📚 Year 13 OCR Statistics: In-Depth Past Paper Analysis | 高三OCR统计：历年真题深度解析
The OCR A-Level Mathematics Statistics component challenges students to apply statistical thinking to real-world contexts. A deep dive into past papers reveals recurring question types, common pitfalls, and essential techniques that distinguish high achievers. This article dissects past exam questions by topic, offering strategic insights to master the syllabus.
OCR A-Level 数学统计模块要求学生将统计思维应用于实际情境。深入剖析历年真题可发现反复出现的题型、常见错误以及取得高分的关键技巧。本文按主题拆解历年考题，提供策略性见解，助力掌握考试大纲。
1. Understanding the Assessment Structure | 理解评估结构
OCR Statistics Paper 2 (H640/02) is 1 hour 30 minutes, worth 60 marks, covering probability, statistical distributions, hypothesis testing, and data interpretation. Questions often combine multiple topics—for instance, a problem might require binomial probability, then test a hypothesis using a normal approximation. Familiarising with the mark scheme expectations is crucial.
OCR 统计试卷 2 (H640/02) 考试时间 1 小时 30 分钟，总分 60 分，涵盖概率、统计分布、假设检验和数据解释。题目常结合多个考点，例如先计算二项概率，再用正态近似进行假设检验。熟悉评分标准的期待至关重要。
Common command words include ‘state’, ‘calculate’, ‘show that’, and ‘interpret in context’. Past papers often demand precise critical values, clear hypotheses, and contextual conclusions. Missing the ‘contextual interpretation’ can lose 1–2 marks per question. The specification emphasises the use of technology, but written working remains essential for method marks.
常见指令词如“陈述”、“计算”、“证明”和“结合情境解释”。历年真题通常要求准确的临界值、清晰的假设及结合情境的结论。忽略“情境解释”每道题可能扣除 1–2 分。考试大纲强调使用技术工具，但书写解题过程仍是获得方法分的关键。
2. Sampling and Data Presentation | 抽样与数据展示
OCR frequently asks about sampling methods—simple random, stratified, systematic, quota, and opportunity sampling. Know their advantages and biases. In past papers, students often confuse stratified sampling with quota sampling. A common question: ‘Explain how to obtain a stratified sample of size 50 from a population of 200 boys and 300 girls.’ The correct response involves proportional allocation: 50 × (200/500) = 20 boys, 50 × (300/500) = 30 girls.
OCR 常考抽样方法——简单随机抽样、分层抽样、系统抽样、配额抽样和便利抽样。需要了解各自的优缺点与偏差。历年真题中，学生常混淆分层抽样和配额抽样。常见的题目：“解释如何从 200 名男生和 300 名女生总体中抽取容量 50 的分层样本。”正确答案是按比例分配：50×(200/500)=20 名男生，50×(300/500)=30 名女生。
Data presentation: cumulative frequency diagrams, box plots, and histograms. A subtle past-paper twist: interpreting outliers using the 1.5 × IQR rule and justifying removal. When calculating class widths for histograms with unequal intervals, always use frequency density = frequency / class width. OCR often hides a missing frequency behind a given histogram bar area, requiring careful reverse calculation.
数据展示：累积频率图、箱线图和直方图。过去试题中的巧妙之处：利用 1.5 倍 IQR 法则判断异常值并说明移除理由。在不等距直方图中，务必使用频率密度 = 频数 / 组距。OCR 常通过直方图条形的面积隐藏缺失的频数，需要逆向仔细计算。
3. Probability and Conditional Probability | 概率与条件概率
Tree diagrams and Venn diagrams regularly appear. A typical past-paper question: ‘Find P(B’ | A).’ Many candidates mistakenly write P(B’ ∩ A) instead of P(B’ ∩ A)/P(A). Use clear denotation. Another pitfall: assuming independence without checking P(A ∩ B) = P(A)P(B). The mark scheme rewards using the multiplication rule only when independence is justified or stated.
树状图和韦恩图经常出现。典型的真题：“求 P(B’ | A)。”许多考生错误地写出 P(B’ ∩ A) 而非 P(B’ ∩ A)/P(A)。需清晰标注。另一个易错点：未检验 P(A ∩ B)=P(A)P(B) 就假设独立。评分标准仅在验证独立性或题中已说明独立时才给乘法法则分数。
When dealing with ‘at least one’ probability, the complement rule 1 – P(none) often simplifies calculations. In harder questions, conditional probability arises after a first selection without replacement, and a two-way table helps structure the problem. Additionally, using set notation correctly is rewarded: P(A ∪ B) = P(A) + P(B) – P(A ∩ B).
遇到“至少一个”的概率，补集法则 1 – P(无) 常能简化计算。在较难题目中，不放回抽取后的条件概率常出现，双向表格有助于梳理问题。此外，正确使用集合符号会得到加分：P(A ∪ B)=P(A)+P(B)–P(A ∩ B)。
4. Discrete Random Variables and Expectation | 离散随机变量与期望
OCR expects students to construct a probability distribution table and verify ∑ P(X = x) = 1. Then compute E(X) and Var(X) using E(X) = ∑ xp and Var(X) = E(X²) – [E(X)]². Many past answers lose marks by misapplying E(aX + b) = aE(X) + b, especially when combining independent variables, and forgetting that Var(aX + b) = a²Var(X).
OCR 期望考生构建概率分布表并验证 ∑P(X=x)=1。然后使用 E(X)=∑ xp 和 Var(X)=E(X²)–[E(X)]² 计算期望与方差。许多真题答案因错用 E(aX+b)=aE(X)+b 而失分，尤其在组合独立变量时，也常忘记 Var(aX+b)=a²Var(X)。
A classic exam scenario: a game costs c to play, with a prize distribution given. Find the expected profit or the fair price c such that E(profit) = 0. Another common request is to find E(2X + 3Y) given independent variables X and Y. Always show that expectation adds, but variance only adds when variables are independent.
经典考题情境：游戏花费 c 参与，给定奖金分布，求期望利润或使得 E(利润)=0 的公平价格 c。另一常见要求是已知独立变量 X 和 Y，求 E(2X+3Y)。务必体现期望可直接相加，而方差仅当变量独立时才能直接相加。
5. Binomial Distribution: Calculation and Conditions | 二项分布：计算与条件
The binomial distribution X ~ B(n, p) requires fixed number of trials, two possible outcomes, constant probability of success, and independence. Past papers ask to justify why a situation is binomial—do not simply state ‘it’s binomial’. Instead, explicitly mention each condition and link it to the context. For instance, ‘Each egg is either broken or not, the probability of a broken egg is constant at 0.03, and eggs are packed independently.’
二项分布 X~B(n,p) 需满足固定试验次数、两种可能结果、恒定成功概率和独立性。真题要求论证为何情况符合二项分布——不要简单说“它是二项分布”。而应逐一说明每个条件并联系情境。例如：“每个鸡蛋要么破损要么完好，破损的概率恒为 0.03，且各个鸡蛋独立包装。”
Exact probabilities may be found using the formula P(X = k) = nCk pᵏ (1 – p)ⁿ⁻ᵏ or a calculator. However, OCR often expects candidates to use cumulative binomial probability tables. A frequent slip: reading P(X ≥ 4) as 1 – P(X ≤ 3) but using the wrong inequality. Where the table gives P(X ≤ x), always double-check the inequality sign. ‘Show that’ questions often guide you to a critical value, rewarding precise handling of the inequality direction.
精确概率可用公式 P(X=k)=nCk pᵏ(1–p)ⁿ⁻ᵏ 或计算器。然而 OCR 常希望考生使用二项累积概率表。常见滑落：将 P(X≥4) 读成 1–P(X≤3) 时误判不等式。表格给出的是 P(X≤x)，一定要反复核对不等号方向。“证明”类问题常引导你找到一个临界值，准确处理不等号方向可得满分。
6. Normal Distribution: Standardization and Inverse | 正态分布：标准化与逆运算
The standard normal variable Z ~ N(0, 1) is fundamental. OCR provides tables of Φ(z). Students must use sketches, symmetry Φ(–z) = 1 – Φ(z), and P(Z > z) = 1 – Φ(z). A typical past-paper error: using the lower-tail z-value when a question asks for the upper 10% point. Always draw a curve and shade the required region.
标准正态变量 Z~N(0,1) 是基础。OCR 提供 Φ(z) 表。考生必须画草图并利用对称性 Φ(–z)=1–Φ(z) 和 P(Z>z)=1–Φ(z)。真题中典型错误：题目要求上侧 10% 分位点，却用了下侧 z 值。务必画出曲线并给目标区域涂阴影。
Inverse normal: find unknown μ or σ given a probability. Set up standardisation (x – μ)/σ = z and solve. For example, P(X < 12) = 0.15 leads to (12 – μ)/σ = –1.04. Context-based questions demand a final statement interpreting the mean lifespan or a warranty cutoff. Common slip: forgetting to invert the inequality sign when the z-value is negative.
逆正态：已知概率求未知均值 μ 或标准差 σ。建立标准化方程 (x–μ)/σ=z 再求解。例如 P(X<12)=0.15 得 (12–μ)/σ=–1.04。情境题要求最后陈述解释平均寿命或保修截止值。常见遗忘：当 z 值为负时未反转不等式方向。
7. Hypothesis Testing: Binomial and Normal | 假设检验：二项分布与正态分布
A five-step structure is essential: define hypotheses (H₀ and H₁), state significance level α, identify test statistic and its distribution under H₀, determine critical region or compute p-value, and write a conclusion in context. Past papers repeatedly test single-tailed versus two-tailed tests. For a binomial test of p, the test statistic is the number of successes under B(n, p₀). If p-value < α, reject H₀. For discrete distributions, many students forget to state the actual significance level (the exact probability of the critical region).
五步结构至关重要：定义假设 (H₀ 和 H₁)，陈述显著性水平 α，确定检验统计量及其在 H₀ 下的分布，定出临界域或计算 p 值，结合情境写出结论。真题反复考查单尾与双尾检验。对于 p 的二项检验，检验统计量是 H₀ 下 B(n,p₀) 的成功次数。若 p 值 < α，则拒绝 H₀。对于离散分布，许多学生忘记陈述实际显著性水平（临界域的确切概率）。
Normal hypothesis testing for the mean: test statistic Z = (x̄ – μ₀)/(σ/√n). Compare with critical z-value from tables. Contextual conclusion must reference the claim: ‘There is insufficient evidence to reject the company’s claim that the mean is 250 g.’ If using a two-tailed test, compare the p-value with α/2 in each tail or double the tail probability. Remember to use the sample mean given in the question, not the population mean.
正态均值假设检验：检验统计量 Z=(x̄–μ₀)/(σ/√n)。与表中的临界 z 值比较。情境结论必须引述主张：“没有充分证据拒绝公司声称的均值为 250 克。”若用双尾检验，需将 p 值与 α/2 比较或加倍尾部概率。切记使用题目给定的样本均值，而非总体均值。
8. Correlation and Linear Regression | 相关与线性回归
The product moment correlation coefficient (PMCC) r measures linear association. OCR asks to interpret r = 0.812 in context: ‘There is a fairly strong positive linear correlation between…’ Hypothesis tests for correlation often use H
Published by TutorHao | Year 13 统计 Revision Series | aleveler.com
更多咨询请联系16621398022（同微信）
屏轩国际教育cambridge primary/secondary checkpoint, cat4, ukiset,ukcat,igcse,alevel,PAT,STEP,MAT, ibdp,ap,ssat,sat,sat2课程辅导，国外大学本科硕士研究生博士课程论文辅导Cancel reply