📚 Topic Test: Common Mistakes in OxfordAQA International AS Level Further Mathematics (9665) Statistics | 专题测试：牛津AQA国际AS进阶数学(9665)统计易错点总结
The OxfordAQA International AS Level Further Mathematics (9665) Statistics topic test assesses a range of discrete and continuous distributions, hypothesis testing, and chi‑squared procedures. Students often lose marks not because they lack understanding, but because small, recurring errors creep into their work. This article summarises the most common pitfalls and shows you how to avoid them.
牛津AQA国际AS进阶数学(9665)统计专题测试涵盖了多种离散与连续分布、假设检验以及卡方检验。学生失分往往不是因为不理解，而是因为一些反复出现的小错误。本文总结最常见的易错点，并说明如何避免这些错误。
1. Misapplying Expectation and Variance Formulas for Discrete Random Variables | 1. 离散随机变量期望与方差公式的误用
Students regularly use the formula Var(X) = E(X²) – [E(X)]² correctly, but slip up when dealing with linear transformations. It is essential to remember that for constants a and b, Var(aX + b) = a² Var(X). The constant b does not affect the variance at all. Many mistakenly believe Var(aX + b) = a Var(X) + b, or forget to square the multiplier.
学生通常能够正确使用 Var(X) = E(X²) – [E(X)]²，但在处理线性变换时容易出错。必须记住，对于常数 a 和 b，有 Var(aX + b) = a² Var(X)。常数 b 完全不影响方差。许多学生误以为 Var(aX + b) = a Var(X) + b，或者忘记对乘数进行平方。
If Y = aX + b, then E(Y) = aE(X) + b and Var(Y) = a² Var(X)
若 Y = aX + b，则 E(Y) = aE(X) + b 且 Var(Y) = a² Var(X)
Another common slip occurs when combining independent random variables: Var(X ± Y) = Var(X) + Var(Y). The ‘+’ sign remains even when subtracting Y, because variances add regardless of the sign. Students occasionally write Var(X – Y) = Var(X) – Var(Y), which is incorrect and can lead to negative variance values.
另一个常见错误发生在独立随机变量的组合：Var(X ± Y) = Var(X) + Var(Y)。即使减去 Y，加号依然保留，因为方差始终相加。有些学生会写成 Var(X – Y) = Var(X) – Var(Y)，这是不正确的，甚至可能导致负的方差值。
2. Confusing the Conditions for a Poisson Distribution | 2. 混淆泊松分布的使用条件
A Poisson distribution is appropriate when events occur independently at a constant average rate in a fixed interval of space or time. Students often apply it when the probability of success is not small, or when the number of trials is not large. The key conditions are: events occur singly, randomly, and at a constant mean rate. Using Poisson when the variance is far greater than the mean (overdispersion) or when events are not independent (e.g. contagious occurrences) is a serious modelling error.
当事件在固定空间或时间区间内以恒定平均发生率独立发生时，适合使用泊松分布。学生常常在成功概率不低或试验次数不大的情况下错误套用泊松分布。关键条件是：事件单独发生、随机发生且均值恒定。如果方差远大于均值（过度离散）或事件不独立（例如传染性发生），却仍使用泊松分布，这是一个严重的建模错误。
In exam questions, always check whether the parameter λ is the mean for the whole interval you are considering. If the interval length changes, λ must be scaled proportionally. For example, if λ = 4 per hour, then for a 30‑minute period λ = 2. Carelessly using the original λ is a frequent mistake.
在考题中，务必检查参数 λ 是否是所考虑的整体区间的均值。如果区间长度改变，λ 必须按比例缩放。例如，若 λ = 4 每小时，那么30分钟时段的 λ = 2。粗心地直接使用原来的 λ 是一个常见错误。
3. Poisson Approximation to the Binomial: Neglecting the Conditions | 3. 二项分布的泊松近似：忽视条件
When n is large and p is small, the binomial distribution B(n, p) can be approximated by a Poisson distribution with parameter λ = np. Students often apply this approximation when p is not small (e.g. p = 0.4), or when n is small, which leads to inaccurate probabilities. The rule of thumb is n > 50 and np < 5, though some specifications accept n > 20 and p < 0.1.
当 n 很大而 p 很小时，二项分布 B(n, p) 可以用参数 λ = np 的泊松分布来近似。学生往往在 p 不小（例如 p = 0.4）或 n 较小的时候使用这个近似，导致概率结果不准确。经验法则要求 n > 50 且 np < 5，虽然部分大纲接受 n > 20 且 p < 0.1。
A related mistake is forgetting that the approximation replaces a discrete distribution with another discrete distribution; no continuity correction is needed. Nevertheless, learners must state the approximation clearly and justify it by checking the conditions.
一个相关的错误是忘记这个近似是用一个离散分布替换另一个离散分布，因此不需要连续性校正。但是，学生必须清晰地陈述这个近似，并通过检验条件来说明理由。
4. Normal Approximation Errors: Missing Continuity Correction | 4. 正态近似错误：遗漏连续性校正
When approximating a discrete distribution (binomial or Poisson) by the normal distribution, a continuity correction is essential. For a binomial X ~ B(n, p), the normal approximation is N(np, np(1-p)). To find P(X ≤ 12), you must use the boundary 12.5, not 12. Many candidates calculate z‑scores without this adjustment and lose marks, especially in hypothesis tests for proportions.
用正态分布近似离散分布（二项或泊松）时，必须进行连续性校正。对于二项分布 X ~ B(n, p)，正态近似为 N(np, np(1-p))。求 P(X ≤ 12) 时，必须用边界 12.5，而不是 12。许多考生不进行这个调整就直接计算 z 值，从而失分，尤其是在比例假设检验中。
P(X ≤ x) ≈ P(Y < x + 0.5) after correction, where Y ~ N(μ, σ²)
校正后 P(X ≤ x) ≈ P(Y < x + 0.5)，其中 Y ~ N(μ, σ²)
Similarly, for a Poisson X ~ Po(λ), the normal approximation is N(λ, λ). The continuity correction must be applied in the same way. Students often overlook the requirement that λ should be reasonably large (usually λ > 15) for the approximation to be valid.
同理，对于泊松分布 X ~ Po(λ)，正态近似为 N(λ, λ)。必须以同样的方式应用连续性校正。学生常常忽略近似有效的条件：λ 应该足够大（通常 λ > 15）。
5. Hypothesis Testing: Mixing p‑value and Critical Region Methods | 5. 假设检验：混淆 p 值法与临界区域法
Two valid approaches exist for hypothesis testing: comparing the p‑value with the significance level α, or comparing the test statistic with the critical value(s). Errors occur when students attempt to combine these methods or write mismatched conclusions. If you use the p‑value method, you must state: ‘reject H₀ if p < α'. If you use the critical region, you must compare the calculated statistic with the critical value directly.
假设检验有两种有效方法：比较 p 值与显著性水平 α，或比较检验统计量与临界值。如果学生试图混合这两种方法，或者写成不匹配的结论，就会出错。如果使用 p 值法，必须陈述：“若 p < α，则拒绝 H₀”。如果使用临界区域法，则必须将计算的统计量与临界值直接比较。
A further pitfall is writing non‑standard conclusions. Acceptable wording includes: ‘There is sufficient evidence to reject H₀’ or ‘There is insufficient evidence to reject H₀’. Never write ‘accept H₀’ or ‘prove H₀’. The test only provides evidence against the null hypothesis, not proof of the alternative.
另一个陷阱是写出不规范的结论。可接受的措辞包括：“有充分证据拒绝 H₀”或“没有充分证据拒绝 H₀”。绝不能写“接受 H₀”或“证明 H₀”。检验只是提供反对零假设的证据，并非证明备择假设。
6. One‑tailed versus Two‑tailed Tests: Choosing the Wrong Alternative | 6. 单尾检验与双尾检验：选错备择假设
Choosing between a one‑tailed and a two‑tailed test depends on the wording of the problem. If the question asks whether a parameter ‘has increased’ or ‘is greater than’, use a one‑tailed test. If it asks whether it ‘has changed’ or ‘is different from’, use a two‑tailed test. Using the wrong test type will halve or double the p‑value region and lead to an incorrect conclusion.
选择单尾还是双尾检验取决于问题的措辞。如果题目问参数是否“增加”或“大于”，则使用单尾检验。如果问是否“改变”或“不同于”，则使用双尾检验。检验类型选错会使 p 值区域减半或加倍，导致结论错误。
When conducting a two‑tailed test of a binomial proportion or Poisson mean, the critical region is split equally between both tails. Students often make the mistake of comparing the p‑value with α/2 instead of α, or comparing the test statistic with only one critical value. Always adjust the interpretation according to the type of test.
在对二项比例或泊松均值进行双尾检验时，临界区域平均分配在两端。学生常犯的错误是用 α/2 而不是 α 与 p 值比较，或者只与一个临界值比较。必须根据检验类型调整解释方式。
7. Chi‑squared Goodness of Fit: Mishandling Expected Frequencies and Degrees of Freedom | 7. 卡方拟合优度：预期频数与自由度处理不当
For a χ² goodness‑of‑fit test, expected frequencies are calculated as n × pᵢ, where n is the total number of observations and pᵢ is the hypothesised proportion for category i. A frequent mistake is using observed frequencies again instead of expected frequencies, or forgetting to ensure all expected frequencies are at least 5. If any expected frequency is below 5, categories must be combined until the requirement is met, and the degrees of freedom adjusted accordingly.
对于 χ² 拟合优度检验，预期频数按 n × pᵢ 计算，其中 n 是观测总数，pᵢ 是类别 i 的假设比例。常见错误是再次使用观测频数代替预期频数，或者忘记确保所有预期频数至少为 5。如果任何预期频数低于 5，必须合并类别直到满足要求，并且相应调整自由度。
Degrees of freedom: For a goodness‑of‑fit test with no estimated parameters from the data, ν = k – 1, where k is the number of categories after merging. If parameters are estimated, subtract one degree of freedom for each estimated parameter. Misjudging ν leads to an incorrect critical value and a false conclusion. Remember to subtract the correct number of restrictions.
自由度：对于没有从数据中估计参数的拟合优度检验，ν = k – 1，其中 k 是合并后的类别数。如果估计了参数，每估计一个参数再减去一个自由度。误判 ν 会导致临界值错误，并得出错误结论。务必减去正确的约束个数。
8. Sampling Distribution of the Mean and the Central Limit Theorem | 8. 样本均值的抽样分布与中心极限定理
The sampling distribution of the sample mean X̄ plays a central role in hypothesis tests for a normal mean. Many candidates incorrectly state that the population must be normally distributed for the sample mean to be normal. In fact, if the sample size n is large (typically n ≥ 30), the Central Limit Theorem ensures X̄ is approximately normally distributed regardless of the population shape, provided the observations are independent and the population variance is finite.
样本均值 X̄ 的抽样分布在正态均值假设检验中起着核心作用。许多考生错误地认为，要让样本均值为正态，总体必须呈正态分布。事实上，如果样本容量 n 较大（通常 n ≥ 30），根据中心极限定理，无论总体形状如何，X̄ 都近似服从正态分布，前提是观测值独立且总体方差有限。
A related mistake is confusing the standard deviation of the population σ with the standard error of the mean, which is σ/√n. When the population variance is unknown and n is large, the sample standard deviation s replaces σ, but candidates sometimes forget to divide by √n when calculating the test statistic. The correct form is Z = (X̄ – μ) / (σ/√n) or approximate using s if σ is unknown and n is large.
一个相关错误是混淆总体标准差 σ 与均值标准误差 σ/√n。当总体方差未知且 n 足够大时，用样本标准差 s 代替 σ，但是考生有时在计算检验统计量时忘记除以 √n。正确形式为 Z = (X̄ – μ) / (σ/√n)，或者若 σ 未知且 n 较大则用 s 近似。
9. Linear Combinations of Independent Normal Variables | 9. 独立正态变量的线性组合
When combining independent normal random variables, the result is also normally distributed. For X ~ N(μ₁, σ₁²) and Y ~ N(μ₂, σ₂²) independently, aX + bY + c is N(aμ₁ + bμ₂ + c, a²σ₁² + b²σ₂²). Students often misuse the variance formula by applying the same sign as the means, leading to a negative variance term. Even when subtracting variables, the variance coefficients are always squared and added.
当组合独立的服从正态分布的随机变量时，其结果也是正态分布。对于相互独立的 X ~ N(μ₁, σ₁²) 和 Y ~ N(μ₂, σ₂²)，有 aX + bY + c ~ N(aμ₁ + bμ₂ + c, a²σ₁² + b²σ₂²)。学生经常在方差公式中使用与均值相同的符号，从而导致负的方差项。即使变量是相减，方差系数也总是平方后相加。
Another oversight is assuming that the sum of two non‑normal random variables is automatically normal without checking the sample size or the Central Limit Theorem. For small samples from non‑normal populations, the sum may not be well approximated by a normal distribution.
另一种疏忽是认为两个非正态随机变量的和自动服从正态分布，而不检查样本容量或中心极限定理。对于来自非正态总体的小样本，其和可能无法很好地用正态分布近似。
10. Misreading the Alternative Hypothesis in Context | 10. 在具体情境中误读备择假设
Context is everything in statistical hypothesis testing. A test of a binomial proportion p may be phrased as ‘test whether the proportion of faulty items has reduced after a process change’. This clearly indicates a one‑tailed test with H₁: p < p₀. Many learners simply set up a two‑tailed test out of habit, or misinterpret 'has reduced' as a two‑sided change. Always highlight the direction keywords: reduced, increased, higher, lower, different, changed.
在统计假设检验中，语境至关重要。某个二项比例 p 的检验可能表述为“检验流程变更后次品率是否降低”。这清楚地表明是单尾检验，备择假设为 H₁: p < p₀。许多学生出于习惯直接设定双尾检验，或将“降低”误读为双边变化。务必标示出方向性关键词：降低、增加、更高、更低、不同、改变。
Similarly, in a chi‑squared test for independence or goodness of fit, the alternative hypothesis is always ‘the model does not fit’ or ‘there is an association’. Writing a one‑sided alternative for a χ² test is a conceptual error because the test statistic measures squared deviations and is inherently non‑directional.
类似地，在独立性或拟合优度卡方检验中，备择假设总是“模型不拟合”或“存在关联”。为 χ² 检验写出单侧备择假设是一个概念性错误，因为检验统计量度量的是平方偏差，本质上就是不区分方向的。
Published by TutorHao | AS Further Mathematics 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