σx²/n – x̄² 计算，Σx² = 45² + 52² + … + 75² = 39402，方差 = 39402/10 – 62² = 3940.2 – 3844 = 96.2，标准差约 9.81。

Example: Class A of 10 students scored 45, 52, 58, 60, 62, 64, 66, 68, 70, 75; Class B of 10 students scored 30, 55, 58, 60, 62, 64, 66, 68, 72, 95. Tasks: (a) find the mean, median and standard deviation of each class; (b) compare the two classes with justification. Class A first: the data are already sorted, so the median is (62+64)/2 = 63; the mean is 620/10 = 62. For the variance use the expanded form Σx²/n – x̄²: Σx² = 45² + 52² + … + 75² = 39402, so variance = 39402/10 – 62² = 3940.2 – 3844 = 96.2, and the standard deviation is about 9.81.

再看乙班：均值为 630/10 = 63，中位数仍为 63，但注意乙班存在极端值 30 和 95。Σx² = 30² + 55² + … + 95² = 42754，方差 = 42754/10 – 63² = 4275.4 – 3969 = 306.4，标准差约 17.50。比较结论：(i) 乙班均值 63 略高于甲班 62，整体水平略好；(ii) 但乙班标准差 17.50 远大于甲班 9.81，说明乙班内部差异大得多，成绩两极分化严重；(iii) 乙班的中位数与均值接近，但分布存在明显离群值（30 与 95），因此用中位数加 IQR 描述乙班更稳健。若用 1.5×IQR 规则检验：乙班 Q1 = 58, Q3 = 68, IQR = 10，离群下界 = 58 – 15 = 43，因此 30 确实是离群值。

Now Class B: the mean is 630/10 = 63 and the median is still 63, but note the extreme values 30 and 95. Σx² = 30² + 55² + … + 95² = 42754, so variance = 42754/10 – 63² = 4275.4 – 3969 = 306.4 and the standard deviation is about 17.50. Comparison conclusions: (i) Class B has a slightly higher mean of 63 against Class A’s 62, so its overall level is marginally better; (ii) but Class B’s standard deviation of 17.50 is far larger than Class A’s 9.81, showing much greater internal variation and polarisation; (iii) Class B’s median and mean are close, yet the distribution contains clear outliers (30 and 95), so the median with the IQR describes Class B more robustly. Testing with the 1.5 × IQR rule: for Class B, Q1 = 58, Q3 = 68, IQR = 10, and the lower fence is 58 – 15 = 43, so 30 is indeed an outlier.

十一、实际应用：用统计量比较两个生产过程 | Real-World Application: Comparing Two Production Processes

统计量的比较能力不仅用于考试，也是真实世界中质量管理的基础。例如两家工厂生产同一规格的螺栓，标称直径 10 mm。工厂 X 抽样测得均值 10.01 mm，标准差 0.02 mm；工厂 Y 均值 10.00 mm，标准差 0.15 mm。从数据看：工厂 X 的均值略偏大，但标准差极小，说明产品高度一致，几乎全部落在公差范围内；工厂 Y 均值虽然更接近标称值，但标准差大 7.5 倍，说明大量产品可能超出公差，废品率更高。结论：单看均值，工厂 Y 似乎更好；结合标准差，工厂 X 的质量控制明显更优。

The power of comparing statistics extends beyond exams into quality control in the real world. Two factories produce bolts of the same specification with a nominal diameter of 10 mm. Factory X samples bolts with a mean of 10.01 mm and a standard deviation of 0.02 mm; Factory Y has a mean of 10.00 mm and a standard deviation of 0.15 mm. Reading the data: Factory X’s mean is slightly high, but its standard deviation is tiny, so its products are highly consistent and almost all fall within tolerance; Factory Y’s mean is closer to the nominal value, but its standard deviation is 7.5 times larger, so many products may exceed tolerance and the defect rate is higher. Conclusion: looking only at the means, Factory Y appears better; combining the standard deviations, Factory X clearly has superior quality control.

这类应用题的答题要点：第一，必须把统计量翻译成业务含义，例如”标准差小意味着产品质量稳定”；第二，比较时要控制变量，同一道题中两组数据要使用同一种统计量；第三，如果题目给出成本或损失信息（如”超出公差每个赔 2 元”），还要结合数值做定量判断。Edexcel 的应用题通常提供真实背景（生产、金融、体育、气象），但统计方法完全相同，关键是不要被冗长的文字吓住，先提取数据再套用标准流程。

Key points for such application questions: first, translate the statistics into business meaning, for example “a small standard deviation means stable product quality”; second, keep the comparison fair by using the same statistic for both groups; third, if the question gives cost or loss information (such as “each item out of tolerance costs 2 yuan”), make a quantitative judgement with the numbers. Edexcel application questions usually carry a realistic context (production, finance, sport, weather), but the statistical method is identical: do not be intimidated by long wording, extract the data first, then follow the standard procedure.

十二、考试常见陷阱与答题技巧 | Common Exam Pitfalls and Answering Techniques

陷阱一：忘记说明单位。均值、标准差等统计量都要带单位（如”分””mm”），解释时也要把数值和情境挂钩。陷阱二：分组数据直接用区间端点代替组中值。必须用组中值（上下限的平均数），否则全题连锁出错。陷阱三：方差开方时漏掉平方根，把方差当标准差写进结论。陷阱四：求四分位数时排序出错，尤其是数据个数为偶数时。陷阱五：比较题只写”甲班均值高”而不写”所以甲班整体更好”，缺少连接数值与结论的解释句，这在评分方案中通常单独占分。

Pitfall one: forgetting units. Statistics such as the mean and standard deviation must carry units (for example “marks” or “mm”), and interpretations must link the numbers to the context. Pitfall two: using interval endpoints instead of midpoints for grouped data. You must use the midpoint (the average of the two bounds), otherwise every later step fails. Pitfall three: forgetting the square root when converting variance to standard deviation, then quoting the variance as the standard deviation. Pitfall four: sorting errors when finding quartiles, especially with an even number of data values. Pitfall five: writing only “Class A has a higher mean” without the concluding sentence “so Class A is better overall”; the sentence linking the number to the conclusion usually earns a separate mark in the mark scheme.

答题技巧总结：(1) 先排序再求位置型统计量；(2) 计算均值方差时用表格列 x、fx、fx²，减少笔误；(3) 比较题按”中心趋势 + 离散程度 + 情境解释”三段式作答；(4) 涉及离群值时明确写出判定规则和计算结果；(5) 最后留 30 秒检查单位与平方根。掌握这些细节，数据比较类题目就能稳定拿满分。

Summary of techniques: (1) sort the data before finding positional measures; (2) use a table with columns x, fx and fx² when computing the mean and variance to reduce arithmetic slips; (3) answer comparison questions in three parts: central tendency + spread + interpretation in context; (4) when outliers are involved, state the rule and show the calculation explicitly; (5) keep the last 30 seconds to check units and square roots. Master these details and data comparison questions become reliable full marks.

Summary | 总结

数据比较是 A-Level 数学统计部分的基础能力。有效的比较必须同时使用中心趋势量（均值、中位数、众数）与离散度量（极差、四分位距、标准差），并根据数据是否对称、是否存在离群值选择合适的组合：对称数据用均值配标准差，偏态或含离群值的数据用中位数配四分位距。箱线图、百分位数和偏度系数提供了直观与定量的比较工具，编码技巧则让计算更加高效。掌握判定离群值的 1.5 倍 IQR 规则、分组数据的组中值处理，以及”数值 + 解释”的答题结构，就能在考试中稳定得分。

Comparing data is a foundational skill in the A-Level mathematics statistics component. An effective comparison must combine a measure of central tendency (mean, median, mode) with a measure of spread (range, interquartile range, standard deviation), and choose the appropriate pairing according to whether the data are symmetric and whether outliers exist: use the mean with the standard deviation for symmetric data, and the median with the interquartile range for skewed data or data containing outliers. Box plots, percentiles and the coefficient of skewness provide visual and quantitative tools for comparison, while coding makes the arithmetic more efficient. Mastering the 1.5 × IQR outlier rule, the midpoint treatment of grouped data, and the “number plus interpretation” answering structure will earn you reliable marks in the exam.

更多咨询请联系16621398022（同微信）