Common Mistakes in A-Level Statistics Unit 2 (MA04) | A-Level 统计学第二单元(MA04)常见易错点总结

📚 Common Mistakes in A-Level Statistics Unit 2 (MA04) | A-Level 统计学第二单元(MA04)常见易错点总结

A-Level Statistics Unit 2 (MA04) for the International A-Level 9660 specification challenges students with discrete probability distributions, hypothesis testing, approximation techniques, and sampling theory. Many candidates lose marks not because they lack understanding, but because they repeatedly fall into the same traps involving misinterpretation of conditions, incorrect use of approximations, confusion between one‑tailed and two‑tailed tests, and mishandling of significance levels. This article summarises the most persistent pitfalls observed in past papers and revision sessions, providing clear explanations and strategies to avoid them.

A-Level 统计学第二单元(MA04)（国际版9660考纲）涉及离散概率分布、假设检验、近似方法以及抽样理论，许多考生丢分并非因为不懂知识点，而是反复掉进相同的陷阱：条件误判、近似使用错误、单双尾混淆、显著性水平处理不当等。本文归纳了历年真题与复习课中最常见的易错点，并给出清晰解释和避坑策略。

1. Misunderstanding the Conditions for Binomial and Poisson Distributions | 混淆二项分布与泊松分布的使用条件

The binomial distribution X ~ B(n, p) requires a fixed number of trials n, each with two outcomes (success/failure), a constant probability p, and independence between trials. Students often apply it when the event can occur multiple times within a continuous interval, which instead calls for a Poisson model. Similarly, a Poisson distribution X ~ Po(λ) requires events occurring independently at a constant average rate in a continuous domain (time, length, area). Failing to recognise that Poisson is used for counts of rare events in a continuous space leads to selecting the wrong distribution from the start.

二项分布 X ~ B(n, p) 要求试验次数 n 固定，每次试验只有两种结果，概率 p 恒定，且试验之间相互独立。学生常常在事件可以在连续区间内多次发生时错误使用二项分布，而这恰恰是泊松分布的适用场景。同样，泊松分布 X ~ Po(λ) 要求事件在连续域（时间、长度、面积等）内以恒定平均率独立发生。没有意识到泊松用于连续空间中稀有事件的计数，从一开始就会选错分布。

• Binomial checklist: fixed n, independent trials, constant p, discrete count of successes.
  二项清单：固定 n，独立试验，恒定 p，成功次数的离散计数。
• Poisson checklist: events in a continuum, constant mean rate, independence, rare events.
  泊松清单：连续域内的事件，恒定的平均率，独立性，稀有事件。

2. Forgetting to Define the Random Variable and Parameters | 忘记定义随机变量与参数

A common mark‑losing habit is jumping straight into calculations without stating the distribution explicitly, e.g. ‘Let X be the number of defective items in a sample of 10, X ~ B(10, 0.2).’ Without this statement, the working lacks context, and examiners cannot award method marks. Students must also specify all parameters, such as n and p for binomial, or λ for Poisson, and indicate when a normal approximation is used by writing N(μ, σ²).

一个常见的丢分习惯是直接代入计算而不先明确写出分布，例如“设 X 为10个样品中次品的数量，X ~ B(10, 0.2)”。缺少这个说明，解题过程就失去上下文，考官无法给予方法分。学生还必须写出所有参数，如二项分布的 n 和 p，或泊松分布的 λ，并在使用正态近似时标明 N(μ, σ²)。

Examiners look for a clear frame: ‘X ~ B(20, 0.15)’, ‘X ~ Po(3.2)’, or ‘X can be approximated by N(30, 30)’. Even when using tables, write down the distribution you are using.
考官期望清晰的框架：“X ~ B(20, 0.15)”、“X ~ Po(3.2)” 或 “X 可由 N(30, 30) 近似”。就算查表，也要写下你使用的分布。

3. Incorrect Use of Normal Approximations and the Continuity Correction | 正态近似与连续性校正的错误使用

When approximating a binomial or Poisson distribution with a normal distribution, a continuity correction is essential because the normal distribution is continuous while the original data are discrete. The most frequent mistake is either omitting the correction entirely or applying it with the wrong sign. For a binomial X ~ B(n, p) approximated by Y ~ N(np, np(1−p)), if we want P(X ≤ a), we use P(Y < a + 0.5). For P(X ≥ a), use P(Y > a − 0.5). Students often try to memorise ±0.5 without understanding the logic, leading to confusion when inequalities are strict or when calculating ‘between’ probabilities.

用正态分布近似二项或泊松分布时，必须进行连续性校正，因为正态分布是连续的而原始数据是离散的。最常见的错误是要么完全忽略校正，要么用错正负号。对于二项 X ~ B(n, p) 用 Y ~ N(np, np(1−p)) 近似，若要计算 P(X ≤ a)，使用 P(Y < a + 0.5)；P(X ≥ a) 则用 P(Y > a − 0.5)。学生往往死记 ±0.5 而不理解逻辑，当不等号包含等号或计算“介于之间”的概率时，极易混淆。

A robust approach: draw a small number line and shade the relevant discrete bars, then see where the continuous boundary should lie. For strict inequalities like P(X < a), treat it as P(X ≤ a−1) and apply correction.
稳妥的做法：画一条小型数轴，给相关离散柱形涂色，再观察连续边界应落在何处。对于严格不等式如 P(X < a)，应视作 P(X ≤ a−1) 后再进行校正。

4. Choosing the Wrong Tail in Hypothesis Tests | 假设检验中选错单双尾

Determining whether a test is one‑tailed or two‑tailed is a fundamental step; yet many students default to a two‑tailed test without reading the alternative hypothesis carefully. If H₁ includes a ‘not equal to’ sign (≠), the test is two‑tailed. If H₁ states a direction (greater than, less than, increase, decrease, improved), the test is one‑tailed. A common pitfall is using a two‑tailed test when the context clearly indicates a directional claim, which halves the critical region on one side and leads to a wrong conclusion or a lost mark for the hypothesis statement itself.

判断单尾还是双尾是假设检验的基础步骤，但许多学生不仔细阅读备择假设就默认双尾。如果 H₁ 含有“不等于”号 (≠)，则检验为双尾；如果 H₁ 指出方向（大于、小于、增加、减少、改善），则为单尾。常见的陷阱是当题目语境明显表明有方向性时，仍用双尾检验，导致临界区域在一侧减半，从而结论错误，或直接因假设陈述失分。

Always write H₀ and H₁ using precise notation, e.g. H₀: p = 0.5, H₁: p > 0.5 for a one‑tailed test. Then match the significance level: in a one‑tailed test, the full α goes into one tail; in a two‑tailed test, halve α for each tail unless the question explicitly uses a two‑tailed critical value.
始终用准确符号写出 H₀ 和 H₁，例如对单尾检验写 H₀: p = 0.5，H₁: p > 0.5。然后匹配显著性水平：单尾检验中整个 α 放入一侧尾部；双尾检验则将 α 均分至两侧，除非题目明确使用双尾临界值。

5. Confusing p‑values with Significance Levels | 混淆 p 值与显著性水平

The p‑value is the probability of obtaining a test statistic at least as extreme as the one observed, assuming H₀ is true. Students often compare it incorrectly: if p < α, reject H₀; if p ≥ α, do not reject H₀. A typical error is rejecting H₀ when p is just above α, or failing to recognise that the p‑value should be compared against the significance level of the test decision. For two‑tailed tests, some candidates forget to double the one‑tailed probability (unless using a direct two‑tailed table).

p 值是在 H₀ 为真的前提下，获得至少与观测值同样极端的检验统计量的概率。学生经常比较错误：若 p < α，则拒绝 H₀；若 p ≥ α，则不拒绝 H₀。典型错误是当 p 刚好略高于 α 时仍拒绝 H₀，或意识不到 p 值需要与检验的显著性水平进行对比。对于双尾检验，一些考生忘记将单尾概率加倍（除非直接使用双尾临界值表）。

When reporting a conclusion, state clearly: ‘Since p = 0.032 < 0.05, we reject H₀ and conclude there is sufficient evidence to support the claim.' Avoid vague phrases like 'the result is significant'.
给出结论时，应明确说明：“由于 p = 0.032 < 0.05，我们拒绝 H₀，认为有充分证据支持该说法。” 避免使用“结果显著”这类模糊表述。

6. Misinterpreting Type I and Type II Errors in Context | 在具体语境中误解第一类与第二类错误

Students can usually recite that a Type I error is rejecting a true H₀, and a Type II error is failing to reject a false H₀, but they struggle to describe these errors in context. For instance, if H₀: a new drug has no effect, a Type I error would mean concluding the drug works when it does not, while a Type II error would mean concluding the drug does not work when it actually does. Losing marks on ‘Describe what a Type I error means in this situation’ is common because candidates give generic definitions instead of context‑rich statements.

学生通常能背诵第一类错误是拒绝了真实的 H₀，第二类错误是未能拒绝错误的 H₀，但难以结合具体语境进行描述。例如，H₀ 为新药无效，第一类错误意味着药实际无效却认为有效；第二类错误意味着药实际有效却认为无效。在“描述此情况下第一类错误的含义”这类题目中丢分很普遍，因为考生给出的是通用定义而非结合情境的表述。

A handy formula: ‘A Type I error is concluding [H₁ in context] when actually [H₀ in context].’ Practice connecting to the scenario every time.
一个实用的句式：“第一类错误是指在实际情况 [H₀ 的情境描述] 时却得出 [H₁ 的情境描述] 的结论。” 每次都练习与情景挂钩。

7. Errors in Sampling and Recognising Bias | 抽样方法与识别偏差的错误

Unit 2 often tests simple random sampling, stratified sampling, systematic sampling, quota sampling, and cluster sampling. A typical mistake is confusing a stratified sample (where the population is divided into groups and a fixed number or proportion is taken from each group) with a quota sample (where interviewers select individuals to fill quotas, introducing potential interviewer bias). Students also mislabel opportunity sampling as random sampling, ignoring that convenience sampling is not random and leads to bias.

第二单元常考简单随机抽样、分层抽样、系统抽样、配额抽样和整群抽样。典型的错误是将分层抽样（将总体分组并从每组抽取固定数量或比例）与配额抽样（调查员自行选择个体填满配额，可能引入调查员偏差）混淆。学生也常把便利抽样标为随机抽样，忽视了便利抽样并非随机且会导致偏差。

When asked to suggest a sampling method, justify based on the need for representativeness, cost, and practicality. Avoid one‑word answers; explain briefly. For example, ‘Stratified sampling would ensure each year group is proportionally represented.’
当被要求建议抽样方法时，需基于代表性、成本和可行性进行说明。避免只写一个词；简单解释。例如：“分层抽样能确保每个年级按比例被代表。”

8. Mishandling the Sum or Difference of Normal Variables | 正态变量和与差的处理错误

If X ~ N(μ₁, σ₁²) and Y ~ N(μ₂, σ₂²) are independent, then aX + bY ~ N(aμ₁ + bμ₂, a²σ₁² + b²σ₂²). The most persistent error is subtracting variances instead of adding when finding the distribution of X − Y. The correct variance is always σ₁² + σ₂² (assuming independence). Students often incorrectly write Var(X − Y) = Var(X) − Var(Y), ending up with an invalid negative variance or a too‑small spread, which subsequently wrecks probability calculations.

若 X ~ N(μ₁, σ₁²) 与 Y ~ N(μ₂, σ₂²) 独立，则 aX + bY ~ N(aμ₁ + bμ₂, a²σ₁² + b²σ₂²)。最常见的顽固错误是求 X − Y 的分布时对方差做减法而非加法。正确的方差总是 σ₁² + σ₂²（假设独立）。学生常错误地写成 Var(X − Y) = Var(X) − Var(Y)，导致出现无效的负方差或过小的离散程度，进而彻底破坏概率计算。

Remember: for independent variables, variances add regardless of whether the linear combination involves + or −. Standard deviations do not add.
牢记：对于独立变量，无论是线性组合中出现 + 还是 −，方差总是相加。标准差则不能相加。

9. Incorrectly Applying the Central Limit Theorem | 错误应用中心极限定理

The Central Limit Theorem states that for a random sample of size n from any distribution with mean μ and variance σ², the sample mean X̄ is approximately N(μ, σ²/n) for sufficiently large n (typically n ≥ 30). Students often misapply CLT by assuming the original population becomes normal, or they use it for very small sample sizes from wildly non‑normal populations. Another mistake is forgetting to divide the variance by n when working with the sample mean, thus using σ² instead of σ²/n for the standard error.

中心极限定理指出，从均值为 μ、方差为 σ² 的任意分布中抽取样本量为 n 的随机样本，样本均值 X̄ 在 n 足够大时（通常 n ≥ 30）近似服从 N(μ, σ²/n)。学生常误用 CLT，以为原始总体因此变得正态，或对来自极非正态总体的小样本也使用该定理。另一个错误是处理样本均值时忘记将方差除以 n，即仍用 σ² 而非 σ²/n 作为标准误。

When using CLT for sample means, always write X̄ ~ N(μ, σ²/n) approximately. For finite populations with sampling without replacement, a finite population correction may be needed, though this is rarely tested at this level.
使用 CLT 处理样本均值时，应始终书写 X̄ 近似服从 N(μ, σ²/n)。对于不重复抽样的有限总体，可能需要有限总体校正因子，但在该层次较少考查。

10. Calculation Slips with Variance and Standard Deviation | 方差与标准差的计算失误

Computing variance using ∑x²/n − (∑x/n)² can lead to arithmetic errors if intermediate results are rounded prematurely. Students also confuse population variance (dividing by n) with sample variance (dividing by n−1) when estimating a population parameter from a sample. In hypothesis testing for a population mean with unknown variance, the t‑distribution requires the sample variance s² with denominator n−1. Using the wrong divisor can shift the test statistic and alter the conclusion.

用公式 ∑x²/n − (∑x/n)² 计算方差时，如果中途过早四舍五入，容易导致算数错误。此外，在用样本估计总体参数时，学生会混淆总体方差（除以 n）和样本方差（除以 n−1）。在总体方差未知的均值假设检验中，t 分布需要使用分母为 n−1 的样本方差 s²。用错分母会改变检验统计量，进而改变结论。

A quick check: if your variance looks unreasonably small or large given the data spread, re‑examine the working and the divisor. When using calculators, know how to extract both σ and s as needed.
快速检查：如果算出的方差相对数据离散程度显得过小或过大，应重新审视计算过程和分母。使用计算器时，要清楚如何根据需要提取 σ 和 s。

11. Misinterpreting Confidence Intervals | 错误解释置信区间

A 95% confidence interval for a mean, such as (48.2, 52.8), means: if we repeated the sampling process many times, 95% of the calculated intervals would contain the true population mean. A frequent misinterpretation is stating that there is a 95% probability that the true mean lies in this particular interval. The true mean is a fixed but unknown constant; it either lies in the interval or it does not. The probability statement describes the long‑run confidence of the method, not a single interval.

均值的95%置信区间，例如 (48.2, 52.8)，其含义为：若多次重复抽样过程，则有95%的计算区间会包含真实的总体均值。常见的错误解释是声称该特定区间含有真实均值的概率为95%。真实均值是一个固定但未知的常数，它要么在区间内，要么不在。概率描述的是方法的长期置信度，而非针对某一个具体区间。

When asked to interpret, use the phrase ‘We are 95% confident that the interval … captures the true mean’ rather than ‘There is a 95% chance …’.
回答解释题时，使用“我们有95%的信心认为该区间……覆盖了真实均值”，而非“有95%的概率……”。

12. Not Using the Correct Statistical Tables | 未使用正确的统计表格

Candidates often get penalised for using the wrong page of the formula booklet or misreading table entries. For binomial cumulative probabilities, they may read the individual probability P(X = k) table when the question requires P(X ≤ k). For Poisson, confusing λ values and reading across wrong columns is also common. In normal distribution tables, mixing up φ(z) for P(Z < z) with the inverse table for percentage points leads to nonsensical z‑values.

考生常因翻错公式手册的表格页或误读表内数值而失分。对于二项分布累积概率，题目要求 P(X ≤ k)，他们却读了个体概率 P(X = k) 的表。对于泊松分布，混淆 λ 值并看错列也屡见不鲜。在正态分布表中，将 P(Z < z) 的 φ(z) 表与百分比点的反查表混用，会得出荒谬的 z 值。

Before any calculation, identify exactly which probability is needed: lower tail, upper tail, between, or critical value. Annotate the table carefully and double‑check the parameter values.
任何计算前，先明确所需概率类型：左尾、右尾、中间还是临界值。仔细在表格上做标注，并反复核对参数值。

---

Published by TutorHao | Mathematics Revision Series | aleveler.com

更多咨询请联系16621398022（同微信）