AS WJEC Statistics: Exam Technique and Marking Criteria | AS WJEC 统计：答题技巧与评分标准

📚 AS WJEC Statistics: Exam Technique and Marking Criteria | AS WJEC 统计：答题技巧与评分标准

Success in AS WJEC Statistics requires not only solid mathematical ability but also a clear understanding of how marks are awarded. The examiner looks for method, accuracy, correct notation and the ability to interpret results in context. This guide unpacks the marking principles and shares practical exam techniques to help you turn your knowledge into a high score.

在 AS WJEC 统计学考试中取得成功，不仅需要扎实的数学能力，还需要清楚地了解评分方式。考官关注的是方法、准确性、正确的符号以及在情境中解释结果的能力。本指南将拆解评分原则，并分享实用的答题技巧，帮助你把知识转化为高分。

1. Understanding the Marking Scheme | 理解评分标准

WJEC uses a clear marking scheme with M marks for method, A marks for accuracy and B marks for independent facts or definitions. The mark scheme often allows ‘ft’ (follow-through) marks, which means you can still earn accuracy marks if you use an earlier incorrect answer correctly in a later part. Knowing this can reduce exam panic.

WJEC 采用清晰的评分方案：M 分代表方法分，A 分代表准确分，B 分代表独立的事实或定义分。评分方案中经常允许“跟进分”（ft），这意味着如果你在后面的部分正确地使用了前面错误的答案，你仍然可以获得准确分。了解这一点可以减少考试时的恐慌。

In a probability question, writing the formula P(A ∪ B) = P(A) + P(B) – P(A ∩ B) might earn an M1. Correctly substituting numbers earns the next M1, and obtaining the right final value earns an A1. Even if you mis-copied a number, you could still get method marks and possibly ft accuracy marks in subsequent parts.

在概率题中，写出公式 P(A ∪ B) = P(A) + P(B) – P(A ∩ B) 可能获得一个 M1。正确代入数字得到下一个 M1，得出正确的最终数值则获得 A1。即使你抄错了一个数字，你仍然可以在后续部分获得方法分，甚至可能获得跟进准确分。

Always check how many marks a question is worth. A 2-mark question often expects one method step and one final answer. A 5-mark hypothesis test may split marks for hypotheses, test statistic, critical value or p-value, comparison and contextual conclusion.

一定要看清题目的分值。一道 2 分题通常期望一个方法步骤和一个最终答案。一道 5 分假设检验题可能会将分数分配给假设、检验统计量、临界值或 p 值、比较以及情境化的结论。

2. Showing Sufficient Working Clearly | 充分且清晰地展示解题步骤

One of the most common reasons for losing marks is omitting steps. WJEC examiners need to see how you reached your answer to award method marks. Even if a calculation seems simple, write down the formula or key substitution. For instance, when calculating the mean from a frequency table, clearly show ∑fx and ∑f before the final division.

失分最常见的原因之一就是省略步骤。WJEC 考官需要看到你是如何得出答案的，才能给方法分。即使一个计算看起来很简单，也要写下公式或关键的代入步骤。例如，当从频数表计算均值时，要先清晰地展示 ∑fx 和 ∑f，再做最后的除法。

When using a calculator for regression line or standard deviation, do not just give the final equation or a value like r = 0.96. Copy intermediate sums such as Sxx, Syy, Sxy from your calculator screen or show them in a table. This not only secures method marks but also makes checking easier if you suspect an error.

当使用计算器计算回归直线或标准差时，不要只给出最终方程或诸如 r = 0.96 这样的值。从计算器屏幕上复制中间和，如 Sxx、Syy、Sxy，或者在表格中展示它们。这不仅能确保拿到方法分，还能在你怀疑有错误时方便检查。

In hypothesis testing, always specify the distribution of the test statistic under H₀, e.g. X ~ B(20, 0.5). Write down the probability you are calculating, such as P(X ≥ 15). This structured layout matches the mark scheme directly.

在假设检验中，要明确写出检验统计量在 H₀ 下的分布，例如 X ~ B(20, 0.5)。写下你正在计算的概率，如 P(X ≥ 15)。这种结构化的书写直接对应评分方案。

3. Handling Probability Questions | 处理概率问题

Probability questions often involve set notation, tree diagrams or Venn diagrams. Marks are awarded for correct structure even if the arithmetic slips. Draw and label Venn diagrams or tree diagrams clearly, showing probabilities on branches. For conditional probability, state the formula P(A|B) = P(A ∩ B) / P(B) before plugging in numbers.

概率题通常涉及集合符号、树状图或维恩图。即使算术有误，正确的结构也能得分。清晰地画出并标注维恩图或树状图，在树枝上标明概率。对于条件概率，先写下公式 P(A|B) = P(A ∩ B) / P(B)，再代入数字。

When asked to find P(A ∪ B), check if events are mutually exclusive or independent first. State your reasoning: ‘since mutually exclusive, P(A ∪ B) = P(A) + P(B)’. This comment often carries a B mark. If events are independent, use P(A ∩ B) = P(A) × P(B) appropriately.

当被要求求 P(A ∪ B) 时，首先检查事件是否互斥或独立。陈述你的理由：“由于互斥，P(A ∪ B) = P(A) + P(B)”。这个说明通常会获得一个 B 分。如果事件独立，则恰当地使用 P(A ∩ B) = P(A) × P(B)。

For discrete random variables, ensure the sum of probabilities equals 1. If you derive a missing probability by subtracting from 1, show that step clearly. Many marks are lost through careless addition of decimal probabilities – double-check with the total.

对于离散随机变量，要确保概率之和等于 1。如果你通过从 1 中减去得到缺失的概率，要清晰地展示这一步。很多分数是因为小数概率加法粗心而丢掉的——用总和再检查一遍。

4. Using Statistical Tables and Calculator Correctly | 正确使用统计表和计算器

WJEC provides statistical tables for the binomial, Poisson and normal distributions. Learn to read them efficiently. For a binomial probability P(X ≤ k), locate n and p directly. If you need P(X = k), compute P(X ≤ k) – P(X ≤ k–1) and show this subtraction. Even if the final answer is misread, the method can earn marks.

WJEC 为二项分布、泊松分布和正态分布提供了统计表。要学会高效地查阅它们。对于二项概率 P(X ≤ k)，直接根据 n 和 p 查找。如果你需要 P(X = k)，则计算 P(X ≤ k) – P(X ≤ k–1) 并展示这一减法。即使最终答案读错，方法仍能得分。

When using a calculator for normal distribution probabilities, be precise with the order of inputs. For a right-tail probability P(X > a), write it as 1 – Φ((a – μ)/σ). Show the standardised value first, then the probability. This boosts method marks and reduces calculator entry errors.

当使用计算器求正态分布概率时，要精确输入顺序。对于右尾概率 P(X > a)，写成 1 – Φ((a – μ)/σ)。先展示标准化值，再展示概率。这能增加方法分，并减少计算器输入错误。

Do not forget to indicate continuity correction if you are approximating a discrete distribution with a normal. For example, P(X ≥ 12) with a normal approximation becomes P(X > 11.5). State the correction explicitly.

如果你正在用正态分布近似一个离散分布，不要忘记指明连续性校正。例如，用一个正态分布近似 P(X ≥ 12) 会变成 P(X > 11.5)。要明确地写出校正过程。

5. Hypothesis Testing: Structure and Conclusion | 假设检验：结构与结论

A hypothesis test is marked rigidly: hypotheses, test statistic distribution, actual test statistic or p-value, comparison and conclusion in context. Use the standard format every time. State H₀ and H₁ clearly using parameters, e.g. H₀: p = 0.3, H₁: p > 0.3 for a one-tailed test. Never use sample statistics in the hypotheses.

假设检验的评分很严格：假设、检验统计量的分布、实际的检验统计量或 p 值、比较以及情境中的结论。每次都使用标准格式。清晰地陈述 H₀ 和 H₁，使用参数，例如对于单尾检验，H₀: p = 0.3, H₁: p > 0.3。绝不要在假设中使用样本统计量。

After calculating the test statistic or p-value, make a comparison with the significance level. For critical value approach, write ‘test statistic = 7, critical value = 9, so not significant’. For p-value, write ‘p = 0.036 < 0.05, so reject H₀'. The final line must be a non-technical conclusion in the context of the problem: 'There is sufficient evidence to suggest that the proportion has increased.' Avoid merely stating 'accept H₁'.

在计算出检验统计量或 p 值后，与显著性水平进行比较。对于临界值法，写“检验统计量 = 7，临界值 = 9，因此不显著”。对于 p 值法，写“p = 0.036 < 0.05，因此拒绝 H₀”。最后一行必须是用问题背景表述的非技术性结论：“有充分证据表明比例有所增加。”避免仅仅说“接受 H₁”。

One common penalty is omitting the conclusion or writing it technically. Marks are specifically available for a clear contextual sentence. Practise writing these for every past paper question.

一个常见的扣分点是遗漏结论或使用技术性语言写结论。清晰的情境性语句是专门有分可得的。为每一道往年试题练习写这个结论句。

6. Regression and Correlation | 回归与相关

When calculating the equation of the regression line y = a + bx, show how b and a are found using the given summary statistics. Even if you use a calculator, note down Sxy, Sxx and perhaps the formulas b = Sxy / Sxx and a = ȳ – bx̄. This demonstrates method even if a calculator output is slightly mis-typed.

当计算回归直线方程 y = a + bx 时，要展示如何用给定的汇总统计量求出 b 和 a。即使你使用计算器，也要记下 Sxy、Sxx，或许还有公式 b = Sxy / Sxx 和 a = ȳ – bx̄。这样，即使计算器输出略有输入错误，也能展现出方法。

Interpreting the gradient: ‘For every one unit increase in x, y increases by b on average.’ Similarly, interpretation of the correlation coefficient r should mention strength and direction. A value of r = –0.85 indicates ‘strong negative linear correlation’. Do not forget to comment on the context, e.g. ‘as age of car increases, its value tends to decrease’.

解释斜率：“x 每增加一个单位，y 平均增加 b。”类似地，解释相关系数 r 时应提及强度和方向。r = –0.85 的值表示“强负线性相关”。不要忘记结合背景评论，例如“随着车龄增加，汽车的价值趋于下降”。

A reliability warning: interpolation is reliable within the data range, but extrapolation may be unreliable. If a question asks you to predict y for an x far outside the given data, state that the estimate is unreliable because the value lies outside the range of the sample data.

可靠性提醒：在数据范围内内插是可靠的，但外推可能不可靠。如果题目要求你对一个远超出给定数据范围的 x 预测 y，要说明这个估计不可靠，因为该值位于样本数据范围之外。

7. Discrete Random Variables and Expected Value | 离散随机变量与期望值

For a discrete random variable X, the probability distribution table must show all possible values and their probabilities. Marks are often reserved for verifying ΣP(X = x) = 1. Write down E(X) = Σ x·P(X = x) and Var(X) = E(X²) – [E(X)]². Show E(X²) = Σ x²·P(X = x) in a separate column.

对于离散随机变量 X，概率分布表必须显示所有可能取值及其概率。通常会为验证 ΣP(X = x) = 1 保留分数。写下 E(X) = Σ x·P(X = x) 和 Var(X) = E(X²) – [E(X)]²。在单独的一列中展示 E(X²) = Σ x²·P(X = x)。

If the question asks for Var(3X – 2), use properties correctly: Var(3X – 2) = 3² Var(X) = 9 Var(X). Show that addition/subtraction does not affect variance. Misapplying this property is a common slip which loses easy marks.

如果题目要求求 Var(3X – 2)，要正确运用性质：Var(3X – 2) = 3² Var(X) = 9 Var(X)。要展示加减常数不影响方差。误用这一性质是常见的失误，会丢掉容易得到的分数。

In problems about games or profit, the expected gain E(X) might be asked. Always define your random variable clearly first, e.g. ‘Let X be the net gain in pounds’. Then construct the distribution table and calculate E(X). A sentence concluding whether the game is fair or the expected profit is positive earns the final mark.

在关于游戏或利润的问题中，可能会要求期望收益 E(X)。始终先清晰地定义你的随机变量，例如“设 X 为净收益（单位为镑）”。然后构建分布表并计算 E(X)。用一句话总结游戏是否公平或期望利润为正，将获得最后一分。

8. Data Presentation and Graph Interpretation | 数据表示与图表解读

Box plots, histograms, stem-and-leaf diagrams and cumulative frequency curves all test specific plotting skills. For a histogram, use frequency density = frequency / class width. Label axes correctly and draw bars without gaps. For a box plot, show outliers as separate crosses if the rule Q1 – 1.5×IQR or Q3 + 1.5×IQR is applied.

箱线图、直方图、茎叶图和累积频率曲线都考查特定的绘图技能。对于直方图，要使用频数密度 = 频数 / 组距。正确标注坐标轴，并画出没有间隙的条形。对于箱线图，如果应用了 Q1 – 1.5×IQR 或 Q3 + 1.5×IQR 的规则，将异常值显示为单独的叉号。

When reading a cumulative frequency curve, draw lines on the graph to indicate how you find the median and quartiles. Even if the graph is provided on the question paper, do mark it clearly. These construction lines are often evidence for method marks.

当读取累积频率曲线时，要在图上画出指示你如何找到中位数和四分位数的线条。即使题目纸上已提供图表，也要清晰地标注出来。这些构建线往往是方法分的证据。

For comparison of data sets, always give both a measure of central tendency and a measure of spread, e.g. median and interquartile range. Make a meaningful comparative statement: ‘The median time is higher for Group A, and the IQR is smaller, suggesting less variation.’ Such contextual comments are essential for interpretation marks.

在比较数据集时，既要给出集中趋势的度量，也要给出离散程度的度量，例如中位数和四分位距。做出一个有意义的比较性陈述：“A组的时间中位数更高，而IQR更小，表明变异较小。”这种结合情境的评论对于获得解释分至关重要。

9. Common Errors and Misconceptions | 常见错误与误区

Many students confuse Pearson’s product-moment correlation coefficient with the equation of the regression line, trying to use r as the gradient. Remember, b = r × (Sy / Sx). The regression line always passes through the mean point (x̄, ȳ). Check this as a verification.

许多学生将皮尔逊积矩相关系数与回归直线方程混淆，试图将 r 用作斜率。请记住，b = r × (Sy / Sx)。回归直线总是通过均值点 (x̄, ȳ)。用这一点来验证你的结果。

In the normal distribution, a typical error is using P(Z > z) as the table value directly when the table gives Φ(z) = P(Z < z). Always sketch a small diagram showing the area you need. For a negative z, use symmetry appropriately, e.g. P(Z < –1.2) = 1 – Φ(1.2). Write this step down.

在正态分布中，一个典型的错误是在表格给出 Φ(z) = P(Z < z) 时，直接将 P(Z > z) 用作表格值。总是画一个小示意图显示你需要的面积。对于负 z 值，适当地利用对称性，例如 P(Z < –1.2) = 1 – Φ(1.2)。写下这个步骤。

For binomial and Poisson approximations to the normal, missing the continuity correction is a recurring mistake. Remember that a discrete integer boundary is adjusted by ±0.5. Also, check the conditions: np and nq > 5 for binomial, and λ > about 10 for Poisson approximation to normal.

对于二项分布和泊松分布的正态近似，遗漏连续性校正是一个反复出现的错误。记住，离散整数边界要通过 ±0.5 进行调整。还要检查条件：二项正态近似要求 np 和 nq > 5，泊松正态近似要求 λ 约大于 10。

10. Time Management and Final Accuracy | 时间管理与终稿准确性

AS Statistics papers are often time-pressured. Skim the paper and start with the questions you feel most confident about. Before moving to the next question, double-check the numerical answers, especially probabilities that must be between 0 and 1, and any negative variances which are impossible.

AS 统计试卷通常时间紧迫。快速浏览试卷，从你最自信的题目开始做。在进入下一题之前，再次检查数值答案，尤其是必须在 0 到 1 之间的概率，以及任何不可能为负的方差。

State all decimal probabilities to 3 or 4 decimal places unless the question specifies otherwise. But in critical value calculations, avoid premature rounding. Use exact binomial or Poisson probabilities from tables, and for normal distribution, use unrounded z-values to find probabilities.

除非题目另有规定，所有小数概率保留 3 或 4 位小数。但在临界值计算中，避免过早四舍五入。使用统计表中精确的二项或泊松概率，对于正态分布，使用未四舍五入的 z 值来查找概率。

If you have time, re-read the question prompt. Often marks are lost because a question asks for ‘an interpretation of the intercept’ but the student talks only about the gradient. Underline command words such as ‘state’, ‘calculate’, ‘interpret’ and ‘comment on reliability’.

如果有时间，重新阅读题目提示。常常有分数是因为题目要求“解释截距”，而学生只讨论了斜率而丢掉的。在诸如“陈述”、“计算”、“解释”和“评论可靠性”这些指令词下划线。

A final check: do your conclusions refer back to the original context? In a hypothesis test, have you mentioned the specific scenario (e.g. ‘the new drug is more effective’)? In a regression prediction, have you cautioned about extrapolation if x is outside the range? These small touches consistently pick up the final mark on a question.

最后检查：你的结论是否回扣了原始情境？在假设检验中，你是否提到了具体的场景（例如“新药更有效”）？在回归预测中，如果 x 超出了范围，你是否警示了外推问题？这些细节能持续地拿到题目中的最后一分。

Published by TutorHao | Statistics Revision Series | aleveler.com

更多咨询请联系16621398022（同微信）

AS WJEC Statistics: Exam Technique and Marking Criteria | AS WJEC 统计：答题技巧与评分标准