Option 7: Statistics and Probability | 选项七：统计与概率

📚 Option 7: Statistics and Probability | 选项七：统计与概率

For many IB Higher Level students, the Statistics and Probability option represents both a fascinating intellectual challenge and a powerful toolkit for real-world problem solving. This branch of mathematics equips you to model uncertainty, draw reliable conclusions from data, and understand the theoretical underpinnings of random phenomena. In the Cambridge 2012 edition by Fannon, Kadelburg, Woolley and Ward, the option is structured to build from fundamental distribution theory through advanced inferential techniques, culminating in modern approaches to estimation and hypothesis testing. The following revision guide covers the essential concepts, ensuring you are thoroughly prepared for the demands of Paper 3.

对于许多 IB 高级水平的学生来说，统计与概率选项既是一个迷人的智力挑战，也是解决实际问题的强大工具。这一数学分支使你能够对不确定性进行建模，从数据中得出可靠的结论，并理解随机现象的理论基础。在剑桥大学出版社 2012 年出版的教材中，作者 Fannon、Kadelburg、Woolley 和 Ward 将该选项的内容从基础分布理论逐步构建到高级推断技术，最终呈现现代估计与假设检验的方法。以下复习指南涵盖了所有核心概念，确保你为第三卷的要求做好全面准备。

1. Foundational Probability Concepts | 概率基础概念

A solid grasp of probability is the bedrock of the entire option. You need to be comfortable with sample spaces, events, and the axioms of probability, including the addition and multiplication rules. Conditional probability and the law of total probability become especially important when handling compound experiments or data that exhibit dependence.

扎实掌握概率是整个选项的基石。你需要熟悉样本空间、事件以及概率公理，包括加法和乘法法则。在处理复合实验或呈现依赖关系的数据时，条件概率与全概率公式显得尤其重要。

checked carefully; do not assume it unless the problem states it or you have verified P(A ∩ urring tool that allows you to update probabilities as new evidence emerges. In examination questions, always identify the prior probabilities and the likelihoods before applying the formula.

事件的独立性必须仔细检验；除非题目明确说明或你已验证 P(A ∩ B) = P(A)P(B)，否则不要假定事件独立。贝叶斯定理是一个反复出现的工具，它允许你在新证据出现时更新概率。在考题中，应用公式之前一定要先识先验概率和似然。

2. Discrete Random Variables and Expectation | 离散随机变量与期望

ty of expectation, E(aX + b) = aE(X) + b, is a time-saving device that appears repeatedly.

离散随机变量由概率质量函数（pmf）刻画。你应能从文字题出发建立 pmf，验证概率之和为 1，并利用定义计算期望值 E(X) 与方差 Var(X)。期望的线性性质 E(aX + b) = aE(X) + b 是一个反复出现、能节省时间的工具。

The variance can be computed efficiently with the formula Var(X) = E(X²) – e involved, joint distributions and covariance become key; remember that Cov(X, Y) = E(XY) – ) and that independence implies zero covariance, but the converse is not necessarily true.

方差可通过公式 Var(X) = E(X²) – [E(X)]² 高效计算。你还必须能够处理随机变量的函数 g(X) 并求出其期望。当涉及多个离散变量时，联合分布和协方差成为关键；记住 Cov(X, Y) = E(XY) – E(X)E(Y)，且独立性蕴含零协方差，但反之未必成立。

3. Common Discrete Distributions | 常见离散分布

on Bin(n, p) models the number of successes in n independent trials, with E(X) = np and Var(X) = np(1 – p).

该选项的考试要求你熟练掌握二项分布、泊松分布、几何分布和负二项分布的详细知识。对于每个分布，你必须知道其产生的背景、概率质量函数的形式、均值和方差，以及任何特殊性质。例如，二项分布 Bin(n, p) 对 n 次独立试验中的成功次数进行建模，且 E(X) = np，Var(X) = np(1 – p)。

cribes the trial number of the first success, while the negative binomial extends this to the number of trials required to achieve a fixed number of successes. Be ready to derive probabilities or to recognise which distribution fits a given scenario.

泊松分布 Po(λ) 是当 n → ∞ 且 p → 0 使得 np = λ 保持常数的二项分布极限。它用于描述在时间或空间中随机发生的事件，且 E(X) = Var(X) = λ。几何分布描述首次成功所需的试验次数，而负二项分布则将之推广到获得固定次数成功所需的试验次数。你要准备好推导概率，或识别哪种分布适合给定的情景。

4. Continuous Random Variables | 连续随机变量

en useful, and you should be able to obtain it by integration and use it to find probabilities.

当转向连续随机变量时，概率密度函数（pdf）取代了概率质量函数。你必须非常清楚，概率由密度曲线下的面积给出，且总面积必须等于 1。累积分布函数（cdf）F(x) = P(X ≤ x) 通常很有用，你需要能够通过积分获得它，并利用它求概率。

e defined using integrals: E(X) = ∫ x f(x) dx, E(g(X)) = ∫ g(x) f(x) dx, and Var(X) = ∫ (x – ge-of-variable technique, remembering to multiply by the absolute value of the derivative of the inverse function.

连续变量的期望与方差通过积分定义：E(X) = ∫ x f(x) dx，E(g(X)) = ∫ g(x) f(x) dx，且 Var(X) = ∫ (x – μ)² f(x) dx。你也将遇到连续随机变量的变换。若 Y = g(X) 是单调的，可利用换元法求得 Y 的概率密度函数，并记住要乘以反函数的导数的绝对值。

5. Important Continuous Distributions | 重要的连续分布

σ²) is central. You must be able to standardise a normal variable to the standard normal Z ~ often used for waiting times, with pdf f(x) = λe⁻λx for x ≥ 0, mean 1/λ and variance 1/λ².

正态分布 N(μ, σ²) 占据中心地位。你必须能够将正态变量标准化为标准正态变量 Z ~ N(0, 1) 并使用统计表。独立正态变量的线性组合仍服从正态分布这一性质你会经常应用。指数分布是另一个关键模型，常用于等待时间，其概率密度函数为 f(x) = λe⁻λx（x ≥ 0），均值为 1/λ，方差为 1/λ²。

Other continuous distributions you should know include the uniform distribution on an interval [a, b] and the gamma distribution, from which the exponential and chi-squared distributions are derived. The chi-squared distribution in particular is crucial for hypothesis testing. Do not forget to check whether a function is a valid pdf: it must be non‑negative and integrate to 1 over its domain.

你还应知道其他连续分布，包括区间 [a, b] 上的均匀分布以及伽马分布，指数分布和卡方分布均是伽马分布的特例。卡方分布在假设检验中尤其重要。别忘了检查一个函数是否为有效的概率密度函数：它必须非负且在其定义域上的积分为 1。

6. Generating Functions | 生成函数

y generating functions (PGFs) are a powerful tool for discrete random variables taking non‑negative integer values. Defined as G(t) = E(tˣ) = Σ tˣ distribution. Its derivatives at t = 1 give factorial moments: G′(1) = E(X), G″(1) = E[X(X – he PGF of a sum of independent discrete variables is the product of their individual PGFs.

概率生成函数（PGF）是一种处理取非负整数值的离散随机变量的有力工具。其定义为 G(t) = E(tˣ) = Σ tˣ P(X = x)，PGF 能唯一识别分布。它在 t = 1 处的导数给出阶乘矩：G′(1) = E(X)，G″(1) = E[X(X – 1)]，由此可求出方差。独立离散变量之和的 PGF 是单个变量 PGF 的乘积。

serve a similar purpose for both discrete and continuous variables, defined as M(t) = E(eᵗˣ pendent, identically distributed variables and for proving limit theorems such as the central limit theorem.

矩生成函数（MGF）对离散和连续变量都有类似作用，其定义为 M(t) = E(eᵗˣ)，前提是期望在 0 的某邻域内存在。第 n 阶原点矩为 M⁽ⁿ⁾(0)。在处理独立同分布变量之和以及证明极限定理（如中心极限定理）时，MGF 尤其有用。

广告 / Sponsored

7. The Central Limit Theorem and Sampling Distributions | 中心极限定理与样本分布

at, for a sufficiently large sample size n, the sampling distribution of the sample mean X̄ lations that are nearly symmetric it works reasonably even for smaller n. The standard error of the mean is σ/√n.

中心极限定理（CLT）指出，当样本量 n 足够大时，无论总体分布形状如何，只要总体具有有限方差，样本均值 X̄ 的抽样分布近似为正态。当 n ≥ 30 时这一近似是可靠的，但对于近乎对称的总体，即便 n 较小也能较好地工作。均值的标准误为 σ/√n。

You must also understand other sampling distributions derived from the normal: the chi‑squared distribution χ²(n), which arises from the sum of squares of independent standard normal variables; the t‑distribution t(n) used when the population variance is estimated; and the F‑distribution for comparing two variances. Knowing how these distributions emerge in confidence intervals and test statistics is essential for the inference sections.

你还必须理解由正态分布导出的其他抽样分布：卡方分布 χ²(n)，它来源于独立标准正态变量的平方和；t 分布 t(n)，在总体方差被估计时使用；以及用于比较两个方差的 F 分布。在置信区间和检验统计量中知晓这些分布如何出现，对推断部分至关重要。

8. Point Estimation and Properties of Estimators | 点估计与估计量的性质

An estimator is a rule, usually a statistic calculated from sample data, used to infer the value of an unknown population parameter. The most common methods are the method of moments and the method of maximum likelihood. Maximum likelihood estimators (MLEs) are often preferred because they possess desirable asymptotic properties: consistency, asymptotic normality, and efficiency.

估计量是一条规则，通常是从样本数据计算出的统计量，用于推断未知总体参数的值。最常见的方法是矩估计法和最大似然估计法。最大似然估计量（MLE）常被优先选用，因为它们具有良好的渐近性质：一致性、渐近正态性和有效性。

d MLEs by setting up the likelihood function L(θ) = ∏ f(xᵢ; θ), often working with the log‑likelihood ℓ(θ) = ln L(θ) for computational ease. An estimator is unbiased if E(θ̂ on, and for MLEs it can be approximated using the observed or expected Fisher information.

你需要能够通过建立似然函数 L(θ) = ∏ f(xᵢ; θ) 来求 MLE，为便于计算常使用对数似然函数 ℓ(θ) = ln L(θ)。若 E(θ̂) = θ，则估计量是无偏的；你可能会被要求证明偏差或加以修正。估计量的标准误给出了其精度度量，而对 MLE 而言，可利用观测或期望的费希尔信息量进行近似。

9. Confidence Intervals | 置信区间

ausible values for a population parameter, with a specified confidence level (typically 95%). For a normal mean with known variance σ², the interval is X̄ ± z* σ/√n, where z* is the critical value from the standard normal distribution. When the variance is unknown, switch to the t distribution and use the sample standard deviation s.

置信区间在指定的置信水平（通常为 95%）下给出了总体参数的可能取值范围。对于已知方差 σ² 的正态均值，区间为 X̄ ± z* σ/√n，其中 z* 是标准正态分布的临界值。当方差未知时，转而使用 t 分布，并用样本标准差 s 代替。

You will also construct confidence intervals for the difference of two means, for a population proportion, and for the variance or standard deviation of a normal population using the chi‑squared distribution. The interpretation is subtle: a 95% confidence interval means that if we repeated the sampling procedure many times, about 95% of the intervals so constructed would contain the true parameter – not that there is a 95% probability that the parameter lies in the specific interval computed.

你还会构造两个均值之差的置信区间、总体比例的置信区间，以及使用卡方分布构造正态总体方差或标准差的置信区间。其解释十分微妙：95% 置信区间意味着，如果我们多次重复抽样过程，所构造的区间中大约有 95% 会包含真实参数，而非该特定区间的参数有 95% 的概率落入其中。

10. Hypothesis Testing | 假设检验

Hypothesis testing is a formal procedure for evaluating claims about a population. You must clearly state the null hypothesis H₀ and the alternative hypothesis H₁, choose an appropriate test statistic, and compute its observed value. The p‑value is the probability, under H₀, of obtaining a result at least as extreme as the one observed; a small p‑value leads to the rejection of H₀.

假设检验是评价关于总体的断言的形式化程序。你必须清晰地陈述原假设 H₀ 和备择假设 H₁，选择合适的检验统计量，并计算其观测值。p 值是在 H₀ 下得到至少与观测结果一样极端的结果的概率；一个很小的 p 值将导致拒绝 H₀。

For a single normal mean, the test statistic is Z = (X̄ – μ₀) / (σ/√n) when σ is known, and t = (X̄ – μ₀) / (s/√n) with n–1 degrees of freedom when σ is estimated. Type I error (rejecting H₀ when it is true) is controlled by the significance level α; Type II error (failing to reject H₀ when it is false) is related to the power of the test. Be ready to perform tests on proportions, differences of means, and variances, as well as paired and two‑sample tests.

对于单个正态均值，当 σ 已知时检验统计量为 Z = (X̄ – μ₀) / (σ/√n)；当 σ 被估计时则用 t = (X̄ – μ₀) / (s/√n)，自由度为 n–1。第 I 类错误（H₀ 为真时拒绝 H₀）由显著性水平 α 控制；第 II 类错误（H₀ 为假时未拒绝 H₀）与检验的功效相关。要准备好对比例、均值之差和方差进行检验，并能进行配对和双样本检验。

11. Chi‑Squared Tests | 卡方检验

Chi‑squared tests are used for categorical data. The goodness‑of‑ quencies conform to a specified theoretical distribution. The test statistic is χ² = Σ (Oᵢ – his statistic follows approximately a χ² distribution with appropriate degrees of freedom.

卡方检验用于分类数据。拟合优度检验检查观测频数是否符合某个指定的理论分布。检验统计量为 χ² = Σ (Oᵢ – Eᵢ)² / Eᵢ，其中 Oᵢ 与 Eᵢ 分别为观测频数和期望频数。在原假设下，该统计量近似服从具有适当自由度的 χ² 分布。

The test of independence for a two‑ ion of independence as (row total × column total) / grand total. Degrees of freedom are (r – 1)(c – pected frequencies exceed 5; if not, combine categories or use an appropriate alternative.

双向列联表的独立性检验考察两个分类变量是否有关联。在独立性假设下，期望频数通过（行和 × 列和）/ 总和计算。自由度为 (r – 1)(c – 1)。显著的结果表明存在关联，但并非因果关系。你必须确保所有期望频数大于 5；否则应合并类别或采用适当的替代方法。

12. Introduction to Bayesian Statistics | 贝叶斯统计导论

The Bayesian approach treats parameters as random variables with prior distributions that reflect our beliefs before observing the data. After data are collected, Bayes’ theorem is used to update the prior to the posterior distribution. The posterior combines prior information and the likelihood from the data. In the context of the IB option, you typically work with Beta prior distributions for a binomial proportion p, or Normal priors for a Normal mean, leading to conjugate posteriors.

贝叶斯方法将参数视为具有先验分布的随机变量，先验分布反映了我们在观测数据之前的信念。收集数据后，利用贝叶斯定理将先验更新为后验分布。后验分布结合了先验信息和来自数据的似然。在 IB 该选项的语境下，你通常会使用二项比例 p 的 Beta 先验分布，或正态均值的正态先验分布，从而得到共轭后验。

Key skills include identifying the posterior parameters from given summary statistics, computing posterior means and variances, and constructing credible intervals – the Bayesian counterpart to confidence intervals. A 95% credible interval is the region where the parameter lies with probability 0.95, given the data and the prior. This direct probability interpretation is one reason Bayesian methods are increasingly popular in modern statistics.

关键技能包括从给定的汇总统计量中识别后验参数、计算后验均值和方差，以及构建可信区间——这是贝叶斯学派对应置信区间的概念。在给定数据和先验的条件下，95% 可信区间是参数以 0.95 的概率落入的区域。这种直接的概率解释是贝叶斯方法在现代统计学中日益流行的原因之一。

---

Published by TutorHao | Mathematics Revision Series | aleveler.com

更多咨询请联系16621398022（同微信）