📚 Statistics 2 Revision: Complete Resource Hub and Exam Guide | Statistics 2 复习完全指南：资源汇总与考点精讲
Statistics 2 (S2) is the unit where A-Level Mathematics stops being about textbook routines and starts demanding genuine statistical thinking. You will meet discrete random variables, the binomial and Poisson distributions, continuous random variables defined by probability density functions, the normal distribution used as an approximation, sampling, estimators and hypothesis tests. This guide gathers the revision material that actually works, organised by exam board, and pairs it with a topic-by-topic walkthrough of the ideas examiners reward most.
Statistics 2（S2）是 A-Level 数学中真正开始考验统计思维的一个单元：离散随机变量、二项分布与泊松分布、由概率密度函数定义的连续随机变量、作为近似工具的正态分布、抽样、估计量与假设检验，都会在这里出现。本文按考试局整理真正好用的复习资料，并配套逐专题讲解考点中最容易得分与最容易失分的细节。
1. What Statistics 2 Actually Tests | S2 到底考什么
S2 is normally assessed by one written paper of about 1 hour 30 minutes. Roughly half of the marks come from short, computational questions and half from longer, multi-part questions that chain several ideas together, such as finding a probability from a distribution and then using it inside a hypothesis test. Calculators are allowed and, on most specifications, expected: the Poisson and normal probabilities are almost always obtained from the calculator's distribution functions rather than from printed tables.
S2 通常由一份约 1 小时 30 分钟的笔试评估。大约一半分数来自短小的计算题，另一半来自多步骤的综合题，例如先从某个分布算出概率，再把它用于假设检验。考试允许且多数考试局默认使用计算器：泊松分布和正态分布的概率几乎都通过计算器的分布函数功能求得，而不是查表。
The examiners' reports repeat the same complaints every session: unstated distribution assumptions, missing continuity corrections, arithmetic slips inside integration limits, and conclusions written without reference to the context of the question. Knowing the syllabus is necessary but not sufficient; the marks are lost in the presentation.
考官的年度报告几乎每次都重复同样的抱怨：没有写明分布假设、漏掉连续性修正、在积分上下限处算错、结论没有回扣题目情境。熟悉大纲是必要条件，但并不充分——丢分往往丢在书写与表述上。
| Topic | 专题 | Typical weight | 常见分值占比 | Question style | 题型特征 |
| Binomial & Poisson | 二项与泊松 | 20-25% | Probability calculations, approximations | 概率计算与近似 |
| Continuous random variables | 连续随机变量 | 20-25% | Find k, E(X), Var(X), median, CDF | 求 k、期望、方差、中位数、分布函数 |
| Normal distribution | 正态分布 | 20-25% | Standardising, inverse normal, approximations | 标准化、逆正态、近似 |
| Sampling & hypothesis tests | 抽样与假设检验 | 25-30% | Unbiased estimators, critical regions, conclusions | 无偏估计、拒绝域、结论表述 |
2. How to Use This Resource Hub | 如何使用这份资源合集
The resources below are organised by the exam boards they are primarily created for, but they are equally handy for other specifications. A binomial question in AQA S2 is the same mathematics as one in Edexcel S2; only the wording of mark schemes and the order of topics differ. If you are short of time, use the notes from your own board and the past papers from every board.
下文资源按主要适用的考试局分类整理，但同样适用于其他考试局。AQA 的二项分布题和 Edexcel 的二项分布题，数学本质完全相同，区别只在于评分方案的措辞和专题顺序。如果时间紧张，就用本考试局的讲义，配合各考试局的历年真题。
A workable study sequence for any one topic is: read the notes, memorise the cheat-sheet formulas, attempt 10 past questions by topic, mark them with the solution bank, then watch a video tutorial only for the questions you still cannot do. Videos feel productive but they are passive; keep them for five minutes at the end, not the start.
任何专题都可以按这样的顺序学习：先读讲义，再记熟速查表公式，然后做 10 道该专题的真题，用解析库自行批改，最后只对仍然做不出的题看视频讲解。看视频感觉高效，其实是被动学习；把它放在学习末端五分钟，而不是开头。
3. Edexcel S2 Resources | Edexcel S2 资源
Edexcel S2 is the most heavily resourced unit of the four. Start with the cheat sheet for orientation, move to the full notes for depth, and keep the revision sheets for the final fortnight when you need speed rather than completeness.
Edexcel S2 是四个考试局中资源最丰富的单元。先用速查表建立整体框架，再用完整讲义深入细节，最后的冲刺两周则使用复习清单，追求速度而不是全面。
- S2 Cheat Sheet: Practical pointers for tackling typical questions and steering clear of common errors, by Dr. Frost.
S2 速查表：Dr. Frost 编写，针对常见题型的实用提示，帮助你避开典型错误。 - Notes: In-depth S2 notes from Simon Baxter (updated June 2016).
讲义：Simon Baxter 编写的深度 S2 笔记（2016 年 6 月更新）。 - Revision Sheets for Edexcel S2.
Edexcel S2 复习清单。 - Papers | 试卷：
- Past papers | 历年真题
- Tiered papers | 分级试卷（Bronze / Silver / Gold）
- Solomon papers | Solomon 系列试卷
- Delphis papers | Delphis 系列试卷
- Past questions by topic | 按专题分类的真题
- Solution Bank | 解析库
- Video Tutorials by ExamSolutions | ExamSolutions 视频讲解
4. AQA S2 Resources | AQA S2 资源
AQA tends to phrase statistical questions in context more heavily than Edexcel, and it expects a concluding sentence that refers back to that context. The summary notes are compact and work well as a final-week checklist.
AQA 的统计题比 Edexcel 更强调情境化表述，并且要求最后一句结论回扣题目情境。AQA 的摘要讲义篇幅紧凑，很适合当作最后一周的检查表使用。
- S2 Summary Notes from mathsbox.org.uk | mathsbox.org.uk 的 S2 摘要讲义
- Revision Sheets for AQA S2 | AQA S2 复习清单
- Papers | 试卷：
- Past papers | 历年真题
- Past questions by topic | 按专题分类的真题
- Video Tutorials by ExamSolutions | ExamSolutions 视频讲解
Because AQA mark schemes reward correct interpretation as well as correct arithmetic, it is worth rewriting every past-paper question in your own words before attempting it. If you cannot state what is being asked in one sentence, you will not know which distribution to choose.
由于 AQA 评分方案同时奖励正确的理解与正确的计算，做题前不妨先用一句话把题目用自己的话复述一遍。如果说不清题目在问什么，就选不出该用哪个分布。
5. OCR S2 Resources | OCR S2 资源
OCR S2 shares the same core content but often reaches it through slightly longer scenarios. The mathshelper notes are concise and pair well with Delphis papers, which are designed as graded practice rather than as full mock papers.
OCR S2 的核心内容与其他考试局一致，但往往通过稍长的情境引入。mathshelper 的笔记简洁，配合 Delphis 试卷使用效果很好——后者是分级练习，而不是完整模拟卷。
- S2 Revision Notes from mathshelper.co.uk | mathshelper.co.uk 的 S2 复习笔记
- Revision Sheets for OCR S2 | OCR S2 复习清单
- Papers | 试卷：
- Past papers | 历年真题
- Delphis papers | Delphis 系列试卷
- Video Tutorials by ExamSolutions | ExamSolutions 视频讲解
6. OCR MEI S2 Resources | OCR MEI S2 资源
MEI is the specification most likely to ask you to justify a modelling assumption or to comment on whether a Poisson model is plausible. Summary sheets are therefore especially valuable: they compress the assumptions into a short list you can quote in an exam.
MEI 是最常要求你说明建模假设、或评论泊松模型是否合理的考试局，因此摘要清单格外有用——它把假设压缩成一份可在考场上直接引用的短清单。
- Revision Sheets for MEI Statistics 2 | MEI Statistics 2 复习清单
- Summary Sheets | 摘要清单
- Papers | 试卷：
- Past papers | 历年真题
- Video Tutorials by ExamSolutions | ExamSolutions 视频讲解
7. Discrete Random Variables and Expectation | 离散随机变量与期望
ctation and variance follow directly from that table, and almost every later topic depends on them.
离散随机变量由它的概率分布定义：一张列出取值 x 与概率 P(X = x) 的表，所有概率之和必须恰好等于 1。期望与方差直接由这张表导出，而后面几乎所有专题都依赖它们。
E(X) = Σ x P(X = x) Var(X) = E(X²) − [E(X)]²
= a²Var(X). Note that adding a constant never changes the variance, and that the coefficient is squared.
两条线性法则能省下大量时间，而且本身常常就是一分方法分。对常数 a 与 b，有 E(aX + b) = aE(X) + b，而 Var(aX + b) = a²Var(X)。注意：加上常数不改变方差，而系数必须平方。
whole table. Compute E(X) and Var(X) from the original distribution and apply the rules. If a question gives you E(X) and Var(X) and asks you to find unknown probabilities, set up simultaneous equations using the fact that the probabilities sum to 1 and the mean matches the stated value.
当题目给出诸如 Y = 3X − 5 的编码变换时，不必重建整张表：先由原分布求出 E(X) 与 Var(X)，再套用上述法则即可。如果题目给出 E(X) 与 Var(X) 要求反求未知概率，就利用"概率之和为 1"以及"均值等于给定值"来建立联立方程。
8. The Binomial and Poisson Distributions | 二项分布与泊松分布
The binomial model applies to a fixed number n of independent trials, each with the same probability p of success. Two conditions do most of the work in exam explanations: a fixed number of trials, and a constant probability with independence between trials.
二项模型适用于固定次数 n 的独立试验，每次成功的概率 p 相同。考试作答时最关键的两个条件是：试验次数固定，以及概率恒定且各次试验相互独立。
X ~ B(n, p) P(X = r) = ⁿCᵣ pʳ (1 − p)ⁿ⁻ʳ E(X) = np Var(X) = np(1 − p)
The Poisson model counts events occurring randomly and independently in a fixed interval of time, length, area or volume, at a constant average rate λ. The key conditions to quote are independence of events, a constant rate, and the fact that events cannot occur simultaneously.
泊松模型统计在固定的时间、长度、面积或体积区间内随机且独立发生的事件次数，平均发生率 λ 恒定。作答时该引用的关键条件是：事件相互独立、发生率恒定，以及事件不可能同时发生。
X ~ Po(λ) P(X = r) = e⁻λ λʳ ÷ r! E(X) = Var(X) = λ
s n > 50 and p < 0.05. The Poisson can be approximated by the normal when λ is large, conventionally λ > 15, using a continuity correction because the Poisson is discrete.
当 n 很大而 p 很小时，可用泊松分布近似二项分布，取 λ = np；常用判定标准是 n > 50 且 p < 0.05。当 λ 较大（通常取 λ > 15）时，可用正态分布近似泊松分布，但由于泊松分布是离散的，必须做连续性修正。
l, symmetry gives P(X ≤ k) = P(X ≥ n − k) when p = 0.5, which can save a long calculation.
记住泊松分布没有上界，所以 P(X ≥ 3) = 1 − P(X ≤ 2) 几乎总是最快的路径。对于二项分布，当 p = 0.5 时对称性给出 P(X ≤ k) = P(X ≥ n − k)，能省去一长串计算。
9. Continuous Random Variables and Probability Density Functions | 连续随机变量与概率密度函数
. Three checks catch most errors: the total area under f(x) must equal 1, f(x) must never be negative, and the median m satisfies F(m) = 0.5.
连续随机变量在任意单点上的概率为零，因此概率来自概率密度函数下方的面积。三个检查可以抓住大部分错误：f(x) 下方的总面积必须等于 1；f(x) 不得取负值；中位数 m 满足 F(m) = 0.5。
∫ f(x) dx = 1 F(x) = ∫₋∞ˣ f(t) dt E(X) = ∫ x f(x) dx Var(X) = E(X²) − [E(X)]²
Finding an unknown constant k is the most common opening part of a question: integrate f(x) over its whole range, set the result equal to 1, and solve. If the range is split into two pieces, integrate each piece separately and add. A piecewise definition often means the mean also needs two integrals.
求未知常数 k 是这类题最常见的开头：在 f(x) 的整个定义域上积分，令结果等于 1 求解。如果定义域分成两段，就分别积分再相加。分段定义往往意味着求期望时也需要做两个积分。
ymmetric about a value c, then E(X) = c and the median is also c, which is a useful check on a long integration.
求众数要看 f(x) 的最大值，而不是 F(x) 的最大值；若密度函数在区间上是线性的，最大值出现在某个端点。当密度函数关于某个值 c 对称时，E(X) = c，中位数也是 c，这可以作为长积分之后的有效检验。
10. The Normal Distribution and Approximations | 正态分布与近似
Every normal question begins with a standardising step, and every error in this topic can be traced back to a standardising step that was skipped or done in the wrong direction.
每道正态分布题都从标准化开始，而这一专题里的所有错误，几乎都能追溯到某一步标准化被跳过，或者方向做反了。
X ~ N(μ, σ²) z = (x − μ) ÷ σ P(X < x) = P(Z < z)
Note carefully that the second parameter of N(μ, σ²) is the variance, not the standard deviation. If a question says the standard deviation is 6 and the mean is 40, then X ~ N(40, 36). Writing N(40, 6) is the single most common careless error in S2.
务必注意：N(μ, σ²) 的第二个参数是方差，不是标准差。若题目给出标准差为 6、均值为 40，则 X ~ N(40, 36)。写成 N(40, 6) 是 S2 中最常见的粗心错误。
For inverse problems, sketch the curve and shade the required region before touching the calculator. If the shaded area is greater than 0.5, your answer lies above the mean; if it is less than 0.5, it lies below. This one habit eliminates most sign errors in questions such as "find the value exceeded by 10% of observations".
做逆问题（已知概率求 x）时，先画草图并标出阴影区域，再动计算器。若阴影面积大于 0.5，答案应在均值右侧；若小于 0.5，则在均值左侧。养成这个习惯，就能消除"求被 10% 观测值超过的数值"这类题目中的绝大多数正负号错误。
tate the approximating distribution and the continuity correction explicitly. A binomial X ~ d by N(np, np(1 − p)), and P(X ≥ 20) becomes P(Y > 19.5) for the continuous variable Y.
使用近似时，要明确写出近似的分布与连续性修正。当 np > 5 且 n(1 − p) > 5 时，X ~ B(n, p) 可用 N(np, np(1 − p)) 近似，此时 P(X ≥ 20) 对连续变量 Y 应写成 P(Y > 19.5)。
11. Sampling, Estimators and Hypothesis Testing | 抽样、估计量与假设检验
t and the most predictable marks. A simple random sample of size n gives the sample mean x̄ as an unbiased estimator of μ, and the estimator s² = Σ(x − x̄ − 1 is what makes it unbiased; dividing by n gives a biased estimator and loses the mark.
这一部分分值最高，也最容易拿分。容量为 n 的简单随机样本中，样本均值 x̄ 是 μ 的无偏估计量，而 s² = Σ(x − x̄)² ÷ (n − 1) 是 σ² 的无偏估计量。除数取 n − 1 才保证无偏；除以 n 会得到有偏估计，直接丢分。
For a hypothesis test, always write the four elements in order: the null hypothesis H₀, the alternative hypothesis H₁, the significance level, and the distribution assumed under H₀. Then compute the test statistic, compare with the critical region or p-value, and finish with a conclusion in the context of the question.
做假设检验时，一定按顺序写出四要素：原假设 H₀、备择假设 H₁、显著性水平，以及在 H₀ 成立时所假设的分布。然后计算检验统计量，与
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