📚 Common Statistical Mistakes and How to Fix Them | 常见统计误区与纠正方法
Statistics can be full of subtle traps. Even students who are comfortable with formulas often fall into the same common errors when interpreting data, selecting the right average, or reasoning about probability. These mistakes can lower exam scores and, more importantly, cloud your understanding of what data is really saying. In this article, we will walk through the most frequent mistakes Year 10 students make in CIE Statistics and, crucially, show you exactly how to correct them with clear reasoning and examples.
统计学充满了微妙的陷阱。即使那些对公式很熟悉的学生，在解读数据、选择正确的平均数或进行概率推理时，也常常会犯同样的错误。这些错误会拉低考试成绩，更重要的是，会模糊你对数据真正含义的理解。本文将梳理 Year 10 学生在 CIE 统计课程中最常犯的错误，并重点通过清晰的推理和实例，告诉你究竟该如何纠正。
1. Confusing Mean, Median and Mode | 混淆平均数、中位数和众数
Many students simply compute all three measures of central tendency and then randomly pick one, or always default to the mean. The error occurs when the choice does not match the data type or the presence of extreme values.
许多学生只是把三种集中趋势度量都算出来，然后随意挑选一个，或者总是默认使用平均数。当选择与数据类型或极端值的存在不匹配时，错误就发生了。
The mean uses every data point and is sensitive to outliers. In a village where most people earn £20,000 a year but one billionaire lives there, the mean income might suggest everyone is well-off, which is misleading. The median, being the middle value, would give a much truer picture of the typical earnings.
平均数使用了每一个数据点，对异常值敏感。如果一个村庄里大多数人年收入为 20,000 英镑，但有一位亿万富翁住在那里，平均收入可能会让人觉得每个人都很富裕，这是误导。中位数是中间值，能更真实地反映典型收入。
Additionally, using the mean for categorical data is meaningless. If you ask 30 students their favourite colour, the modal colour is the only sensible measure. You cannot calculate a ‘mean colour’.
此外，对分类数据使用平均数毫无意义。如果你问 30 个学生他们最喜欢的颜色，众数颜色是唯一合理的度量。你无法计算“平均颜色”。
| Measure | Best used when | Watch out for |
|---|---|---|
| Mean | Symmetric, no outliers | Skewed data, extreme values |
| Median | Skewed data, ordinal data | Not using all data information |
| Mode | Categorical data, most frequent | Bimodal or multimodal sets |
修正方法：先观察数据的分布。如果有异常值，就用中位数。如果是分类数据，就用众数。如果分布对称且没有异常值，平均数最佳。一定要说明为什么选择这种度量。
2. Misunderstanding Standard Deviation | 误解标准差
A common error is to think that a larger standard deviation automatically means bad data, or to confuse the standard deviation with the range. Students often fail to link standard deviation to the spread of data around the mean.
一个常见错误是认为标准差大就自动意味着数据不好，或者将标准差与全距混淆。学生常常未能将标准差与数据围绕平均数的分散程度联系起来。
Standard deviation measures the typical distance of data points from the mean. A small standard deviation tells you the data points are tightly clustered; a large one shows wide dispersion. Neither is ‘good’ or ‘bad’ without context. In a factory, a small standard deviation of product lengths is desired; in a test designed to differentiate students, a moderate standard deviation is needed.
标准差衡量数据点与平均数之间的典型距离。标准差小意味着数据点紧密聚集；标准差大则表示分散程度高。脱离背景，无法说哪个“好”或“坏”。在工厂里，希望产品长度的标准差小；而在旨在区分学生的考试中，则需要适度的标准差。
Another mistake is misinterpreting the formula. Remember that for a population the standard deviation symbol is σ and we divide by n, while for a sample we use s and divide by n−1. Mixing these up leads to biased estimates.
另一个错误是误解公式。记住，对于总体，标准差符号为 σ，我们除以 n；而对于样本，我们用 s，并除以 n−1。混淆二者会导致有偏估计。
纠正方法：把标准差看作“平均距离”，而非绝对的好坏标志。练习从散点或直方图目测标准差的大小，并确保在样本情形下使用 n−1。
3. Interpreting Histograms Incorrectly | 错误解读直方图
In histograms with unequal class widths, students frequently read the height of the bar as the frequency. This mistake is one of the most consistently penalised on CIE papers.
在组距不等的直方图中，学生常常将条形的高度读作频数。这是在 CIE 考试中最常被扣分的错误之一。
The area of each bar, not its height, represents the frequency. When class widths differ, we plot frequency density on the vertical axis, where frequency density = frequency ÷ class width. A taller bar may actually represent fewer data points if its class width is very narrow.
每个条形的面积（而非高度）表示频数。当组距不同时，我们在纵轴上绘制频率密度，频率密度 = 频数 ÷ 组距。一个较高的条形可能实际上代表较少的数据点，如果它的组距非常窄的话。
始终检查横轴的刻度是否均匀。计算频率密度并识别众数组（最高频率密度）时，要看面积最高的条形，而不是单纯看高度。对于等宽直方图，高度才直接对应频数。
4. Thinking Correlation Implies Causation | 认为相关性蕴含因果关系
This logical trap appears in scatter graphs and correlation questions. Students see a strong positive correlation and immediately claim one variable causes the other to change.
这个逻辑陷阱出现在散点图和相关性题目中。学生看到强正相关，就立刻声称一个变量的变化导致了另一个变量的变化。
Correlation measures the strength and direction of a linear relationship. However, causation requires a mechanism. For example, a data set may show that ice cream sales and drowning incidents both increase in summer. It would be wrong to say eating ice cream causes drowning; the hidden variable here is hot weather, which independently raises both.
相关性度量线性关系的强度和方向。然而，因果关系需要一个机制。例如，一个数据集可能显示冰淇淋销量和溺水事件在夏季都增加。说吃冰淇淋导致溺水是错误的；这里的隐藏变量是炎热天气，它独立地推高了二者。
When a question asks you to interpret a strong r value, always use language like ‘suggests an association’ and explicitly state ‘correlation does not imply causation’. Look out for possible confounding variables.
当题目要求你解释一个很强的 r 值时，务必使用“表明存在关联”这样的语言，并明确说明“相关性不意味着因果性”。要留意可能的混杂变量。
5. Sampling Bias in Surveys | 调查中的抽样偏差
Students often design a sampling method that is convenient rather than representative, then wonder why their conclusion is invalid. One classic error is asking only friends or people in one location.
学生常常设计一种方便而非具有代表性的抽样方法，然后好奇为什么他们的结论无效。一个经典错误是只询问朋友或某个地点的人。
A sample must reflect the population. If you want to estimate the average screen time of all Year 10 students in a city, interviewing only members of the computer club will oversample heavy users. This leads to selection bias. Similarly, a voluntary response sample (e.g., an online poll) attracts strong opinions and can skew results.
样本必须反映总体。如果你想估算一个城市所有 Year 10 学生的平均屏幕时间，只采访计算机俱乐部成员会使重使用者被过度抽样，从而导致选择偏差。同样，自愿回应样本（如在线投票）会吸引强烈意见，可能使结果产生偏斜。
纠正方法：尽可能使用简单随机抽样，给总体中每个个体一个已知且非零的入选概率。如果必须使用分层抽样，要确保各层比例与总体一致。在考试中，提及样本的规模和随机性，并指出任何潜在的偏差来源。
6. The Gambler’s Fallacy | 赌徒谬误
Probability questions often reveal the gambler’s fallacy: the belief that if something happens more frequently than normal now, it will happen less frequently in the future to ‘balance out’.
概率题常常暴露赌徒谬误：相信如果某件事现在发生的频率高于正常水平，未来它就会发生得少一些以“扯平”。
If you toss a fair coin and get five heads in a row, the chance of a tail on the sixth toss is still ½. Coins have no memory. Each toss is independent. The misconception comes from confusing the probability of a specific sequence with the probability of the next single event.
如果你抛一枚公平硬币，连续得到五个正面，第六次抛掷出现反面的概率仍然是 ½。硬币没有记忆。每次抛掷都是独立的。这一误解来源于混淆了特定序列的概率与下一次单一事件的概率。
To avoid this, always state explicitly whether events are independent. Show the calculation: P(tail on 6th toss | 5 heads) = 0.5, not something smaller. Use tree diagrams to visualise independence.
为避免此错误，务必明确说明事件是否独立。展示计算过程：P(第6次为反面 | 前5次为正面) = 0.5，而不是某个更小的值。用树状图来可视化独立性。
7. Mixing Up Independent and Mutually Exclusive Events | 混淆独立事件与互斥事件
This is a vocabulary error that causes students to use the wrong formula. ‘Mutually exclusive’ means two events cannot happen at the same time; ‘independent’ means the occurrence of one does not affect the probability of the other.
这是一个词汇错误，会导致学生用错公式。“互斥”意味着两个事件不能同时发生；“独立”意味着一个事件的发生不影响另一个事件的概率。
For mutually exclusive events, P(A ∩ B) = 0, so P(A ∪ B) = P(A) + P(B). For independent events, P(A ∩ B) = P(A) × P(B). These are entirely different concepts. For example, drawing a red card and drawing a black card from a deck in one draw are mutually exclusive but not independent: if one happens, the other cannot.
对于互斥事件，P(A ∩ B) = 0，所以 P(A ∪ B) = P(A) + P(B)。对于独立事件，P(A ∩ B) = P(A) × P(B)。这是完全不同的概念。例如，一次从一副牌中抽到红牌和抽到黑牌是互斥的但不独立：如果一件事发生，另一件事就不会发生。
纠正方法：每次遇到 A 和 B 时，先问自己：“它们能同时发生吗？” 如果不能，就是互斥的。再问：“A 的发生会改变 B 的概率吗？” 如果不会，就是独立的。这两个属性没有必然联系。
8. Conditional Probability Confusions | 条件概率的混淆
Students often reverse the condition or fail to update the sample space when an event has occurred. The question ‘Given that a student studies French, what is the probability they also study Spanish?’ is different from ‘Given that they study Spanish, what is the probability they study French?’
学生经常颠倒条件，或者在事件发生后未能更新样本空间。问题“已知一个学生学法语，他同时也学西班牙语的概率是多少？”与“已知他学西班牙语，他学法语的概率是多少？”是不同的。
The formula for conditional probability is P(A|B) = P(A ∩ B) / P(B). The denominator is the probability of the condition, which reduces the sample space. Many students forget to reduce the denominator, or they use the wrong event in the denominator.
条件概率的公式是 P(A|B) = P(A ∩ B) / P(B)。分母是条件的概率，它缩小了样本空间。许多学生忘记缩小分母，或者在分母中用了错误的事件。
Draw a two-way table or a Venn diagram. Shade in the condition first, and only consider that restricted set. Practise rewriting the problem: ‘Out of all the students who do B, what fraction also do A?’ This linguistic check helps prevent reversals.
画一个双向表或韦恩图。先把条件部分涂上阴影，只考虑那个受限集合。练习改写问题：“在所有做 B 的学生中，有多少比例也做 A？”这种语言检查有助于防止颠倒。
9. Ignoring Outliers | 忽略异常值
When calculating the mean and range, students often mechanically include every value without noticing that one extreme point can distort the whole summary. Worse, they delete outliers without justification.
在计算平均数和全距时，学生常常机械地纳入每一个值，没有注意到一个极端点会扭曲整个汇总。更糟糕的是，他们在没有任何理由的情况下删掉异常值。
An outlier is a value that lies far from the main body of data, typically more than 1.5 × IQR below the lower quartile or above the upper quartile. Before discarding an outlier, investigate whether it is a measurement error or a genuine rare event. In many exam contexts, you should report the mean both with and without the outlier and discuss its influence.
异常值是指远离数据主体的值，通常低于下四分位数减去 1.5 倍 IQR 或高于上四分位数加上 1.5 倍 IQR。在摒弃异常值之前，要调查它是测量错误还是真实的罕见事件。在许多考试情境中，你应该同时报告包含和不包含异常值的平均数，并讨论其影响。
Always check for outliers using a box-and-whisker plot or the IQR criterion. Comment on resistance: the median and IQR are resistant to outliers, while the mean and range are not.
一定要使用箱线图或 IQR 准则检查异常值。要评价抗性：中位数和 IQR 对异常值有抗性，而平均数和全距则没有。
10. Cumulative Frequency Graph Errors | 累积频率图的错误
Cumulative frequency diagrams cause two classic mistakes: plotting at the wrong end of the class interval and misreading the median and quartiles.
累积频率图会造成两个经典错误：在组区间错误的一端描点，以及误读中位数和四分位数。
Cumulative frequency is plotted against the upper class boundary, not the midpoint. If the interval is 10 ≤ x < 20, you plot the point at x = 20. Connecting points with a smooth curve rather than straight line segments is required. When finding the median, read across from ½ of the total frequency on the vertical axis, then down to the horizontal axis.
累积频率应该对应上组限描点，而不是中点。如果区间是 10 ≤ x < 20，你就在 x = 20 处描点。需要用平滑曲线连接点，而非直线段。找中位数时，从纵轴上总频数的一半处水平读过去，再向下读到横轴。
Practice drawing the curve, and use a ruler to project lines across and down. Remember that the interquartile range is Q₃ − Q₁, not the range of the whole graph. Be careful with the scale on the frequency axis.
练习绘制曲线，并用尺子横向和纵向投影。记住四分位距是 Q₃ − Q₁，而不是整个图形的全距。注意频数轴的刻度。
11. Misreading Scales on Charts | 图表刻度的误读
Under time pressure, students glance at a bar chart or line graph and quote incorrect values because the scale does not start at zero, or because they misjudge intermediate divisions.
在时间压力下，学生匆匆扫一眼条形图或折线图，就报出错误的值，因为刻度并非从零开始，或者他们错判了中间的分格。
A bar chart with a truncated vertical axis can make differences appear much larger than they really are. Always check the starting value and the increments. If the axis begins at 50 instead of 0, a bar that is twice as tall might represent only a small absolute increase.
纵轴被截断的条形图会使差异看起来比实际大得多。一定要检查起始值和增量。如果轴从 50 开始而不是 0，高一倍的条形可能只代表很小的绝对增长。
Read values by carefully estimating between grid lines. Do not just guess half-way if the scale is in multiples of 3 or 7. Write down the value as precisely as the graph allows, and use it in further calculations only after confirming.
通过仔细估计网格线之间的值来读取。如果刻度是 3 或 7 的倍数，不要仅仅猜测一半。尽量按照图形允许的精度写下数值，并在确认后用于进一步计算。
12. Confusing Population and Sample Statistics | 混淆总体与样本统计量
In surveys and experiments, students often treat a sample statistic as if it were the exact population parameter. This leads to overconfidence and incorrect conclusions about variability.
在调查和实验中，学生常常把样本统计量当作确切的总体参数来看待。这会导致过度自信，以及对变异性得出错误的结论。
A sample mean x̄ is an estimate of the population mean µ. Different samples yield different x̄ values; this is sampling variability. If you calculate the standard deviation of a sample using the population formula (dividing by n instead of n−1), you will on average underestimate the true population standard deviation σ.
样本平均数 x̄ 是总体平均数 µ 的估计值。不同的样本会产生不同的 x̄；这就是抽样变异性。如果你用总体公式（除以 n 而不是 n−1）来计算样本的标准差，平均而言你会低估真实的总体标准差 σ。
Always use sample notation (x̄, s) when dealing with a sample, and population symbols (µ, σ) only for the whole group. When reporting a sample statistic, it is good practice to mention the sample size and acknowledge that it is an estimate.
处理样本时，始终使用样本记号（x̄, s），仅当涉及整个群体时才使用总体符号（µ, σ）。在报告样本统计量时，最好提及样本量，并承认这是一个估计值。
Published by TutorHao | Statistics 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Cancel reply