📚 Mathematics HL Options: Statistics and Probability Key Concepts | 数学HL选修：统计与概率知识点精讲
For IB Higher Level students, the Statistics and Probability option extends core ideas into powerful tools for inference, modelling and decision‐making. This article explains the essential concepts—discrete and continuous random variables, probability functions, common distributions, confidence intervals and hypothesis testing—with a bilingual approach designed to reinforce both technical precision and conceptual understanding.
对 IB 高等级学生而言，统计与概率选修将核心思想延伸为推断、建模与决策的强大工具。本文以中英双语阐释关键概念——离散与连续随机变量、概率函数、常见分布、置信区间与假设检验，在强化技术准确性的同时加深概念理解，帮助建立严谨的知识体系。
1. Discrete Random Variables | 离散随机变量
A discrete random variable X takes a countable set of possible values. Each value x has an associated probability P(X = x), and the sum of all probabilities equals 1. The probability distribution can be displayed in a table or bar chart. For instance, the number of heads obtained when flipping three fair coins is a discrete random variable taking values 0, 1, 2, 3.
离散随机变量 X 取可数个可能值。每个值 x 对应一个概率 P(X = x)，且所有概率之和为 1。概率分布可用表格或条形图表示。例如，抛掷三枚公平硬币得到的正面个数就是一个离散随机变量，取值为 0、1、2、3。
The probability mass function (PMF) is the function f(x) = P(X = x). For any discrete variable, we require f(x) ≥ 0 for all x and ∑ f(x) = 1. Questions often ask you to find an unknown constant in a PMF or compute probabilities of compound events using addition rules.
概率质量函数（PMF）是函数 f(x) = P(X = x)。对于任何离散变量，要求对所有 x 满足 f(x) ≥ 0 且 ∑ f(x) = 1。题目常要求找出 PMF 中的未知常数，或利用加法规则计算复合事件的概率。
2. Continuous Random Variables | 连续随机变量
A continuous random variable can take any value in an interval on the real line. Instead of giving point probabilities, we describe it with a probability density function (PDF) f(x) such that the area under the curve equals 1. For any specific value x, P(X = x) = 0; probabilities are computed over intervals as integrals of f(x).
连续随机变量可取实轴上某区间内的任意值。与点概率不同，我们用概率密度函数（PDF）f(x) 描述它，且曲线下的面积等于 1。对任何特定值 x，P(X = x) = 0；概率通过求 f(x) 在区间上的积分得到。
A key skill is verifying that a given function is a valid PDF: it must be non‑negative everywhere and its integral over the entire domain must be 1. The probability P(a < X < b) is then given by ∫ₐᵇ f(x) dx. Unlike discrete variables, inequalities are strict or non‑strict without affecting the probability.
一个关键技能是验证给定函数是否为有效的 PDF：它必须处处非负且在整个定义域上的积分为 1。此时概率 P(a < X < b) 由 ∫ₐᵇ f(x) dx 给出。与离散变量不同，不等式严格与否不影响概率大小。
3. Cumulative Distribution Function (CDF) | 累积分布函数
The cumulative distribution function F(x) = P(X ≤ x) works for both discrete and continuous cases. It is a non‑decreasing function with F(x) → 0 as x → –∞ and F(x) → 1 as x → +∞. For continuous variables, F(x) = ∫₋∞ˣ f(t) dt, and the PDF can be recovered by differentiating: f(x) = F′(x) wherever F is differentiable.
累积分布函数 F(x) = P(X ≤ x) 对离散和连续情形均适用。它是一个非递减函数，且当 x → –∞ 时 F(x) → 0，当 x → +∞ 时 F(x) → 1。对于连续变量，F(x) = ∫₋∞ˣ f(t) dt，而 PDF 可通过求导还原：在 F 可微处 f(x) = F′(x)。
In exam problems you often need to find median m from F(m) = 0.5, quartiles from F(Q₁) = 0.25, F(Q₃) = 0.75, or to obtain probabilities such as P(X > a) = 1 – F(a). The CDF can also handle mixed distributions where some probabilities are concentrated at points.
在考题中常需从 F(m) = 0.5 求中位数 m，从 F(Q₁) = 0.25、F(Q₃) = 0.75 求四分位数，或计算形如 P(X > a) = 1 – F(a) 的概率。CDF 还可处理某些概率集中于单个点的混合分布。
4. Expectation and Variance | 期望与方差
The expected value (mean) of a random variable is the long‑run average. For discrete X, E(X) = ∑ x·P(X = x); for continuous X, E(X) = ∫ x·f(x) dx over the domain. Linearity of expectation holds: E(aX + b) = aE(X) + b, whether or not X and Y are independent.
随机变量的期望值（均值）是长期平均值。对离散 X，E(X) = ∑ x·P(X = x)；对连续 X，E(X) = ∫ x·f(x) dx 在整个定义域上计算。期望的线性性质成立：E(aX + b) = aE(X) + b，无论 X 与 Y 是否独立。
The variance measures spread: Var(X) = E[(X – μ)²] = E(X²) – [E(X)]². The standard deviation is σ = √Var(X). Key properties: Var(aX + b) = a²Var(X). For sums of independent random variables, Var(X + Y) = Var(X) + Var(Y). Expectation and variance are the foundation for later inference procedures.
方差衡量离散程度：Var(X) = E[(X – μ)²] = E(X²) – [E(X)]²。标准差为 σ = √Var(X)。重要性质：Var(aX + b) = a²Var(X)。对于独立随机变量之和，Var(X + Y) = Var(X) + Var(Y)。期望与方差是后续推断方法的基础。
5. The Binomial Distribution | 二项分布
When we count the number of successes in n independent Bernoulli trials, each with success probability p, the random variable X follows a binomial distribution: X ~ B(n, p). Its PMF is P(X = k) = ⁿCₖ pᵏ (1 – p)ⁿ⁻ᵏ for k = 0, 1, …, n. Here ⁿCₖ = n!/(k!(n – k)!).
当我们计算 n 次独立伯努利试验中成功的次数，每次成功概率为 p，则随机变量 X 服从二项分布：X ~ B(n, p)。其 PMF 为 P(X = k) = ⁿCₖ pᵏ (1 – p)ⁿ⁻ᵏ，k = 0, 1, …, n。此处 ⁿCₖ = n!/(k!(n – k)!)。
For X ~ B(n, p), the mean is E(X) = np and the variance is Var(X) = np(1 – p). As n grows large, the binomial distribution becomes approximately normal (provided np and n(1–p) are sufficiently large). This normal approximation is a central tool in hypothesis testing of proportions.
对 X ~ B(n, p)，均值 E(X) = np，方差 Var(X) = np(1 – p)。当 n 很大时，二项分布近似正态分布（只要 np 和 n(1–p) 足够大）。这一正态近似是比例假设检验的核心工具。
6. The Poisson Distribution | 泊松分布
The Poisson distribution models the number of rare events occurring in a fixed interval of time or space, given a constant mean rate λ. X ~ Po(λ) has PMF P(X = k) = (λᵏ e⁻λ) / k! for k = 0, 1, 2, … . Both the mean and variance equal λ. This makes Poisson useful for counts such as phone calls arriving per minute or defects per metre of fabric.
泊松分布用于描述单位时间或空间内稀有事件发生的次数，给定恒定平均率 λ。X ~ Po(λ) 的 PMF 为 P(X = k) = (λᵏ e⁻λ) / k!，k = 0, 1, 2, …。其均值和方差均等于 λ。因此泊松分布适用于诸如每分钟到达的电话数或每米织物上的疵点数等计数场景。
The Poisson distribution can approximate the binomial B(n, p) when n is large and p is small, by setting λ = np. A common criterion is n ≥ 20 and p ≤ 0.05, or n ≥ 100 and np ≤ 10. Additionally, the sum of independent Poisson variables is also Poisson: if X ~ Po(λ₁) and Y ~ Po(λ₂) are independent, then X + Y ~ Po(λ₁ + λ₂).
当 n 大而 p 小，令 λ = np，泊松分布可近似二项分布 B(n, p)。常用判据为 n ≥ 20 且 p ≤ 0.05，或 n ≥ 100 且 np ≤ 10。此外，独立泊松变量之和仍服从泊松分布：若 X ~ Po(λ₁)、Y ~ Po(λ₂) 独立，则 X + Y ~ Po(λ₁ + λ₂)。
7. The Normal Distribution | 正态分布
The normal distribution, X ~ N(μ, σ²), is the most important continuous model in statistics. Its bell‑shaped PDF is f(x) = (1/(σ√(2π))) e^{–½((x–μ)/σ)²}. The curve is symmetric about μ, and the total area under the curve is 1. Linear transformations preserve normality: if X ~ N(μ, σ²), then aX + b ~ N(aμ + b, a²σ²).
正态分布 X ~ N(μ, σ²) 是统计中最重要的连续模型。其钟形 PDF 为 f(x) = (1/(σ√(2π))) e^{–½((x–μ)/σ)²}。曲线关于 μ 对称，曲线下总面积为 1。线性变换保持正态性：若 X ~ N(μ, σ²)，则 aX + b ~ N(aμ + b, a²σ²)。
We standardise to Z ~ N(0, 1) using Z = (X – μ)/σ. Probability statements then involve the standard normal CDF Φ(z). Tables or calculators give Φ(z) = P(Z ≤ z). For any X ~ N(μ, σ²), P(X < a) = Φ((a – μ)/σ). The empirical rule says about 68% of data lie within μ ± σ, 95% within μ ± 2σ, and 99.7% within μ ± 3σ.
我们用 Z = (X – μ)/σ 标准化至 Z ~ N(0, 1)。概率表述进而涉及标准正态 CDF Φ(z)。表格或计算器给出 Φ(z) = P(Z ≤ z)。对任意 X ~ N(μ, σ²)，P(X < a) = Φ((a – μ)/σ)。经验法则表明约 68% 的数据落在 μ ± σ 内，约 95% 在 μ ± 2σ 内，99.7% 在 μ ± 3σ 内。
8. Sampling Distribution of the Mean | 均值的抽样分布
When independent observations are drawn from a population with mean μ and variance σ², the sample mean X̄ is a random variable. Its expected value is E(X̄) = μ, and its variance is Var(X̄) = σ²/n. The Central Limit Theorem (CLT) states that for large n (typically n ≥ 30), the distribution of X̄ approaches N(μ, σ²/n) regardless of the population shape.
当我们从均值为 μ、方差为 σ² 的总体中抽取独立观测值时，样本均值 X̄ 是一个随机变量。其期望为 E(X̄) = μ，方差为 Var(X̄) = σ²/n。中心极限定理（CLT）指出，对于大 n（通常 n ≥ 30），无论总体形状如何，X̄ 的分布趋近于 N(μ, σ²/n)。
If the population itself is normal, then X̄ is exactly normal for any sample size. This result underpins confidence intervals and hypothesis tests for a mean. When σ is unknown, we estimate it with the sample standard deviation s and use the t‑distribution with (n – 1) degrees of freedom.
若总体本身为正态，则对于任何样本容量 X̄ 都是精确正态的。这一结果为均值的置信区间和假设检验提供了依据。当 σ 未知时，我们用样本标准差 s 估计它，并使用自由度为 (n – 1) 的 t 分布。
9. Confidence Intervals for the Mean | 均值的置信区间
A 95% confidence interval for the population mean μ, when σ is known, is x̄ ± z* × σ/√n, where z* is the critical value from N(0,1) corresponding to the middle 95%. For 95% confidence, z* ≈ 1.96. The interpretation: if we took many samples and constructed intervals, about 95% of them would capture the true μ.
当 σ 已知时，总体均值 μ 的 95% 置信区间为 x̄ ± z* × σ/√n，其中 z* 是 N(0,1) 中对应中间 95% 的临界值。对于 95% 置信度，z* ≈ 1.96。其解释为：若我们多次抽样并构造区间，约 95% 的区间将包含真实的 μ。
When σ is unknown (common in reality), the interval uses the t‑distribution: x̄ ± t* × s/√n, with t* having n – 1 degrees of freedom. Wider intervals reflect the extra uncertainty from estimating σ. The margin of error decreases as sample size n grows, because the standard error σ/√n becomes smaller.
当 σ 未知时（现实中常见），区间使用 t 分布：x̄ ± t* × s/√n，t* 的自由度为 n – 1。区间更宽反映了由估计 σ 带来的额外不确定性。因标准误 σ/√n 随样本容量 n 增大而减小，误差边际也随之缩小。
10. Hypothesis Testing for a Mean | 均值的假设检验
A hypothesis test evaluates evidence against a null hypothesis H₀ about a population parameter. For a mean, we often test H₀: μ = μ₀ against a one‑ or two‑sided alternative. The test statistic is Z = (x̄ – μ₀)/(σ/√n) if σ is known, or T = (x̄ – μ₀)/(s/√n) if σ is unknown. The p‑value is the probability of observing a statistic as extreme as, or more extreme than, the one obtained, under H₀.
假设检验评估反对总体参数零假设 H₀ 的证据。对均值，常检验 H₀: μ = μ₀，对立假设为单侧或双侧。若 σ 已知，检验统计量为 Z = (x̄ – μ₀)/(σ/√n)；若 σ 未知，则为 T = (x̄ – μ₀)/(s/√n)。p 值是在 H₀ 为真的条件下，观察到与所得统计量一样极端或更极端的统计量的概率。
In the critical‑value approach, we compare the test statistic with a threshold (e.g. ±1.96 for a two‑sided test at 5% significance). If |Z| > z*, we reject H₀. In the p‑value method, we reject H₀ if p‑value < α (e.g. α = 0.05). Note that failing to reject H₀ does not prove H₀ is true; it only indicates insufficient evidence against it.
在临界值方法中，我们将检验统计量与阈值比较（如双侧检验 5% 显著性水平下为 ±1.96）。若 |Z| > z*，则拒绝 H₀。在 p 值方法中，若 p < α（如 α = 0.05），则拒绝 H₀。注意，未拒绝 H₀ 并不证明 H₀ 为真，仅表示缺乏足够的反对证据。
For paired data (before/after scenarios), we test the mean difference using one‑sample techniques on the differences. Always check assumptions: random sampling, independence, normality of the underlying population (or large n via CLT).
对于配对数据（前后对比），我们对差值应用单样本技术检验均值差。始终应检查假定：随机抽样、独立性、总体正态性（或通过 CLT 满足的大 n 条件）。
11. Type I and Type II Errors | 第 I 类与第 II 类错误
A Type I error occurs when H₀ is true but we reject it; its probability is α (the significance level). A Type II error occurs when H₀ is false but we fail to reject it; its probability is β. The power of the test is 1 – β, the probability of correctly rejecting a false H₀. Power increases with larger sample size, larger effect size, and higher α.
第 I 类错误发生在 H₀ 为真但被拒绝时；其概率为 α（显著性水平）。第 II 类错误发生在 H₀ 为假但未被拒绝时；其概率为 β。检验的功效为 1 – β，即正确拒绝错误 H₀ 的概率。功效随样本容量增大、效应量增大及 α 提高而增加。
Exam problems may ask you to interpret these errors in context: for example, concluding a new drug works when it really doesn’t (Type I), or failing to detect a genuine effect (Type II). Designing a study often involves choosing n so that power is at least 80% for a given effect size.
考题常要求在语境中解释这些错误：例如，在新药实际无效时却推断其有效（第 I 类），或未能检测到真实效应（第 II 类）。研究设计常涉及选择 n 以使在给定效应量下功效至少达到 80%。
12. Correlation and Linear Regression | 相关与线性回归
Pearson’s product‑moment correlation coefficient r measures the strength and direction of a linear relationship between two variables X and Y. It is calculated from paired data (xᵢ, yᵢ) and satisfies –1 ≤ r ≤ 1. A value r close to +1 indicates strong positive linear association; r close to –1 indicates strong negative linear association; r ≈ 0 suggests no linear pattern.
皮尔逊积矩相关系数 r 衡量两变量 X 和 Y 之间线性关系的强度与方向。它由成对数据 (xᵢ, yᵢ) 计算得出，且满足 –1 ≤ r ≤ 1。r 接近 +1 表明强正线性相关；r 接近 –1 表明强负线性相关；r ≈ 0 表示没有线性模式。
The least‑squares regression line of y on x has equation y = a + bx, where the slope b = r × (sᵧ/sₓ) and the intercept a = ȳ – b x̄. The line minimises the sum of squared vertical residuals. The coefficient of determination R² = r² tells us the proportion of variation in y explained by the linear relationship with x.
y 关于 x 的最小二乘回归直线方程为 y = a + bx，其中斜率 b = r × (sᵧ/sₓ)，截距 a = ȳ – b x̄。该直线使竖直残差平方和最小。决定系数 R² = r² 表示 y 的变异中可由与 x 的线性关系解释的比例。
Residual plots help assess model appropriateness: residuals should be randomly scattered around zero with no clear pattern. A curved pattern suggests a non‑linear model may be better. Outliers and influential points can distort the regression line and should be examined carefully.
残差图有助于评估模型适宜性：残差应围绕零随机散布，无明显模式。出现弯曲模式提示非线性模型可能更好。异常值和强影响点可扭曲回归直线，应仔细审查。
Published by TutorHao | Mathematics HL 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