Common Misconceptions in Statistics and How to Correct Them | 统计常见误区与纠正方法

📚 Common Misconceptions in Statistics and How to Correct Them | 统计常见误区与纠正方法

Statistics can be tricky at Year 10, and even the most confident students fall into common traps. From misreading graphs to confusing probability rules, these errors can cost valuable marks in Cambridge assessments. This article identifies ten persistent misconceptions and shows you exactly how to overcome them, so you can analyse data with accuracy and confidence.

统计在 Year 10 阶段可能充满陷阱，即使最自信的学生也会落入常见的误区。从误读图表到混淆概率规则，这些错误会在剑桥考评中让你失分。本文指出十个顽固的误解，并教你如何纠正，让你能够准确而自信地分析数据。

1. Confusing Correlation with Causation | 将相关性误解为因果关系

A classic mistake is to see two variables moving together and immediately decide that one causes the other. A strong correlation simply means that as one variable increases, the other tends to increase or decrease, but there may be a third lurking variable driving both. For example, ice cream sales and drowning incidents are positively correlated, yet eating ice cream does not cause drowning — both rise in hot weather because more people swim and buy ice cream.

一个经典错误是看到两个变量一起变化就立刻判定一个导致另一个。强相关仅仅意味着一个变量增加时，另一个倾向于增加或减少，但背后可能隐藏着第三个变量同时驱动两者。例如，冰淇淋销量和溺水事件呈正相关，但吃冰淇淋并不会导致溺水——两者都在炎热天气下增加，因为更多人游泳和购买冰淇淋。

To avoid this, always ask whether there is a plausible mechanism linking cause and effect, and whether a controlled experiment could rule out other factors. In observational studies, use phrases like ‘is associated with’ rather than ’causes’. Remember that correlation is a useful clue but never proof of causation.

要避免这个错误，务必先问是否存在合理的因果机制，以及能否通过对照实验排除其他因素。在观察性研究中，使用“与……相关”而不是“导致”的表述。记住，相关性是有用的线索，但绝不是因果关系的证明。

2. Blindly Trusting the Mean | 盲目信赖算术平均数

Many students calculate the mean for any dataset without considering the shape of the distribution. The mean is sensitive to extreme values (outliers). In a salary survey where most employees earn around £25,000 but a few executives earn millions, the mean gives a distorted picture of a typical worker’s pay. A better average here is the median, which is the middle value and resists the pull of outliers.

许多学生计算任何数据集的平均数时都不考虑分布形状。均值对极端值（异常值）非常敏感。在一项薪资调查中，大多数员工收入约 25,000 英镑，但少数高管收入数百万，均值就会扭曲典型员工的收入状况。这种情况下更好的平均数是中位数，也就是中间值，它能抵抗异常值的拉动。

When data is symmetric, the mean and median are close, and the mean is perfectly fine. But if a box plot or histogram shows a clear skew, report the median and the interquartile range alongside. Always ask: which measure best represents a typical value for this context?

当数据对称时，均值和中位数接近，使用均值完全没问题。但如果箱线图或直方图显示出明显偏态，就应报告中位数和四分位距。始终问自己：哪种度量最能代表此背景下的典型值？

3. Misreading Histograms: Height vs. Area | 误读直方图：高度与面积

A common blunder is treating a histogram like a bar chart and assuming the bar height directly tells you the frequency. In a histogram, it is the area of each bar that is proportional to frequency. If the class intervals are not all equal, the bar height represents frequency density, not frequency. For instance, a bar with class width 10 and frequency 30 has frequency density 3, while a wider bar of width 20 and frequency 30 has density only 1.5. The second bar would be shorter despite having the same frequency.

一个常见错误是把直方图当作条形图，并假设柱子的高度直接表示频数。在直方图中，每个柱子的面积与频数成正比。如果组距不全相等，柱子的高度代表的是频数密度，而不是频数。例如，组距为 10、频数为 30 的柱子频数密度为 3，而组距为 20、频数同为 30 的柱子密度仅为 1.5。第二个柱子虽然频数相同，但会更矮。

Frequency density = Frequency ÷ Class width

频数密度 = 频数 ÷ 组距

Always check the horizontal axis for unequal intervals. If they vary, calculate the frequency density before drawing or reading the graph. To find the frequency of any interval, multiply the bar’s height by its width. This ensures you correctly interpret data presented in histograms.

务必检查横轴是否存在不等距区间。如果组距不同，在绘制或阅读图形前要先计算频数密度。要找出任一组段的频数，用柱子的高度乘以其宽度。这样才能正确解读直方图中的数据。

4. The Gambler’s Fallacy | 赌徒谬误

When flipping a fair coin, seeing five heads in a row makes many people believe that a tail is ‘due’ next. This is the gambler’s fallacy — the mistaken belief that past outcomes in independent events affect future probabilities. In reality, each coin flip is independent, and the probability of a tail remains exactly 0.5 every single time, regardless of what happened before.

抛一枚均匀硬币时，连续看到五个正面会让许多人认为下一次“该”出反面了。这就是赌徒谬误——错误地认为独立事件的过去结果会影响未来概率。事实上，每一次抛硬币都是独立的，无论之前发生了什么，反面的概率每次都是 0.5。

To correct this, remind yourself that probability has no memory. A roulette wheel, a die, or a lottery machine does not ‘compensate’ for earlier results. The long-run pattern only emerges over thousands of trials, but each individual trial is governed by the same fixed probabilities. Practice spotting independence in questions, and resist the urge to predict the next outcome based on a short sequence.

要纠正这个错误，提醒自己概率没有记忆。轮盘、骰子或彩票机不会为之前的结果“补偿”。长期规律仅在成千上万次试验后才会显现，而每一次试验都由相同的固定概率支配。练习在题目中识别独立性，并克制基于短期序列预测下一个结果的冲动。

5. Mixing Up Independent and Mutually Exclusive Events | 混淆独立事件与互斥事件

Independent events and mutually exclusive events are often treated as synonyms, but they are profoundly different. Two events are independent if the occurrence of one does not affect the probability of the other. For example, getting a head on a coin flip and rolling a 6 on a die are independent: P(head AND 6) = P(head) × P(6) = 0.5 × 1/6. In contrast, events are mutually exclusive if they cannot happen at the same time, such as rolling a 1 and rolling a 2 on a single die. Their intersection probability is 0.

独立事件和互斥事件常被当作同义词，但它们截然不同。如果一件事发生不影响另一件事的概率，两个事件就是独立的。例如，抛硬币得到正面和掷骰子得到 6 是独立事件：P(正面且6) = P(正面) × P(6) = 0.5 × 1/6。相反，如果两件事不可能同时发生，它们就是互斥的，例如掷一个骰子同时得到 1 和 2。它们的交事件概率为 0。

A dangerous misunderstanding occurs when students apply the multiplication rule P(A ∩ B) = P(A) × P(B) to mutually exclusive events. That formula is only valid for independent events. For mutually exclusive events, P(A ∩ B) = 0 by definition. When working with probability trees or Venn diagrams, always ask: can these two outcomes occur together? If not, they are mutually exclusive; if yes, then test for independence using the definition P(A|B) = P(A).

当学生将乘法规则 P(A ∩ B) = P(A) × P(B) 用于互斥事件时，就会出现危险的误解。这个公式仅对独立事件有效。对于互斥事件，根据定义 P(A ∩ B) = 0。在使用概率树或维恩图时，务必先问：这两个结果能否同时发生？如果不能，它们互斥；如果能，则用 P(A|B) = P(A) 的定义检验是否独立。

6. Ignoring Sampling Bias | 忽略抽样偏差

A statistical conclusion is only as good as the sample it comes from. A frequent error is to collect data from a convenient but unrepresentative group and then generalise to the whole population. If you survey only your classmates about school lunch preferences, you cannot claim the results reflect the entire school. The sample is biased because it over-represents one age group and possibly one set of tastes.

统计结论的好坏取决于其来源样本。一个常见错误是从方便但不具代表性的群体中收集数据，然后推广到整体。如果你只调查班上同学对学校午餐的偏好，就不能声称结果反映全校情况。该样本有偏差，因为它过度代表某一年龄段和可能的某种口味。

To avoid sampling bias, use some form of random sampling where every member of the population has an equal chance of being selected. In Cambridge questions, look out for who is left out. A survey about internet habits conducted via an online poll misses people without internet access. Always consider whether the sample truly mirrors the population before trusting the results.

要避免抽样偏差，可采用某种随机抽样方法，确保总体中每个成员都有同等机会被选中。在剑桥考题中，注意谁被遗漏了。通过网络投票进行的互联网使用习惯调查会遗漏没有网络接入的人。在相信结果之前，始终考虑样本是否真正代表了总体。

7. Misusing Pie Charts | 饼图的误用

Pie charts are attractive but easily abused. A common error is using a pie chart when there are too many categories, making thin slices almost impossible to compare. If you have eight or ten sectors, human eyes struggle to judge angles accurately. Pie charts also fail when you need to compare two distributions, because viewers must mentally align angles from two separate circles.

饼图很吸引人，但容易被滥用。一个常见错误是类别过多时使用饼图，导致细薄的扇区几乎无法比较。如果有八个或十个扇区，人眼很难准确判断角度。当需要比较两个分布时，饼图也不适用，因为观察者必须在脑中对齐两个独立圆的扇区角度。

A better choice is often a horizontal bar chart, which displays categories clearly and allows exact length comparison. Use a pie chart only when you have 3-5 categories that clearly sum to a meaningful whole, and the primary goal is to show proportions of that whole. Always label percentages or frequencies on the slices to help the viewer.

更好的选择通常是水平条形图，它能清晰显示类别并允许精确的长度比较。只有当有 3 到 5 个类别、明显构成一个有意义的整体，且主要目标是展示各部分占比时，才使用饼图。务必在扇区上标注百分比或频数以帮助阅读。

8. Misinterpreting Cumulative Frequency Graphs | 累积频数图的误解

Cumulative frequency curves are a powerful tool for finding medians and quartiles, but many Year 10 students confuse the steepness of the curve with high frequency. A steep section does not indicate a high frequency of data; it indicates that the data is densely packed within a small interval, meaning many values are added over a narrow class width. The frequency of any interval is simply the difference between cumulative frequencies at its boundaries.

累积频数曲线是寻找中位数和四分位数的强大工具，但许多 Year 10 学生会把曲线的陡峭程度与高频率混淆。陡峭的段落并不表示数据频率高，而是表示数据密集地分布在一个小区间内，意味着在狭窄的组距上叠加了许多数值。任一组段的频数就是该组上下界累积频数之差。

To read the median correctly, locate the halfway mark on the cumulative frequency axis (total frequency ÷ 2), draw a horizontal line to the curve, then drop vertically to the horizontal axis. That value is the median. The lower quartile is at total frequency ÷ 4, and the upper quartile at 3 × total frequency ÷ 4. Practice sketching these lines so you don’t rely on guesswork.

要正确读取中位数，先在累积频数轴上找到总频数的一半（总频数 ÷ 2），画一条水平线连到曲线，再垂直向下交于横轴，该值即为中位数。下四分位数在总频数 ÷ 4 处，上四分位数在 3 × 总频数 ÷ 4 处。练习绘制这些线条，避免仅凭猜测。

9. Assuming All Data Follows a Normal Distribution | 假定所有数据都符合正态分布

Textbooks often show bell-shaped symmetric distributions, leading some students to assume that real-world data is always ‘normal’. In reality, many datasets are skewed. Reaction times, house prices, and rainfall amounts are often skewed to the right, with a long tail of high values. Treating such data as symmetric will lead to poor choices, such as using the mean and standard deviation when the five-number summary would be far more informative.

教科书常展示钟形对称分布，使得一些学生以为现实数据总是“正态”的。实际上，许多数据集是偏态的。反应时间、房价和降雨量往往右偏，有一个高值长尾。把这类数据当作对称处理会导致错误选择，例如使用均值和标准差，而实际上五数概括能提供更多信息。

Before choosing summary statistics, look at the shape. Draw a quick stem-and-leaf diagram or a box plot. If the median lies closer to one end of the box, or whiskers are very unequal, the data is skewed. In such cases, the median and interquartile range are more robust and meaningful than the mean and range.

在选择汇总统计量之前，先观察分布形状。快速绘制茎叶图或箱线图。如果中位数靠近箱体一端，或者须线长短极不相等，数据就是偏态的。这种情况下，中位数和四分位距比均值和全距更稳健、更有意义。

10. Confusing the Range with the Interquartile Range | 混淆全距与四分位距

The range (maximum – minimum) is simple but extremely sensitive to outliers. A single typing error in a dataset of heights — recording 189 cm as 819 cm — blows the range out of proportion. The interquartile range (IQR = upper quartile – lower quartile) clears away the extreme ends and measures the spread of the middle 50% of data. It is therefore a much more reliable measure of spread when outliers are present.

全距（最大值减最小值）简单但极易受异常值影响。一组身高数据中一个录入错误——把 189 cm 录成 819 cm——就会让全距不成比例地膨胀。四分位距（IQR = 上四分位数 – 下四分位数）去除了极端两端，衡量中间 50% 数据的分散程度。因此当存在异常值时，它是更可靠得多的散布度量。

Use the range only when you are sure the data has no extreme anomalies and you need a quick sense of total variability. In formal analysis, always report the IQR alongside the median. In box plots, the IQR is the length of the box, making it a visual indicator of consistency. Always pair the median with the IQR, just as you pair the mean with the standard deviation.

仅当确定数据没有极端异常值且需要快速了解总变异程度时，才使用全距。在正式分析中，务必在中位数旁报告 IQR。在箱线图中，IQR 就是箱体长度，直观地显示数据的一致性。始终将中位数与 IQR 配对，就像将均值与标准差配对一样。

Published by TutorHao | Statistics Revision Series | aleveler.com

更多咨询请联系16621398022（同微信）

Common Misconceptions in Statistics and How to Correct Them | 统计常见误区与纠正方法