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

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

Mastering AS Statistics goes beyond knowing the formulas — it requires a clear understanding of how examiners award marks and how to present your work to maximise those marks. This guide blends practical exam technique with the specific marking principles used by OCR, helping you gain every possible mark on your paper.

掌握 AS 统计不仅需要熟记公式，更需要透彻理解评分员如何给分、如何展示解题过程以获得最多分数。本指南融合了实用的答题技巧与 OCR 特有的评分原理，帮助你在试卷上拿下每一分。

1. Understanding OCR Mark Schemes | 理解 OCR 评分方案

Every mark on your paper belongs to a category. Recognising these categories can change the way you structure your answers.

试卷上的每一分都属于某个类别。识别这些类别能改变你组织答案的方式。

Mark type	Symbol	What it rewards
Method	M1	A correct and clearly shown method
Accuracy	A1	The correct numerical answer following a method
Independent	B1	A statement or value not dependent on method
Explanation	E1	A clear verbal interpretation or justification

Method marks (M1) are awarded even if the final answer is incorrect, provided the examiner can see a valid route. A1 usually depends on getting the right outcome from that method. B1 marks often test facts like stating hypotheses or giving a definition. E1 marks reward interpreting a result in context, such as the conclusion of a hypothesis test.

方法分（M1）即使最终答案错误，只要评分员能看到有效的解题路径，就能得到。A1 通常依赖该方法下得到的正确结果。B1 分常测试事实，如陈述假设或给出定义。E1 分则奖励在上下文中解释结果，例如假设检验的结论。

Always write enough working so that a method can be identified. If you give only a final answer and it is wrong, you risk losing both M and A marks.

一定要写出足够的步骤，以便评分员能看到用的方法。如果你只写出最终答案，并且错了，你可能会同时失去方法分和答案分。

---

2. Show Clear Working, Not Just Answers | 展示清晰的求解步骤，而不只是答案

The single most effective technique to secure method marks is to write down the formula you intend to use, substitute the numbers carefully, and then evaluate.

确保方法分最有效的技巧是：写下你要使用的公式，仔细代入数字，然后再计算。

Example: x̄ = Σx / n or s = √[ Σ(x − x̄)² / (n−1) ]

Do not skip steps such as squaring differences or summing frequencies. A clear layout makes it easy for the examiner to follow your thinking, and if you make a slip, they can still award the M mark.

不要跳过计算差值平方或频率求和等步骤。清晰的布局能让评分员轻松跟上你的思路，即使你出现失误，他们仍然可以给方法分。

When you use a calculator to obtain a summary statistic or a probability, still note the expression you are evaluating. For instance, write ‘P(X ≤ 3) = 0.1719 (from Bcd)’ rather than just the number.

当你用计算器获得汇总统计量或概率时，仍然要写下你正在计算的表达式。例如，写“P(X ≤ 3) = 0.1719（来自 Bcd）”，而不只是数字。

---

3. Data Presentation and Summary Statistics | 数据呈现与汇总统计量

OCR questions often provide data in frequency tables or as coded values. You must handle coding correctly and present your answers with proper units.

OCR 考题常以频数表或编码值的形式提供数据。你必须正确处理编码，并以恰当的单位呈现答案。

For coded data y = (x − a) / b, the reverse transformations are x̄ = b·ȳ + a and sₓ = |b|·sᵧ. A common mistake is to forget that the standard deviation is multiplied by the absolute value of b, while the mean is linearly transformed.

对于编码数据 y = (x − a) / b，反向变换为 x̄ = b·ȳ + a 以及 sₓ = |b|·sᵧ。一个常见错误是忘记标准差要乘以 b 的绝对值，而均值是线性变换。

When drawing diagrams such as box plots or histograms, label axes clearly, indicate the scale, and mark key values like quartiles or boundaries. Diagram accuracy can earn independent B marks.

绘制箱线图或直方图等图形时，要清楚标示坐标轴、标注尺度，并标记四分位数或组界等关键值。图表的准确性可以赢得独立的 B 分。

In questions that ask for comparison of data sets, always refer to a measure of central tendency and a measure of spread, such as ‘The median is higher, so …’ and ‘The interquartile range is smaller, indicating less variation.’

在要求比较数据集的题目中，一定要同时提及集中趋势量数和离散程度量数，如“中位数更高，所以…”以及“四分位距更小，表明变异较小”。

---

4. Probability and the Binomial Distribution | 概率与二项分布

Binomial situations must be checked against the conditions: a fixed number of trials n, each trial independent, only two outcomes, and a constant probability of success p.

二项分布的情形必须验证其条件：固定的试验次数 n，每次试验独立，只有两种结果，且成功率 p 恒定不变。

When working with X ~ B(n, p), use the correct calculator function: Bpd for P(X = r) and Bcd for cumulative probabilities such as P(X ≤ r). Write the expression first.

在处理 X ~ B(n, p) 时，使用正确的计算器功能：Bpd 用于 P(X = r)，Bcd 用于累积概率如 P(X ≤ r)。要先写出表达式。

P(X = r) = ⁿCᵣ · p^r · (1 − p)^(n − r)

For inequalities, remember that P(X > r) = 1 − P(X ≤ r) and P(X ≥ r) = 1 − P(X ≤ r−1). A sketch of the required region on a small bar chart can prevent careless errors.

对于不等式，记住 P(X > r) = 1 − P(X ≤ r) 且 P(X ≥ r) = 1 − P(X ≤ r−1)。在简单条形图上画出所需区域可以避免粗心错误。

---

5. Hypothesis Testing Framework | 假设检验框架

OCR expects a structured approach to every hypothesis test. Missing a step can cost E1 or B1 marks even if the calculations are correct.

OCR 期望每道假设检验题都有条理清晰的解答。遗漏某个步骤就可能失去 E1 或 B1 分，即便计算正确。

Always write: ‘Let p be the probability that …’ Then state the null hypothesis H₀: p = … and the alternative H₁: p < ..., p > … or p ≠ … as appropriate. Define the significance level, usually given in the question.

务必写出：“令 p 为…的概率”。然后陈述原假设 H₀: p = … 和备择假设 H₁: p < ...、p > … 或 p ≠ …。定义显著性水平，通常题干会给出。

Specify the distribution of the test statistic. For a binomial test, write ‘Under H₀, X ~ B(n, p).’ Then calculate the p-value or find the critical region. Use clear notation like P(X ≤ observed value).

明确检验统计量的分布。对于二项检验，写“在 H₀ 下，X ~ B(n, p)”。然后计算 p 值或找出临界域。使用清楚的记号，如 P(X ≤ 观测值)。

End with a conclusion in context: ‘Since the p-value (0.023) < 0.05, there is sufficient evidence to reject H₀. The data suggests that the proportion has increased.' Never just say 'accept H₁' without interpreting.

最后结合背景给出结论：“因为 p 值 (0.023) < 0.05，有充分证据拒绝 H₀。数据表明比例有所增加。”不要只是说“接受 H₁”而不作解释。

---

6. Normal Distribution Calculations | 正态分布计算

For X ~ N(μ, σ²), probabilities are found by standardising or using the calculator’s normal Cdf. Write the standardised value z, even if your calculator does it internally.

对于 X ~ N(μ, σ²)，通过标准化或使用计算器的正态 Cdf 来求概率。即使你的计算器内置了标准化过程，也要写出标准值 z。

z = (X − μ) / σ

When an unknown mean or standard deviation is required, set up an equation using the inverse normal. For example, if P(X < k) = 0.95, find the z-value such that Φ(z) = 0.95 and then solve (k − μ)/σ = z.

当需要求未知的均值或标准差时，利用逆正态建立方程。例如，若 P(X < k) = 0.95，找到满足 Φ(z) = 0.95 的 z 值，然后解 (k − μ)/σ = z。

Always draw a labelled normal curve, shading the region of interest. This not only helps you avoid errors but also provides evidence of your method, potentially earning an M mark if the final answer is lost.

始终画出带标注的正态曲线，并给关注区域涂上阴影。这不仅有助于避免错误，还能为你的方法提供证据，万一最后答案丢失，仍有可能获得方法分。

---

7. Correlation and Regression Essentials | 相关与回归要点

When calculating Pearson’s product moment correlation coefficient r, record the value from your calculator to at least three significant figures. Interpret r in the context of the data: a value close to +1 indicates strong positive linear correlation.

计算皮尔逊积矩相关系数 r 时，记下计算器给出的值，至少保留三位有效数字。结合数据背景解释 r：接近 +1 的值表示强正线性相关。

Ensure you identify the explanatory variable (x) and response variable (y) correctly. The regression line is y = a + bx, where b describes the estimated change in y for a one-unit increase in x.

确保正确识别解释变量（x）和响应变量（y）。回归直线为 y = a + bx，其中 b 描述的是 x 每增加一个单位时 y 的估计变化量。

Use the regression line only for prediction within the observed range of x. Predicting far outside this range is unreliable, and OCR may ask you to comment on the danger of extrapolation.

回归直线只能用于在观测到的 x 范围内进行预测。太远超出该范围的预测并不可靠，OCR 可能会要求你评论外推的危害。

When discussing reliability, mention the correlation coefficient and the sample size. When discussing validity, refer to whether the data really measures what it claims and whether the model is appropriate.

讨论可靠性时，要提及相关系数和样本大小。讨论效度时，要说明数据是否真正测量了它所声称的内容，以及模型是否适用。

---

8. Common Pitfalls and How to Avoid Them | 常见陷阱及避免方法

Pitfall 1: Confusing one-tailed and two-tailed tests. Always check the wording of the alternative hypothesis — ‘increased’ or ‘higher’ suggests one-tailed; ‘changed’ or ‘different’ implies two-tailed.

陷阱一：混淆单尾与双尾检验。务必检查备择假设的措辞——“增加了”或“更高”暗示单尾；“改变了”或“不同”暗示双尾。

Pitfall 2: Losing accuracy marks through premature rounding. Keep intermediate results to at least four significant figures; only round the final answer to the required degree of accuracy, typically three significant figures.

陷阱二：因过早舍入而失去答案分。中间结果至少保留四位有效数字；只有最终答案才按要求舍入到指定的精度，通常是三位有效数字。

Pitfall 3: Not stating assumptions. In binomial and normal problems, you may be asked to comment on the assumption of independence, constant probability, or normality. Missing this can cost B or E marks.

陷阱三：未说明假设。在二项分布和正态分布问题中，可能会要求你评论独立性、概率恒定或正态性等假设。遗漏这些可能会丢失 B 或 E 分。

Pitfall 4: Mixing up p-value and significance level. The p-value is calculated from your test statistic; compare it with the stated significance level to make your decision, and state that comparison explicitly.

陷阱四：混淆 p 值与显著性水平。p 值是根据检验统计量计算得出的；将其与给定的显著性水平比较以作出判断，并明确写出这一比较。

---

9. Using Your Calculator Effectively | 有效使用计算器

Before the exam, ensure you are fully familiar with the statistical functions on your specific calculator model. Know how to enter data lists, compute two-variable statistics, and access binomial and normal distributions.

考前要确保你完全熟悉自己计算器型号的统计功能。知道如何输入数据列表、计算双变量统计量以及使用二项分布和正态分布功能。

Clear any old data before starting a new calculation. A residual list can distort your summary statistics and cost you all accuracy marks.

开始新的计算前，清除所有旧数据。残留的列表会扭曲汇总统计量，让你失去全部答案分。

Use the calculator’s memory or store functions to hold intermediate values such as a mean or standard deviation. This reduces rounding errors when these values are used in later parts of the question.

使用计算器的存储功能保存均值或标准差等中间值。当这些值用于题目的后续部分时，这能减少舍入误差。

---

10. Final Checks Before Submission | 提交前的最终检查

Re-read every sub-question to confirm you have answered exactly what was asked. If the question asks for a probability ‘greater than 5’, make sure you used the correct inequality and included or excluded the boundary correctly.

重新阅读每一个子问题，确认你准确回答了所问。如果题目要求“大于 5”的概率，确保你使用了正确的不等式，并正确包含或排除边界。

Verify that all probabilities and p-values are reported in the correct format, usually to three significant figures, and that your conclusion is written in plain English within the context of the problem.

核实所有概率和 p 值都已按正确格式报告，通常是三位有效数字，并且结论是用通俗英语在题目背景下写出的。

A quick check for impossible probabilities — anything negative or greater than 1 — can immediately catch a mistake. Similarly, standard deviations must be non-negative.

快速检查不可能的概率——任何负值或大于 1 的值——能立即捕捉到错误。同样，标准差必须是非负的。

Published by TutorHao | Statistics Revision Series | aleveler.com

更多咨询请联系16621398022（同微信）