📚 A-Level Maths Unit 3 Mark Scheme June 2022: Common Mistakes Summary | A-Level 数学 2022年6月单元3评分方案易错点总结
Analysing examiner reports and mark schemes from the June 2022 Unit 3 assessment reveals a consistent set of errors that cost candidates valuable marks. Whether the paper focused on statistics, mechanics or pure maths applications, many mistakes stem from the same conceptual misunderstandings, sloppy algebraic manipulation or misinterpretation of question wording. This article compiles the most frequent pitfalls identified in the mark scheme for that session, explains why they occur and shows how to avoid them. By addressing these weak spots, you can sharpen your technique and boost your final score significantly.
分析2022年6月单元3考试的考官报告和评分方案后可以发现，有一系列反复出现的错误让考生白白丢分。无论试卷侧重于统计、力学还是纯数应用，许多扣分点都源于相同的概念误解、粗心的代数运算或对题干措辞的误读。本文汇总了该场考试评分方案中暴露出的最高频错误，解释其成因并给出避错方法。补齐这些短板能够有效提升你的答题质量，从而显著提高最终得分。
1. Misreading Conditional Probability in Tree Diagrams | 树图中条件概率的误读
A typical question presented a scenario with successive events and asked for P(A given B). Many candidates correctly drew a probability tree but then used the wrong branch values when applying the conditional formula. The mark scheme required P(A∩B)/P(B), but students often took P(B) directly from a branch that was itself conditional on another event, forgetting that the denominator must be the total probability of B occurring across all possible pathways. This mistake was especially common in 'without replacement' contexts, where the probabilities on the second layer of branches are inherently conditional.
一道典型题目给出了连续发生的事件并要求计算 P(A given B)。许多考生正确画出了概率树，但在套用条件概率公式时错误地选用了分支上的数值。评分方案要求使用 P(A∩B)/P(B)，但学生经常直接从某条分支取 P(B)，而该分支本身是以另一事件为条件的，他们忘记了分母必须是 B 在所有可能路径中发生的总概率。这种错误在“无放回”情境中尤为常见，因为此时第二层分支上的概率本身就是条件概率。
To avoid this, always list all the ways B can happen, sum their probabilities and then divide the intersection by that total. Draw a separate working line for the denominator to remind yourself that it is not simply the first branch figure you see.
要避免这个错误，始终罗列出事件 B 发生的所有途径，求和得到 P(B)，再用交集除以该总值。单独写出分母的计算行，提醒自己它并非你看到的第一根分支上的数字。
2. Confusing Independent and Mutually Exclusive Events | 混淆独立事件与互斥事件
In questions testing probability laws, candidates often used P(A∪B) = P(A)+P(B) for events that were only known to be independent, or applied P(A∩ xclusive. The mark scheme explicitly penalised this confusion. For independent events, P(A∩B) = P(A)P(B), but for mutually exclusive events, P(A∩ drawing a card simultaneously versus a single draw that cannot yield two outcomes at once.
在考查概率法则的题目中，考生常常对仅已知为独立的事件错误套用 P(A∪B) = P(A)+P(B)，或是在事件互斥时使用 P(A∩B) = P(A)×P(B)。评分方案明确针对此类混淆进行了扣分。对于独立事件，P(A∩B) = P(A)P(B)，但对于互斥事件，P(A∩B) = 0。2022年6月的试卷中有若干小问要求学生根据情境判断应使用哪条规则，例如掷骰子与抽牌同时进行的情境对比一次抽取得不到两个结果的互斥情境。
Read the question stem carefully: words like 'rolled at the same time' or 'independently' signal independence; phrases like 'either ... or' and 'cannot both occur' indicate mutual exclusivity. Write down the property you are using next to your working to help the examiner follow your logic.
仔细阅读题干：如 “同时掷出” 或 “独立地” 这样的字眼暗示独立性；而 “要么……要么……” 以及 “不能同时发生” 则表明互斥。在你的解题步骤旁注明所使用的性质，以便考官理解你的推理。
3. Binomial Distribution Calculator Misuse: n and p Confusion | 二项分布计算器误用：n 与 p 的混淆
Many candidates lost marks by inputting the wrong parameters into the binomial probability formula or their calculator. The mark scheme showed that errors most often occurred when the question gave the number of trials and the probability of failure instead of success. Students either failed to convert the given proportion, or swapped n and p, or used 1-p incorrectly. Another recurring issue was using the binomial distribution when the trials were not independent, e.g. selections without replacement from a small population, where the hypergeometric model is more appropriate, though in A-level statistics the population is usually considered large enough for a binomial approximation.
许多考生因为向二项概率公式或计算器输入了错误的参数而丢分。评分方案显示，当题目给出试验次数和失败概率而非成功概率时，错误最为频发。学生要么未转换所给比例，要么将 n 与 p 颠倒了，要么错误地使用了 1-p。另一个反复出现的问题是在试验并不独立的情况下直接套用二项分布，比如从一个小总体中进行无放回抽样，此时更适宜采用超几何模型，不过在A-Level统计学中通常默认总体足够大，可作二项近似。
Be systematic: write X ~ problem defines them. Check whether the given probability is for the outcome you count as 'success'. If a question says 'probability of a defective item is 0.02' on-defectives out of 10, you can redefine success or calculate the equivalent with p=0.98.
务必条理清晰：按题目定义写出 X ~ B(n, p)，并填入正确的 n 与 p。核实给定的概率是否对应于你所计数的“成功”结果。若题目说“次品率为 0.02”，而你计数的是次品，则 p=0.02；若需计算 10 个产品中恰好有 3 个合格品的概率，你可以重新定义成功，或改用 p=0.98 计算等价概率。
4. Omitting Continuity Corrections in Normal Approximations | 正态近似中忽略连续性校正
When a binomial distribution was approximated by a normal distribution, the mark scheme required a continuity correction. A large number of scripts either omitted it entirely or applied it incorrectly, e.g. using P(X < 10) becoming P(Z < (10-μ)/σ) instead of P(Z < (9.5-μ)/σ). The exact half-unit adjustment depends on the inequality: 'less than' becomes 'less than or equal to n-0.5', 'at least' becomes 'greater than or equal to n-0.5', etc. Candidates who skipped this step lost accuracy marks even if their standardisation was otherwise correct.
当用正态分布近似二项分布时，评分方案要求进行连续性校正。大量答卷要么完全遗漏了校正，要么应用失当，例如将 P(X < 10) 转换为 P(Z < (10-μ)/σ) 而非 P(Z < (9.5-μ)/σ)。精确的半单位调整取决于不等号类型：“小于” 应变为 “小于或等于 n-0.5”，“至少” 变为 “大于或等于 n-0.5” 等。跳过这一步的考生即使标准化过程正确，也会失去答案准确度分。
Draw a small number line and shade the region of interest. For a discrete value 10, 'P(X < 10)' means 9 or fewer, so the continuous cut-off is 9.5. Memorising the rule is less reliable than visualising the gap between bars on a histogram. Always state the continuity correction explicitly in your working.
画一条小小的数轴并涂出你所关注的区域。对于离散值 10，“P(X < 10)” 意味着取 9 或更低的值，因此连续分割点为 9.5。相比死记硬背规则，可视化直方图条形间的间隙更为可靠。务必在步骤中明确写出连续性校正的过程。
5. Hypothesis Test Conclusions: Reject H₀ or not? | 假设检验结论：拒绝 H₀ 还是不能拒绝？
The June 22 mark scheme highlighted that many students performed the correct critical value or p-value comparison but then wrote a conclusion that contradicted their findings. Common wording pitfalls included stating 'accept H₀' instead of 'do not reject H₀', or claiming there is 'sufficient evidence for H₁' when the test statistic did not fall in the critical region. Examiners also penalised answers that lacked context, e.g. simply writing 'reject H₀' without relating it back to the original claim about the population mean or proportion.
2022年6月的评分方案强调，许多学生虽正确进行了临界值或 p 值的比较，但随后得出的结论却与他们的发现相矛盾。措辞上的常见陷阱包括写成“接受 H₀” 而非 “不拒绝 H₀”，或是在检验统计量未落入拒绝域时声称“有充分证据支持 H₁”。考官还扣去了缺乏情境的答案的分数，例如仅仅写下“拒绝 H₀” 而未将其与原始关于总体均值或比例的说法相联系。
A bulletproof conclusion template: 'Since [test statistic] >/< [critical value] (or p < α), there is sufficient evidence at the 5% significance level to reject H₀ and conclude that [restate H₁ in context].' Follow this structure and you will always secure the final mark.
一个不会出错的结论模板：“由于 [检验统计量] >/< [临界值]（或 p < α），在 5% 的显著性水平下有充分证据拒绝 H₀，并可得出……[在上下文重申 H₁]。” 遵循此结构你总能拿到最终的结论分。
6. Expected Value and Variance Formula Slips | 期望值与方差公式的滑动
k scheme showed errors such as writing Var(3X) = 3Var(X) instead of 9Var(X), or forgetting that adding a constant does not change variance. In contexts where the probability distribution was given in a table, many students omitted the step of subtracting the square of E(X) when computing variance directly, i.e. they left Var(X) as Σx²P(X=x) instead of Σx²P(X=x) - [E(X)]².
在处理离散随机变量，尤其是线性组合时，考生经常错误地套用 E(aX+b) 与 Var(aX+b)。评分方案显示的错误包括将 Var(3X) 写成 3Var(X) 而非 9Var(X)，或是忘记加上常数并不改变方差。在给定了表格形式的概率分布时，许多学生在直接计算方差时省略了减去 E(X) 平方这一步，即他们将 Var(X) 保留为 Σx²P(X=x)，而没有减去 [E(X)]²。
) = E(X²) - [E(X)]². Show both intermediate values before the final subtraction so you can spot an error if the variance appears negative.
牢记核心规则：E(aX+b) = aE(X)+b，Var(aX+b) = a²Var(X)。对于离散分布，始终先算出 E(X) 与 E(X²)，然后利用 Var(X) = E(X²) - [E(X)]²。在最终做差之前写出这两个中间值，这样一旦出现方差为负你就能立刻发现错误。
7. Tree Diagram Branch Labels: Missing Probabilities or Incorrect Sum | 树图分支标注：漏标概率或总和错误
Examiners noted that many tree diagrams were incomplete or contained branches whose probabilities at a node did not sum to 1. In a question involving two stages of selection, students sometimes forgot to adjust the probabilities for the second stage after a 'without replacement' event, leaving constants from the first stage. Additionally, when the tree had more than two possible outcomes at a node, some candidates wrote probabilities that summed to more than 1, suggesting they had misinterpreted the given information. The mark scheme awarded method marks only if the tree was correctly structured and labelled.
考官注意到，许多树图不完整，或者同一节点各分支的概率加起来不等于 1。在一道涉及两阶段抽样的题目中，学生有时忘记在“无放回”事件后调整第二阶段的概率，而是原封不动地保留第一阶段的数值。此外，当某个节点存在超过两种可能的结果时，一些考生所写概率之和超过了 1，这表明他们误解了所给信息。评分方案仅对结构正确且标注无误的树图给予方法分。
Before moving to the calculation, check that every set of branches emanating from a single point adds up to 1. Write the sum in the margin if it helps. For without-replacement scenarios, update both the numerator and denominator for the second layer. Where exact fractions become messy, leave them as unsimplified but correct expressions.
在进入计算前，检查从同一个点发出的每一组分支概率是否总和为 1。如有必要，可在页边空白处写出总和。对于无放回情境，要同时更新第二层分支的分子和分母。若分数形式较为繁琐，可以保留未化简但正确的表达式。
8. Misreading Statistical Tables: Degrees of Freedom and Tail Areas | 统计表格的误读：自由度与尾部面积
Questions that supplied critical values from the t-distribution, chi-squared or binomial cumulative tables tripped many candidates. Typical mistakes included using degrees of freedom n instead of n-1 for a sample, looking up one-tailed values when the test was two-tailed, or reading the significance level in the wrong column. The mark scheme revealed that a large proportion of students correctly set up the test statistic but then compared it against an incorrect critical value, leading to a wrong conclusion and loss of accuracy marks.
提供 t 分布、卡方分布或二项累积分布表求临界值的题目绊倒了许多考生。典型错误包括：对样本使用自由度 n 而非 n-1，在双尾检验中查找单尾的临界值，或是在错误的列中读取显著性水平。评分方案显示，很大一部分学生虽正确设定了检验统计量，却与错误的临界值进行比较，从而导致错误结论并失掉答案准确度分。
Practise navigating the exact tables your exam board provides. Annotate your table booklet during revision: highlight row labels for common degrees of freedom and mark which columns correspond to 0.05, 0.025 etc. Always sketch a bell curve and shade the rejection region before picking the table entry.
练习翻阅你考试局提供的准确表格。在复习时为你使用的表册加上标注：高亮常见自由度所在的行，标出哪一列对应 0.05、0.025 等。在选择表值之前，务必先画出钟形曲线并涂出拒绝域。
9. Correlation and Regression: Extrapolation and Causation | 相关与回归：外推与因果关系的误解
In the unit covering bivariate data, a common error was predicting y values for x far outside the observed range without commenting on the dangers of extrapolation. The mark scheme required a statement acknowledging that the prediction would be unreliable. Furthermore, when describing a strong correlation, many candidates wrote statements implying causation, e.g. 'the increase in x causes y to rise', which the mark scheme explicitly rejected unless the context justified it. Only a comment on association was appropriate.
在涉及双变量数据的单元中，一个常见错误是在远超出观测范围的 x 值上预测 y，却没有对外推的危险加以说明。评分方案要求给出声明，承认该预测不可靠。另外，在描述强相关时，许多考生写出暗示因果关系的语句，例如“x 的增加导致 y 上升”，除非题目情境能够支持，否则评分方案会明确否定这类表述。此时只应对关联性进行评论。
If the question asks for an estimate of y for an x-value outside the data range, always add a sentence: 'This is an extrapolation and may be unreliable.' For correlation, use wording like 'there is a positive association’ or ‘as x increases, y tends to increase’, and avoid causal verbs.
如果题目要求针对超出数据范围的 x 值估计 y，务必加上一句话：“此为外推结果，可能不可靠。” 对于相关性，使用诸如“存在正向关联”或“随着 x 增加，y 倾向于增加”这样的措辞，并避免因果性的动词。
10. Sampling Distribution: Confusing Standard Deviation and Standard Error | 抽样分布：标准差与标准误的混淆
In questions on the distribution of the sample mean, a significant number of students used the population standard deviation σ instead of the standard error σ/√n. When calculating probabilities for X̄, the mark scheme showed that correct standardisation required dividing by √n to obtain the appropriate spread. Another frequent slip was computing probabilities for a single observation when the question clearly asked for the mean of a sample: many candidates ignored the sample size entirely, treating X̄ as a single normal variable with variance σ² instead of σ²/n.
在涉及样本均值分布的题目中，数量可观的学生使用了总体标准差 σ，而非标准误 σ/√n。计算关于 X̄ 的概率时，评分方案显示，正确的标准化需要除以 √n 以得到恰当的离散程度。另一个频繁出现的疏漏是：当题目明确要求样本均值的概率时，考生却去计算单个观测值的概率，彻底忽略了样本容量，将 X̄ 当作方差为 σ² 而非 σ²/n 的单个正态变量来处理。
Every time you see X̄ (or a question about 'the mean of n observations'), immediately write X̄ ~ N(μ, σ²/n) and underline the denominator. Then standardise with (x̄ - μ) / (σ/√n). This habit will prevent a large class of mark-losing mistakes.
每当你看到 X̄（或“n 个观测值的均值”字眼），立即写下 X̄ ~ N(μ, σ²/n) 并在分母底下画线。然后用 (x̄ - μ) / (σ/√n) 进行标准化。养成这一习惯可以杜绝一大类失分错误。
Published by TutorHao | 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