📚 Oxford AQA FM04 Final MS Jan 21 Question Type Analysis | 牛津AQA FM04 2021年1月评分标准题型全解析
The January 2021 Oxford AQA Further Mathematics Unit 4 (FM04) mark scheme reveals exactly how examiners award marks for Statistics questions. By looking closely at the final mark scheme, we can identify the most common question types, the specific method marks, and the accuracy points that candidates often miss. This analysis breaks down the key patterns so that you can approach similar problems with confidence and a clear understanding of what is required to gain full marks.
牛津AQA 2021年1月进阶数学第四单元（FM04）的评分标准明确显示了考官如何为统计学题目给分。通过仔细分析这份最终评分方案，我们可以识别出最常见的题型、具体的方法分以及考生容易错失的准确度分数。本文将拆解这些关键模式，让你能够自信地应对类似题目，并清楚知道如何拿到满分。
1. Binomial Hypothesis Testing | 二项假设检验
quently in FM04. The mark scheme expects you to define the random variable clearly, e.g. X ~ probability (like P(X ≤ observed)), and A1 for a conclusion in context that uses the word 'significant' appropriately.
二项假设检验在FM04中经常出现。评分标准要求你清楚地定义随机变量，例如 X ~ B(n, p)，并用正确符号写出原假设和备择假设，如 H₀: p = 0.4, H₁: p < 0.4。然后你需要计算临界域或p值。评分标准通常为正确假设给B1分，为找到正确的概率（如 P(X ≤ 观测值)）给M1分，为在上下文中恰当地使用“显著”一词得出结论给A1分。
- Always state the distribution of the test statistic under H₀.
- Check whether the test is one-tailed or two-tailed — this affects how you halve probabilities for comparison.
- A common mistake is to forget to compare the p-value with the significance level, or to write a conclusion that does not refer to the original claim.
- 务必写出在H₀下检验统计量的分布。
- 检查检验是单尾还是双尾——这会影响你如何将概率减半来比较。
- 常见错误包括忘记将p值与显著性水平进行比较，或写了没有回应原声称的结论。
2. Normal Distribution and Critical Regions | 正态分布与临界域
values the use of the standardised test statistic. For a mean test, you will write Z = (x̄ − μ) / (σ/√n) and then find critical z-values from tables. Often you are required to find the critical region in terms of the sample mean. The mark scheme gives M1 for attempting to standardise and A1 for the correct critical value(s). If the population variance is unknown, the t-distribution may be used, and the degrees of freedom must be stated.
当数据用正态分布建模时，评分标准重视标准化检验统计量的使用。对于均值检验，你会写出 Z = (x̄ − μ) / (σ/√n)，然后从表格中查找临界z值。经常要求找出用样本均值表示的临界域。评分标准为尝试标准化给M1分，为正确的临界值给A1分。如果总体方差未知，可能会使用t分布，并且必须说明自由度。
Z = (x̄ − μ₀) / (σ/√n)
The mark scheme highlights that candidates should explicitly check conditions such as normality of the population or a large sample size via the Central Limit Theorem. When a continuity correction is needed (e.g., approximating binomial with normal), failing to include it usually loses the final accuracy mark.
评分标准明确指出，考生应该明确检查条件，例如总体的正态性，或通过中心极限定理得到的大样本量。当需要连续性校正时（比如用正态近似二项），遗漏它通常会失去最后的准确度分。
3. Chi-squared Tests | 卡方检验
− E)² / E, and for the degrees of freedom. You must also combine categories if expected frequencies fall below 5. The final conclusion requires comparing χ² with a critical value from the χ² table. Examiners award M1 for the summation process and A1 for the correct test statistic.
卡方拟合优度检验和独立性检验是FM04常考题。评分标准为计算期望频数、检验统计量 χ² = Σ (O − E)² / E 以及自由度分配分数。如果期望频数低于5，你必须合并分类。最终结论需要将χ²值与卡方分布表中的临界值进行比较。考官为求和过程给M1分，为正确的检验统计量给A1分。
χ² = Σ (O − E)² / E
A mark scheme point frequently missed is the interpretation of the conclusion: 'There is insufficient evidence to reject H₀' is different from 'H₀ is true' e . Stating the result in the context of the problem is essential for the final A1 mark. Also, the Yates' correction is sometimes expected for 2×2 tables.
评分标准中一个经常被忽略的要点是结论的解释：“证据不足以拒绝H₀”与“H₀为真”是不同的。在题目背景下陈述结果对于拿到最后的A1分至关重要。此外，2×2表有时需要使用耶茨校正。
4. Poisson Distribution Applications | 泊松分布应用
ents occurring independently at a constant average rate. The mark scheme rewards stating X ~ ts with Poisson also follow the same structure as binomial: identify λ under H₀, find the probability, and conclude. Sketched distributions or using the Poisson process are sometimes required.
泊松分布题目测试你对以恒定平均速率独立发生的事件进行建模的能力。评分标准奖励写出 X ~ Po(λ)，并使用正确的概率公式或累积表来找到 P(X = k) 或 P(X ≥ k)。用泊松分布进行假设检验也遵循与二项相同的结构：确定H₀下的λ，求出概率并得出结论。有时需要画出分布草图或使用泊松过程。
- When two Poisson variables are combined, use X + Y ~ Po(λ₁ + λ₂) if independent.
- In hypothesis tests, remember to apply the significance level correctly for one-tail or two-tail alternatives.
- The mark scheme may require an exact p-value from tables; interpolation is rarely needed but must be done accurately if specified.
- 当两个泊松变量独立时，用 X + Y ~ Po(λ₁ + λ₂)。
- 在假设检验中，要记得根据单尾或双尾备择假设正确应用显著性水平。
- 评分标准可能要求从表格中获取精确p值；虽然很少需要插值，但如果指定了就必须精确进行。
5. Correlation and Regression | 相关与回归
Product moment correlation coefficient (PMCC) and least squares regression lines are frequently examined. The mark scheme gives method marks for applying the correct formula for Sxx, Syy, Sxy, and for calculating r. You must then test the significance of the population correlation coefficient ρ, using a table of critical values for r. The mark scheme often awards E1 for interpreting the regression coefficient as the change in the response variable per unit change in the explanatory variable.
积矩相关系数（PMCC）和最小二乘回归线是常考内容。评分标准为正确使用 Sxx, Syy, Sxy 的公式、计算 r 等方法步骤给方法分。然后你必须利用 r 的临界值表检验总体相关系数 ρ 的显著性。评分标准常将E1分授予对回归系数的解释：解释变量每变化一个单位，响应变量的平均变化。
| Key formulas | Sxx = Σx² − (Σx)²/n |
| Sxy = Σxy − (Σx)(Σy)/n |
Examiners note that candidates often lose marks by not stating the rounding rules for r, or by forgetting to mention that a causal relationship cannot be inferred solely from correlation. The mark scheme usually contains a specific A1 for 'correlation does not imply causation'.
考官指出，考生常因没有说明 r 的修约规则，或忘记说明不能仅凭相关性推断因果关系而丢分。评分标准通常包含一个具体的A1分，要求写出“相关性并不意味着因果关系”。
6. Probability from Discrete and Continuous Distributions | 离散与连续分布的概率计算
tegration or summation, and for using properties like ΣP(X=x)=1 or ∫f(x)dx=1 to find unknown constants. When dealing with continuous random variables, always sketch the region and clearly write limits.
FM04有些题目纯粹评估计算概率的能力，而不涉及正式的假设检验。你可能会遇到联合概率分布、概率密度函数（pdf）或累积分布函数（cdf）。评分标准为正确的积分或求和，以及利用 ΣP(X=x)=1 或 ∫f(x)dx=1 等性质求出未知常数而给分。处理连续随机变量时，务必画出区域并清晰写出积分上下限。
For discrete uniform or geometric distributions, the mark scheme expects the use of standard formulas for expectation and variance, which gives quick method marks. Always write down E(X) and Var(X) before plugging in numbers.
对于离散均匀分布或几何分布，评分标准期望使用期望和方差的标准公式，这能快速拿到方法分。在代入数字之前，一定要先写出 E(X) 和 Var(X)。
7. Confidence Intervals | 置信区间
core skill assessed in FM04. For a population mean with known variance, the interval is x̄ ± z×(σ/√n). The mark scheme assigns B1 for the correct z-value, M1 for substituting correctly, and A1 for the interval expressed with appropriate endpoints. When variance is unknown, the t-distribution is used, and the degrees of freedom must be correct.
构建并解释置信区间是FM04考核的核心技能。对于已知方差的总体均值，区间为 x̄ ± z×(σ/√n)。评分标准给B1分于正确的z值，M1分于正确代入，A1分于表达得当的端点区间。当方差未知时使用t分布，自由度必须正确。
CI = x̄ ± zα/2 · (σ/√n)
A common missing mark is the interpretation: 'We are 95% confident that the true mean lies between ...' rather than 'There is a 95% probability that the mean is in the interval'. The mark scheme treats the latter as incorrect because the interval is random, not the parameter.
一个常见的丢分点是解释：“我们有95%的信心认为真实均值落在……之间”，而不是“均值有95%的概率落在该区间”。评分标准视后者为错误，因为区间是随机的，而参数不是。
8. Interpreting P-values and Significance | 解释P值与显著性
The January 2021 mark scheme emphasises the correct contextual interpretation of the p-value. Simply stating 'p < 0.05 so reject H₀' is often not enough; you must link the conclusion to the original problem, for example, 'there is evidence to suggest that the new treatment increases recovery rate'. The mark scheme typically awards a final B1 or A1 for a contextualised conclusion that matches the alternative hypothesis.
2021年1月的评分标准强调了在上下文中正确解释p值。仅仅陈述“p < 0.05 所以拒绝 H₀”往往不够；你必须将结论与原始问题联系起来，比如，“有证据表明新处理方法提高了康复率”。评分标准通常会为符合备择假设的上下文结论给一个最终的B1或A1分。
The mark scheme also penalises double negatives or statements that suggest accepting H₀. You should write 'do not reject H₀' or 'insufficient evidence to reject H₀'. Using the actual p-value from a calculator, rounded to two or three significant figures, is preferred.
评分标准还对双重否定或暗示接受H₀的陈述扣分。你应该写“不拒绝H₀”或“证据不足以拒绝H₀”。优先使用计算器给出的实际p值，修约到两位或三位有效数字。
9. Working with Probability Density Functions (pdfs) and Medians | 概率密度函数与中位数
sult equals 1, which typically earns an M1. The mark scheme then awards M1 for setting ∫_am n m, and A1 for the correct value. Remember to consider the piecewise nature of many pdfs.
FM04常包含要求验证f(x)是有效概率密度函数，然后求中位数、众数或四分位数的题目。验证时，在定义域上积分f(x)并展示结果等于1，通常可得M1分。接着评分标准为设 ∫_am f(x) dx = 0.5 以求中位数 m 给M1分，为正确值给A1分。注意许多pdf是分段函数。
If the pdf is given in two parts, you must check which section the median falls into. A frequent error is to integrate the wrong function or to forget the lower bound constant. The mark scheme appreciates a clearly drawn diagram to support the reasoning.
如果pdf分成两段给出，你需要检查中位数落在哪一段。常见错误是积分了错误的函数，或忘记了积分常数的下限。评分标准欣赏能清晰画图辅助推理的答案。
10. Marking Point Strategy: Method and Accuracy | 评分点策略：方法与准确性
Understanding how examiners allocate B, M, and A marks can transform your exam technique. The FM04 mark scheme shows that B marks are given for independent pieces of knowledge, such as stating the correct distribution or a condition. M marks are for method — you must show a clear attempt at the correct process. A marks depend on the accuracy of the final answer and are often only awarded if the corresponding M mark has been earned. An answer given without working may earn zero method marks.
理解考官如何分配B分、M分和A分可以改变你的考试策略。FM04评分标准显示，B分是给独立的知识点，比如写出正确的分布或一个条件。M分是方法分——你必须展示正确过程的一个清晰尝试。A分取决于最终答案的准确性，通常只有在相应的M分已拿到的情况下才给。没有解题过程的答案可能拿不到任何方法分。
- Always write down the distribution and parameters before any calculation; this often secures a B1.
- For hypothesis tests, structure your answer as: Hypotheses, Test statistic, Critical region/p-value, Conclusion.
- If you use a calculator to find a critical value directly, show the input or sketch to prove your method.
- Check rounding: the mark scheme often expects final answers to 3 significant figures unless stated otherwise.
- 在任何计算之前，写下分布及其参数；这通常能保住一个B1分。
- 对于假设检验，将答案结构化为：假设，检验统计量，临界域/p值，结论。
- 如果你直接使用计算器得出临界值，展示输入过程或画草图来证明你的方法。
- 检查修约：除非另有说明，评分标准通常要求最终答案保留3位有效数字。
The mark scheme also reveals that partially correct methods can still earn marks if the error is not fundamental. However, early arithmetic mistakes that simplify the problem too much may prevent subsequent marks. Practise showing all steps clearly, exactly as the mark scheme outlines.
评分标准还揭示，即使方法有错误，如果不是根本性的，部分正确的方法仍能得分。然而，过度简化问题前期的算术错误可能导致后续分数全丢。要按照评分标准的要求，清晰展示所有步骤进行练习。
Published by TutorHao | Further Mathematics Revision Series | aleveler.com
更多咨询请联系16621398022（同微信）
或截图后用「扫一扫 → 相册」选择本图