Year 13 OCR Statistics: High-frequency Exam Topics & Common Mistake Analysis | OCR Year 13 统计高频考点与易错题分析

📚 Year 13 OCR Statistics: High-frequency Exam Topics & Common Mistake Analysis | OCR Year 13 统计高频考点与易错题分析

This article identifies the most frequently examined topics in Year 13 OCR Statistics (both the A Level Mathematics specification H240 and the Further Statistics components) and analyses the typical errors that cost students marks. By highlighting common misconceptions in conditional probability, Poisson approximations, hypothesis testing, correlation and chi-squared tests, we aim to sharpen your exam technique and conceptual clarity.

本文梳理了 Year 13 OCR 统计（涵盖 A Level Mathematics H240 及 Further Statistics 模块）中最高频的考点，并深入解析历年导致学生失分的典型错误。我们将突出条件概率、泊松近似、假设检验、相关性及卡方检验等主题中的常见误区，帮助你提升解题技巧与概念清晰度。

1. Conditional Probability and Bayes’ Theorem | 条件概率与贝叶斯定理

Conditional probability and Bayes’ theorem appear regularly in OCR papers, often disguised within real-world contexts such as medical testing or quality control. The key formula is P(A|B) = P(A ∩ B)/P(B). A high-frequency error is confusing the ‘given’ direction – students often treat P(A|B) as P(B|A) without adjusting the sample space. Another common slip is failing to compute the total probability P(B) correctly when using the law of total probability.

条件概率和贝叶斯定理在 OCR 试题中频繁出现，往往隐藏在医学检测或质量控制等现实情境中。核心公式是 P(A|B) = P(A ∩ B)/P(B)。高频错误是混淆“给定”方向——学生经常不加调整地将 P(A|B) 当作 P(B|A) 处理。另一个常见疏忽是在使用全概率公式时未能正确计算 P(B)。

P(A|B) = P(A ∩ B) / P(B) and P(B) = P(B|A)P(A) + P(B|A’)P(A’)

P(A|B) = P(A ∩ B) / P(B) 且 P(B) = P(B|A)P(A) + P(B|A’)P(A’)

Common mistake: Setting P(Defective|Test positive) = P(Test positive|Defective). Without considering the base rate, this leads to wildly incorrect posterior probabilities. Always draw a tree diagram and carefully label the branches with probabilities given in the question.

常见错误：将 P(缺陷|检测阳性) 直接等同于 P(检测阳性|缺陷)。不考虑基准率会导致后验概率完全错误。务必画出概率树形图，并仔细根据题目给定概率标注分支。

Exam questions often ask for ‘the probability that a randomly selected item is faulty given a positive test’. Many candidates stop after computing the joint probability and forget the division by the total probability of testing positive. Remember: the denominator must include all paths that lead to the given event.

试题常要求计算“已知检测阳性，随机抽选物品有故障的概率”。许多考生在算出联合概率后就停下来，忘了除以检测阳性的总概率。记住：分母必须包含所有导致给定事件的路径。

2. Poisson Distribution and Its Approximations | 泊松分布及其近似

The Poisson distribution models the number of events occurring in a fixed interval, provided events happen independently at a constant average rate λ. For OCR, you must be able to use the Poisson formula P(X = r) = e^{-\lambda} λ^{r}/r! and know when a binomial can be approximated by a Poisson (large n, small p). A persistent error is using the Poisson approximation when np is not small – the rule of thumb is n > 50 and np < 5.

泊松分布用于模拟固定区间内发生的事件数量，前提是事件以恒定平均率 λ 独立发生。在 OCR 考试中，你必须会用泊松公式 P(X = r) = e^{-λ} λ^{r}/r!，并了解何时可用泊松近似二项分布（n 大，p 小）。一个顽固错误是当 np 不小的情况下仍使用泊松近似——经验法则是 n > 50 且 np < 5。

P(X = r) = e^{-λ} λ^{r} / r! for r = 0, 1, 2, …

P(X = r) = e^{-λ} λ^{r} / r! 对于 r = 0, 1, 2, …

Common mistake: Misinterpreting ‘rate per unit’. If a question gives a rate of 4 calls per hour but asks about a 15-minute interval, you must scale λ to 1 call. Many students forget to adjust λ and use λ = 4, leading to incorrect probabilities for the shorter interval.

常见错误：误读“每单位发生率”。如果题目给出每小时 4 次呼叫的速率，却要求 15 分钟间隔内的概率，你必须将 λ 按比例调整为 1。很多学生忘记调整 λ，直接用 λ = 4，导致短时段的概率计算错误。

Another trap involves adding independent Poisson variables. The sum of two independent Poisson variables X ~ Po(λ₁) and Y ~ Po(λ₂) is Po(λ₁ + λ₂). Students occasionally add the Poisson means incorrectly when modelling the total number of defects from two independent machines.

另一个陷阱涉及独立泊松变量的求和。两个独立泊松变量 X ~ Po(λ₁) 和 Y ~ Po(λ₂) 的和为 Po(λ₁ + λ₂)。学生在建立两台独立机器总缺陷模型时，有时会错误地对泊松均值求和。

3. Normal Distribution, Continuity Corrections and Approximations | 正态分布、连续性修正与近似

Normal approximations to binomial and Poisson distributions are core OCR A2 topics. The continuity correction is essential when approximating a discrete distribution with a continuous normal distribution. The most frequent mistake is omitting the continuity correction entirely, or applying it incorrectly by adding an extra ±0.5 beyond the correct adjustment. For X ~ B(n, p), the normal approximation uses X ~ N(np, np(1-p)) and P(X ≤ r) becomes P(X < r + 0.5).

正态分布对二项和泊松的近似是 OCR A2 的核心主题。当用连续正态分布近似离散分布时，连续性修正是必不可少的。最常见的错误是彻底忽略连续性修正，或者错误地应用，在正确调整之外又多加了 ±0.5。对于 X ~ B(n, p)，正态近似使用 X ~ N(np, np(1-p))，而 P(X ≤ r) 转化为 P(X < r + 0.5)。

If Y ~ Po(λ), approximate with Y ~ N(λ, λ) and use P(Y ≤ k) ≈ P(Y < k + 0.5)

若 Y ~ Po(λ)，则近似为 Y ~ N(λ, λ) 并使用 P(Y ≤ k) ≈ P(Y < k + 0.5)

Common mistake: Using the normal approximation without checking the validity conditions: np > 5 and n(1-p) > 5 for binomial, or λ > 10 for Poisson. OCR examiners frequently report that candidates apply the approximation when the conditions are not met, leading to marks being lost even if the rest of the working is correct.

常见错误：未检查有效性条件就使用正态近似：二项分布需 np > 5 且 n(1-p) > 5，泊松需 λ > 10。OCR 考官经常反馈，考生在不满足条件时硬用近似，导致即使后续计算正确也被扣分。

Another subtle error occurs when the question asks for a two-tail critical region but the student only calculates a one-tail boundary. Always read the wording carefully: ‘test whether the proportion has changed’ implies a two-tailed test and you must split the significance level symmetrically.

另一个细微错误是题目要求双尾拒绝域，学生却只计算了单尾边界。务必仔细审题：“检验比例是否改变”意味着双尾检验，必须将显著性水平对称分配。

4. Hypothesis Testing for Binomial and Poisson Models | 二项与泊松模型的假设检验

Hypothesis tests on a binomial proportion or a Poisson mean are standard OCR questions. You need to state null and alternative hypotheses precisely, calculate the probability of the observed test statistic, and compare with the significance level. A high-frequency pitfall is incorrectly defining the p-value – many candidates think the p-value is the probability that H₀ is true, rather than the probability of the observed result (or more extreme) given H₀.

对二项比例或泊松均值的假设检验是 OCR 标准题型。你需要准确陈述原假设和备择假设，计算观测检验统计量的概率，并与显著性水平比较。一个高频陷阱是错误定义 p 值——许多考生认为 p 值是 H₀ 为真的概率，而不是在 H₀ 下得到观测结果（或更极端结果）的概率。

Common mistake: Writing hypotheses as ‘H₀: p = 0.3, H₁: p ≥ 0.3’ for a one-tailed test when the context suggests an increase. This puts the rejection region on the wrong side. Always ensure H₁ reflects the claim you are testing. A typical error is using ‘≥’ instead of ‘>’ in H₁, which may not directly match the critical region setup.

常见错误：在单尾检验中将假设写成“H₀: p = 0.3, H₁: p ≥ 0.3”，但上下文暗示的是增加。这导致拒绝域方向错误。务必确保 H₁ 反映你所检验的主张。典型错误是在 H₁ 中用“≥”而不是“>”，这可能与临界区设定不直接匹配。

For a binomial test, calculating the exact tail probability requires care. A common slip is summing probabilities for X < observed value incorrectly, or failing to include the observed value in the tail when considering 'as or more extreme'. Always check whether the test is one-tailed or two-tailed; for two-tailed, double the smallest tail probability if the distribution is symmetric (but be cautious with asymmetric binomials).

对于二项检验，精确计算尾部概率需要细致。常见疏忽是错误地对 X < 观测值求和，或者在考虑“至少同样极端”时漏掉了观测值本身。务必确认检验是单尾还是双尾；对于双尾，如果分布对称则将较小尾概率加倍（但对于不对称的二项分布要格外小心）。

5. Type I and II Errors, Power and Sample Size | 第一类错误、第二类错误与检验功效

Understanding Type I error (rejecting H₀ when true) and Type II error (failing to reject H₀ when false) is tested through both calculation and interpretation. OCR questions frequently ask for the probability of making a Type II error given a specific alternative, as well as the power of the test. The most common error is confusing the two error types or calculating the wrong probability because the alternative distribution is not correctly identified.

理解第一类错误（拒真）和第二类错误（取伪）既考计算也考解读。OCR 试题经常要求给定特定备择值时计算犯第二类错误的概率，以及检验的功效。最常见的错误是混淆两类错误，或因未能正确识别备择分布而计算出错误的概率。

Common mistake: Using the significance level (α) as the Type II error probability. Type II error probability (β) depends on the actual alternative value. Candidates must set up the distribution of the test statistic under the specific H₁, not under H₀, when finding β. For a binomial test, this means recalculating with a new p.

常见错误：把显著性水平 (α) 当成第二类错误概率。第二类错误概率 (β) 取决于实际的备择值。在求 β 时，考生必须在特定 H₁ 下而非 H₀ 下建立检验统计量的分布。对于二项检验，这意味着用新的 p 重新计算。

Power analysis (1 – β) often appears linking sample size and effect size. A frequent oversight is not stating the power in context – examiners expect a concluding sentence such as ‘There is a 73% chance of correctly detecting an increase to 0.4 if such an increase exists’.

功效分析 (1 – β) 经常与样本量和效应量关联出现。常见的疏忽是没有结合背景陈述功效——考官期待一句结论，如“如果确实增加到 0.4，则该检验有 73% 的机会正确检测出这一增加”。

6. Product Moment Correlation Coefficient and Regression Analysis | 积矩相关系数与回归分析

OCR frequently asks for the calculation of Pearson’s product moment correlation coefficient (r) and a least squares regression line (y = a + bx). The formulas are provided, but misinterpretation of the relationship between correlation and causation is a classic trap. A correlation, even a strong one, does not imply causation. Many students write ‘as X increases, Y increases’ as if X causes Y, when the question only asks for a description of the linear association.

OCR 常考皮尔逊积矩相关系数 (r) 以及最小二乘回归线 (y = a + bx) 的计算。公式已给出，但对相关与因果关系的曲解是经典陷阱。相关性，即便很强，也不意味着因果关系。许多学生写“随着 X 增大，Y 增大”，仿佛 X 导致 Y，而题目只要求描述线性关联。

Common mistake: Using the regression line to predict for a value of X well outside the observed range (extrapolation). OCR examiners often set a follow-up part asking whether a prediction is reliable. Candidates who fail to mention that the prediction is unreliable due to extrapolation lose easy marks.

常见错误：使用回归线对远超出观测范围的 X 值进行预测（外推）。OCR 考官经常设置后续小题询问该预测是否可靠。考生若未提及因外推而不可靠，就会丢掉简单分数。

Another pitfall is misinterpreting the residual. A residual is observed y minus predicted y. Students sometimes subtract in the wrong order, which can invert the sign of the residual and affect their judgement about whether the model overestimates or underestimates.

另一个陷阱是曲解残差。残差是观测值 y 减去预测值。有时学生搞错减法次序，导致残差符号反转，并影响他们对模型高估还是低估的判断。

7. Chi-Squared Tests for Independence and Goodness of Fit | 独立性卡方检验与拟合优度检验

Chi-squared (χ²) tests are examined both for contingency tables (test of independence) and for testing whether observed frequencies fit an expected distribution (goodness of fit). The formula χ² = Σ (O – E)² / E must be used with the correct degrees of freedom. The most persistent error is misidentifying degrees of freedom: for a contingency table with r rows and c columns, df = (r-1)(c-1); for goodness of fit, df = (number of categories – 1 – number of estimated parameters).

卡方 (χ²) 检验既考查列联表（独立性检验），也考查观测频数是否拟合预期分布（拟合优度）。必须正确使用公式 χ² = Σ (O – E)² / E 并确定合适的自由度。最顽固的错误是错误识别自由度：对 r 行 c 列的列联表，df = (r-1)(c-1)；对拟合优度，df = 类别数 – 1 – 参数估计个数。

Common mistake: Combining categories to achieve expected frequencies above 5, but then forgetting to reduce the degrees of freedom accordingly. If you pool two categories, the number of categories decreases, which changes df. Always recalculate df after any pooling.

常见错误：合并类别以使预期频数达到 5 以上，却忘记相应减少自由度。若合并了两个类别，类别总数减少，从而改变 df。任何合并后务必重新计算 df。

Another typical slip involves the decision rule. Some candidates state ‘reject H₀ if χ² > critical value’ but then compare their test statistic with a p-value incorrectly, or use a two-tailed critical value for a test that is inherently one-tailed (χ² tests are one-tailed on the upper side).

另一个典型疏漏涉及决策规则。有些考生陈述“若 χ² > 临界值则拒绝 H₀”，但在比较时却错误地使用 p 值，或者对这种本质上为上侧单尾的检验使用了双尾临界值（χ² 检验是上侧单尾的）。

8. Continuous Random Variables and Probability Density Functions | 连续随机变量与概率密度函数

In Further Statistics, working with probability density functions (pdf) f(x) and cumulative distribution functions (cdf) F(x) is a major topic. Key properties: f(x) ≥ 0, total area under curve = 1, and F(x) = ∫ f(t) dt from lower bound to x. OCR errors often stem from confusion between the pdf and cdf – students might treat F(x) as the pdf when calculating probabilities or finding the median. Remember that the median m satisfies F(m) = 0.5.

在 Further Statistics 中，概率密度函数 (pdf) f(x) 与累积分布函数 (cdf) F(x) 是重要主题。关键性质：f(x) ≥ 0，曲线下总面积为 1，且 F(x) = ∫ 下界到 x 的 f(t) dt。OCR 常见错误来自混淆 pdf 与 cdf——学生在计算概率或求中位数时可能把 F(x) 当作 pdf 使用。请记住中位数 m 满足 F(m) = 0.5。

Common mistake: When calculating P(a < X < b), simply plugging a and b into f(x). This gives a meaningless value; you must integrate f(x) or use F(b) - F(a). Another error is forgetting to find the constant of integration when deriving the cdf from the pdf – always use the condition that F(lower bound) = 0.

常见错误：计算 P(a < X < b) 时，直接将 a 和 b 代入 f(x)。这样得到的值无意义；必须对 f(x) 积分或使用 F(b) - F(a)。另一个错误是在从 pdf 推导 cdf 时忘记确定积分常数——务必利用条件 F(下界) = 0。

Questions on the expectation and variance of continuous variables, E(X) = ∫ x f(x) dx and Var(X) = E(X²) – [E(X)]², require careful integration. A frequent slip is losing a negative sign or mishandling piecewise pdf limits. Always clearly sketch the domain and check that the integrated probability equals 1.

涉及连续变量期望和方差，E(X) = ∫ x f(x) dx 及 Var(X) = E(X²) – [E(X)]²，的题目需要仔细积分。常见失误是丢失负号或错误处理分段 pdf 的界限。务必清晰画出定义域并检查积分总概率是否等于 1。

9. Sampling Distributions and the Central Limit Theorem | 抽样分布与中心极限定理

The concept of the sampling distribution of the sample mean is crucial. For a sample of size n from a normal population, X̄ ~ N(μ, σ²/n). When the population is not normal, the Central Limit Theorem (CLT) states that X̄ is approximately normal for large n (usually n ≥ 30). OCR questions often test whether the CLT can be applied. A common error is applying the CLT without checking sample size, or applying it to a single observation rather than a sample mean.

样本均值的抽样分布概念至关重要。对于来自正态总体的容量为 n 的样本，X̄ ~ N(μ, σ²/n)。当总体非正态时，中心极限定理 (CLT) 指出对大样本（通常 n ≥ 30），X̄ 近似正态。OCR 试题常考查 CLT 能否应用。常见错误是没有检查样本量就应用 CLT，或将其应用于单个观测值而非样本均值。

Common mistake: Confusing standard deviation with standard error. When constructing a confidence interval or conducting a hypothesis test for a mean, candidates often use σ instead of σ/√n. This drastically alters the interval width and may lead to incorrect conclusions. Always recall that the standard error of the mean is σ/√n (or s/√n when σ is unknown).

常见错误：混淆标准差与标准误。在构造均值的置信区间或进行假设检验时，考生常常使用 σ 而不是 σ/√n。这会显著改变区间宽度，并可能导致错误结论。始终记住均值的标准误为 σ/√n（或当 σ 未知时用 s/√n）。

When σ² is unknown and estimated by s², the t-distribution (in Further Statistics) should be used. However, many students ignore this and keep using z-values. This error is particularly damaging for small samples where the t-distribution’s heavier tails make a noticeable difference.

当 σ² 未知而用 s² 估计时，应使用 t 分布（在 Further Statistics 中）。然而众多学生忽略这点，继续使用 z 值。这一错误对于小样本尤为严重，因为 t 分布较重的尾部会带来显著差异。

10. Common Pitfalls and Examiner Tips | 常见失分点与考官建议

Beyond topic-specific errors, there are general patterns that lead to lost marks. These include not showing full working when using statistical tables, misreading the significance level, and failing to interpret results in the context of the problem. OCR examiners consistently emphasise that a final statement such as ‘There is sufficient evidence at the 5% level to reject the null hypothesis and conclude that…’ is required for full marks in hypothesis testing questions.

除了主题特定错误，还有一些共性的失分模式。包括使用统计表时未展示完整计算过程、读错显著性水平，以及未将结果置于题目背景中解读。OCR 考官一贯强调，假设检验题目若要拿满分，必须给出如“在 5% 显著性水平下有充分证据拒绝原假设，并得出……的结论”这样的最终陈述。

Key pitfalls to avoid:

需避免的关键陷阱：

Failing to define the random variable and its distribution at the start of a calculation. Clearly state: ‘Let X ~ B(20, 0.4)’.

在计算开头未定义随机变量及其分布。务必清晰陈述：’Let X ~ B(20, 0.4)’。
Using the wrong tail of the distribution. For a left-tail test, check that you are finding the lower probability, not simply halving the upper tail.

使用了分布的尾部方向。对于左尾检验，确认你求的是下侧概率，而不是简单地将上尾对半分。
Omitting the conclusion or giving a conclusion that does not match the test decision. ‘Do not reject H₀’ is not the same as ‘Accept H₀’.

漏写结论或给出与检验决策不符的结论。“不拒绝 H₀”并不等同于“接受 H₀”。
Forgetting to check the assumptions of the chosen test, e.g., independence, expected frequencies above 5, or normality of residuals in regression.

忘记核查所选检验的假定，例如独立性、期望频数大于 5、或回归中残差的正态性。

Practising with OCR past papers and marking schemes reveals that many mistakes are repetitive. Write out your steps clearly, label hypotheses precisely, and always tie your final answer back to the given scenario. Attention to these details will transform borderline marks into secure grades.

练习 OCR 历年真题和评分方案会发现许多错误是重复出现的。清晰地写出你的步骤，准确标注假设，并将最终答案始终与给定情景联系起来。注意这些细节可以把边缘分数转化为稳定的等级。

Published by TutorHao | Statistics Revision Series | aleveler.com

更多咨询请联系16621398022（同微信）

Year 13 OCR Statistics: High-frequency Exam Topics & Common Mistake Analysis | OCR Year 13 统计高频考点与易错题分析