A-Level Further Maths Unit 5 June 2022: Common Mistakes Summary | A-Level 进阶数学单元5 2022年6月考试易错点总结

📚 A-Level Further Maths Unit 5 June 2022: Common Mistakes Summary | A-Level 进阶数学单元5 2022年6月考试易错点总结

The June 2022 A-Level Further Mathematics Unit 5 paper, typically covering Statistics in the AQA specification, challenged many students with subtle conceptual pitfalls. Whether you sat this exam or are preparing for a future sitting, understanding where marks are most often lost is essential. This article distils recurring errors from student scripts, focusing on hypothesis testing, probability distributions, the Poisson model, normal approximations, chi‑squared tests, conditional probabilities, and the interpretation of statistical language.

2022年6月的A-Level进阶数学单元5试卷（AQA考试局通常对应统计学内容）给不少考生挖下了概念陷阱。无论是已经考过还是正在备考，弄清楚最容易丢分的地方都非常关键。本文提炼了答卷中反复出现的错误，重点涵盖假设检验、概率分布、泊松模型、正态近似、卡方检验、条件概率以及对统计术语的理解。

1. Misreading the Direction of a Hypothesis Test | 混淆假设检验的方向

Many candidates correctly stated H₀ and H₁ but then selected the wrong tail for the test. In a two‑tailed test, the significance level is halved; if the question describes a one‑tailed scenario (e.g. “test whether the mean has increased”), the entire significance level goes into the appropriate tail. In the June 2022 paper, some students used a two‑tailed critical value when a one‑tailed test was required, losing the first method mark immediately.

很多考生能正确写出 H₀ 和 H₁，却在选择检验方向上出错。在双侧检验中，显著性水平要平分；如果题目明确是单侧场景（例如“检验均值是否上升”），显著性水平就应全部放在相应的尾部。在2022年6月试卷中，部分同学在需要单侧检验时却用了双侧临界值，一下子丢掉了第一个方法分。

2. Confusing the Null and Alternative Hypotheses | 混淆原假设与备择假设

Hypotheses must be about population parameters, not sample statistics. A common mistake was writing H₀: x̄ = 50 instead of H₀: μ = 50. Another error was reversing H₀ and H₁ – for instance, putting the claim to be tested in the null rather than the alternative. The null hypothesis always contains the “no change” or “no effect” statement with equality.

假设必须针对总体参数，而不能针对样本统计量。一个常见错误是把 H₀ 写成 x̄ = 50，而不是 μ = 50。还有同学把 H₀ 和 H₁ 的位置弄反了——例如把要检验的主张放在了原假设中，而原假设应始终表示“没有变化”或“没有效应”且包含等号。

3. Mishandling the Poisson Distribution’s Conditions | 误用泊松分布的前提条件

When modelling with the Poisson distribution, events must be independent, occur singly, and at a constant average rate. In a question about accidents per week, some students used a binomial model or failed to state why Poisson was suitable. Even when the model was given, forgetting to mention that the mean rate is constant cost marks in the interpretation part.

运用泊松分布建模时，事件必须独立、单个发生且平均发生率恒定。在一道关于每周事故数的题目中，有的学生用了二项分布模型，或者没有说明为什么适合用泊松分布。即使模型已给出，忘记说明平均发生率是恒定的也会在解释部分失分。

4. Normal Approximation to Poisson: Continuity Correction | 泊松分布的正态近似：连续性校正

When approximating a Poisson distribution with a normal distribution, a continuity correction is essential. A typical error was to write P(X > 10) ≈ P(Y > 10) without the half‑unit shift. The corrected form should be P(X > 10) ≈ P(Y > 10.5). Missing the continuity correction led to completely inaccurate probabilities and no credit for the method.

用正态分布近似泊松分布时，连续性校正必不可少。一个典型错误是把 P(X > 10) 写成 P(Y > 10)，缺乏半个单位的平移。正确形式应为 P(X > 10) ≈ P(Y > 10.5)。遗漏连续性校正会导致概率完全不准确，方法分全无。

5. Forgetting the Variance Equals the Mean in Poisson | 忘记泊松分布中方差等于均值

In the Poisson distribution, E(X) = Var(X) = λ. Many students used λ for the mean but incorrectly took the variance as λ/n or some other expression. When calculating probabilities or carrying out a hypothesis test for a Poisson rate, using the wrong variance made the standard deviation incorrect and the final conclusion invalid.

在泊松分布中，E(X) = Var(X) = λ。很多学生把均值设为 λ，却错误地将方差写成了 λ/n 或其他表达式。在计算概率或进行泊松率的假设检验时，方差写错会使标准差出错，最终导致结论无效。

6. Degrees of Freedom in the Chi‑Squared Test | 卡方检验中的自由度

The chi‑squared goodness‑of‑fit test’s degrees of freedom are (number of categories) – 1 – (number of estimated parameters). In the June 2022 paper, a question involved estimating the Poisson parameter λ from the data. Many candidates forgot to subtract an extra degree of freedom for the estimated parameter, giving ν = k – 1 instead of ν = k – 2. This error changed the critical value and often the test conclusion.

卡方拟合优度检验的自由度为（类别数）– 1 –（估算的参数个数）。2022年6月试卷中有一题需要从数据中估算泊松参数 λ。很多考生忘记为估算的参数减去一个额外的自由度，得到 ν = k – 1 而不是 ν = k – 2。这个错误会改变临界值，常常也改变了检验结论。

7. Combining Categories in Chi‑Squared Incorrectly | 错误合并卡方检验中的类别

When expected frequencies are less than 5, categories must be combined. Errors included combining non‑adjacent categories without justification, failing to combine when needed, or combining in a way that left a single category with an expected frequency below 5. The rule is to keep combining until all expected values are at least 5, and the new categories make contextual sense.

当期望频数小于 5 时，必须合并类别。常见错误包括：没有理由地合并不相邻的类别；需要合并时没有合并；或合并后仍有个别类别的期望频数低于 5。正确做法是持续合并，直到所有期望值都至少为 5，并且新的类别在背景上说得通。

8. Conditional Probability Formula Reversal | 条件概率公式颠倒

Given P(A|B) and asked for P(B|A), many students simply swapped the numbers or used the wrong form of Bayes’ theorem. The correct relationship is P(B|A) = [P(A|B) × P(B)] / P(A). In one exam question, candidates confused P(A ∩ B) with P(A|B) and lost all the marks for a seemingly simple calculation.

9. Misinterpreting p‑values and Conclusions | 误解 p 值与结论

A p‑value less than the significance level leads to rejection of H₀. Some students stated “accept H₀” or “prove H₀” when they should have said “do not reject H₀”. Others claimed a result was significant when the p‑value was actually above the significance threshold. Correct phrasing is critical in the non‑technical conclusions: “There is sufficient evidence to reject H₀” or “There is insufficient evidence to reject H₀.”

p 值小于显著性水平才能拒绝 H₀。有些考生在不应拒绝时却说“接受 H₀”或“证明 H₀”，而应说“没有充分证据拒绝 H₀”。还有些人明明 p 值高于阈值却声称结果显著。非技术性结论的表述极为重要：“有充分证据拒绝 H₀”或“没有充分证据拒绝 H₀”。

10. Confusing Standard Deviation and Standard Error | 混淆标准差与标准误

When testing a sample mean, the standard deviation of the sample mean is σ/√n (standard error), not σ. In a question about the mean weight of packets, candidates frequently used σ directly in the test statistic formula, forgetting to divide by √n. This oversight produced a test statistic that was too small, often leading to a non‑significant result when the correct calculation gave a significant one.

在检验样本均值时，样本均值的标准差是 σ/√n（标准误），而不是 σ。在一道关于包装件平均重量的题目中，考生频繁直接在检验统计量公式中用 σ，忘了除以 √n。这个疏忽使得检验统计量过小，常常得到不显著的结果，而正确计算本应是显著的。

11. Using the Wrong Distribution for Sums or Differences | 对于和或差使用错误的分布

When dealing with the sum or difference of two independent normal variables, variances add (or add in the difference). A common mistake was subtracting variances when finding the distribution of X – Y. The variance of X – Y is Var(X) + Var(Y), not Var(X) – Var(Y). This error appeared in questions about comparative weights or lengths, costing several marks in distribution setup.

处理两个独立正态变量之和或差的分布时，方差是相加的（在求差时也是相加）。一个常见错误是在求 X – Y 的分布时减去方差。X – Y 的方差应为 Var(X) + Var(Y)，而非 Var(X) – Var(Y)。这一错误常出现在比较重量或长度的题目中，导致分布建立的数分尽失。

12. Lack of Precision in Final Answers and Units | 最终答案和单位的精度缺失

The mark scheme frequently demands answers to a specified degree of accuracy (e.g. 3 significant figures). Many students lost accuracy marks by rounding too early in intermediate steps, or by omitting units in contextual questions. For probabilities, giving a percentage when a decimal is required, or vice versa, also lost the final answer mark. Always read the question’s instructions carefully.

评分方案常常要求答案达到指定的精度（例如 3 位有效数字）。很多学生因为中间步骤过早四舍五入而丢了精度分，或者在情境题中遗漏单位。对于概率题，要求写小数却给出百分数（或相反）也会失去最终答案分。务必仔细阅读题目要求。

Published by TutorHao | Further Maths Revision Series | aleveler.com

更多咨询请联系16621398022（同微信）

A-Level Further Maths Unit 5 June 2022: Common Mistakes Summary | A-Level 进阶数学单元5 2022年6月考试易错点总结