Core Knowledge Points for Pre-U CAIE Statistics | Pre-U CAIE 统计：核心知识点梳理

📚 Core Knowledge Points for Pre-U CAIE Statistics | Pre-U CAIE 统计：核心知识点梳理

The Pre-U CAIE Statistics course builds a rigorous foundation in statistical thinking, from exploratory data analysis to formal inference. This article distills the essential knowledge points across the syllabus, designed to support both your conceptual understanding and exam readiness. Whether you are revising probability distributions, hypothesis tests, or regression, the following sections provide a concise yet comprehensive walkthrough of the core topics every candidate must master.

Pre-U CAIE 统计课程旨在为学生打下严谨的统计思维基础，涵盖从探索性数据分析到正式推断的各个方面。本文梳理了考纲中的必备知识点，旨在帮助读者巩固概念理解并做好应试准备。无论你是在复习概率分布、假设检验还是回归分析，以下各节都将对每位考生必须掌握的核心内容进行简明而全面的梳理。

1. Data Summarisation: Central Tendency and Dispersion | 数据汇总：中心趋势与离散程度

Understanding a dataset begins with measures of central tendency. The mean, median and mode each capture a different aspect of the data’s centre. The sample mean is given by x̄ = Σxᵢ / n, while the median is the middle value when data are ordered. The mode is simply the most frequent observation. In grouped data, linear interpolation is used to estimate the median and quartiles.

理解一组数据须从中心趋势的度量开始。均值、中位数和众数各自刻画了数据中心的不同侧面。样本均值由 x̄ = Σxᵢ / n 给出，而中位数是数据排序后的中间值，众数则是最常出现的观测值。对于分组数据，常用线性插值来估计中位数和四分位数。

Equally important are measures of dispersion. The range, interquartile range (IQR = Q₃ – Q₁) and standard deviation tell us how spread out the observations are. The sample variance s² = Σ(xᵢ – x̄)² / (n – 1) uses n–1 to provide an unbiased estimate of the population variance. For grouped data, the calculation uses midpoints and frequencies, and coding may be applied to simplify arithmetic.

离散程度的度量同样重要。极差、四分位距（IQR = Q₃ – Q₁）和标准差告诉我们观测值的分散情况。样本方差 s² = Σ(xᵢ – x̄)² / (n – 1) 采用 n–1 作为分母，以便给出总体方差的无偏估计。对于分组数据，计算时使用组中值和频数，并可通过编码简化运算。

Outliers can be identified using the 1.5 × IQR rule: any value below Q₁ – 1.5 IQR or above Q₃ + 1.5 IQR is considered an outlier. Box-and-whisker plots provide a visual summary of these five-number summary statistics.

异常值可用 1.5 × IQR 法则识别：任何小于 Q₁ – 1.5 IQR 或大于 Q₃ + 1.5 IQR 的值均被视为异常值。箱线图则对这些五数概括统计量进行可视化呈现。

2. Probability | 概率

Probability underpins all inferential statistics. The basic rules include the addition rule for mutually exclusive events, P(A ∪ B) = P(A) + P(B), and the general addition rule, P(A ∪ B) = P(A) + P(B) – P(A ∩ B). Conditional probability is expressed as P(A|B) = P(A ∩ B) / P(B), which leads to the multiplication rule P(A ∩ B) = P(A|B) × P(B).

概率是所有推断统计的基石。基本法则包括互斥事件的加法法则 P(A ∪ B) = P(A) + P(B)，以及一般加法法则 P(A ∪ B) = P(A) + P(B) – P(A ∩ B)。条件概率表示为 P(A|B) = P(A ∩ B) / P(B)，由此可得乘法法则 P(A ∩ B) = P(A|B) × P(B)。

Two events are independent if P(A ∩ B) = P(A) × P(B), or equivalently if P(A|B) = P(A). Tree diagrams and Venn diagrams are powerful tools for visualising combined events. When events are conditional on multiple pathways, a tree diagram helps multiply probabilities along branches and sum across mutually exclusive paths.

若 P(A ∩ B) = P(A) × P(B)，或等价地 P(A|B) = P(A)，则两事件独立。树状图和维恩图是可视化复合事件的有力工具。当事件的发生依赖多条路径时，树状图有助于沿分支相乘概率并对互斥路径求和。

Pre-U candidates should also be comfortable with the use of permutations and combinations: the number of ways of arranging r items from n is ⁿPᵣ = n!/(n–r)!, and the number of combinations is ⁿCᵣ = n!/[r!(n–r)!]. These are essential for calculating probabilities in equally likely sample spaces.

Pre-U 考生还应熟练运用排列与组合：从 n 个物品中取出 r 个进行排列的方法数为 ⁿPᵣ = n!/(n–r)!，组合数为 ⁿCᵣ = n!/[r!(n–r)!]。这对计算等可能样本空间中的概率至关重要。

3. Discrete Random Variables | 离散随机变量

A discrete random variable X takes countable values. Its probability distribution is described by the probability mass function P(X = x). The expected value E(X) = Σ x·P(X = x) represents the long-run average, while the variance Var(X) = E(X²) – [E(X)]² measures spread. For a linear transformation Y = aX + b, we have E(Y) = aE(X) + b and Var(Y) = a² Var(X).

离散随机变量 X 取可数值。其概率分布由概率质量函数 P(X = x) 描述。期望值 E(X) = Σ x·P(X = x) 代表长期平均值，而方差 Var(X) = E(X²) – [E(X)]² 衡量离散程度。对于线性变换 Y = aX + b，有 E(Y) = aE(X) + b 及 Var(Y) = a² Var(X)。

Three special discrete distributions are central to the syllabus:

Binomial distribution X ~ B(n, p): models the number of successes in n independent trials each with probability p. P(X = k) = ⁿCₖ pᵏ (1–p)ⁿ⁻ᵏ, with mean np and variance np(1–p).
Poisson distribution X ~ Po(λ): models rare events occurring randomly in a fixed interval. P(X = x) = e⁻λ λˣ / x!, with mean λ and variance λ. It can approximate the binomial when n is large and p is small (λ ≈ np).
Geometric distribution X ~ Geo(p) (two variants are used): models the number of trials up to and including the first success. The probability function P(X = x) = p(1–p)ˣ⁻¹ for x = 1, 2, …, with mean 1/p and variance (1–p)/p².

考纲中三种特殊的离散分布占据核心地位：

二项分布 X ~ B(n, p)：描述 n 次独立试验中成功次数的分布，每次成功的概率为 p。P(X = k) = ⁿCₖ pᵏ (1–p)ⁿ⁻ᵏ，均值为 np，方差为 np(1–p)。
泊松分布 X ~ Po(λ)：对固定区间内随机发生的稀有事件建模。P(X = x) = e⁻λ λˣ / x!，均值与方差均为 λ。当 n 大 p 小且 λ ≈ np 时，可用于近似二项分布。
几何分布 X ~ Geo(p)（两种定义均常使用）：描述直到首次成功（含成功那次）所经历的试验次数。概率函数为 P(X = x) = p(1–p)ˣ⁻¹，x = 1, 2, …，均值为 1/p，方差为 (1–p)/p²。

Calculations with these distributions, including cumulative probabilities from tables and the use of the Poisson as an approximation to the binomial, are routine exam requirements.

使用这些分布进行计算，包括查表求累积概率以及用泊松分布近似二项分布，是考试中的常规要求。

4. Continuous Random Variables and the Normal Distribution | 连续随机变量与正态分布

Continuous random variables are described by a probability density function (pdf) f(x), where probabilities correspond to areas under the curve. The total area under f(x) is 1, and P(a ≤ X ≤ b) = ∫ₐᵇ f(x) dx. The cumulative distribution function F(x) = P(X ≤ x) gives the area to the left of x.

连续随机变量由概率密度函数 f(x) 描述，概率对应曲线下的面积。f(x) 下的总面积为 1，且 P(a ≤ X ≤ b) = ∫ₐᵇ f(x) dx。累积分布函数 F(x) = P(X ≤ x) 给出 x 左侧的面积。

The normal distribution is the most important continuous model. If X ~ N(μ, σ²), then its pdf is symmetric and bell-shaped. Standardising to Z = (X – μ)/σ yields Z ~ N(0, 1), allowing the use of standard normal tables. Calculations often involve finding probabilities P(X > a), P(X < b) or symmetric intervals. The inverse normal method is used to find the value x corresponding to a given cumulative probability.

正态分布是最重要的连续模型。若 X ~ N(μ, σ²)，其概率密度函数呈对称的钟形。标准化为 Z = (X – μ)/σ 后，Z ~ N(0, 1)，便可使用标准正态分布表。计算常涉及求 P(X > a)、P(X < b) 或对称区间。逆正态方法用于寻找与给定累积概率对应的 x 值。

The normal distribution can also serve as an approximation to the binomial (when np and n(1–p) are large) and to the Poisson (when λ is large), applying a continuity correction for discrete data. Understanding when and how to apply these approximations is essential for Pre-U Statistics.

正态分布还可用于近似二项分布（当 np 和 n(1–p) 较大时）和泊松分布（当 λ 较大时），对离散数据需使用连续性校正。理解何时以及如何应用这些近似，对 Pre-U 统计至关重要。

5. Sampling and Estimation | 抽样与估计

Statistical inference relies on understanding sampling distributions. The sample mean X̄ from a population with mean μ and variance σ² has E(X̄) = μ and Var(X̄) = σ²/n. The Central Limit Theorem states that for sufficiently large samples, the distribution of X̄ is approximately normal regardless of the population distribution. This justifies the use of normal-based confidence intervals.

统计推断依赖于对抽样分布的理解。来自均值为 μ、方差为 σ² 的总体的样本均值 X̄ 满足 E(X̄) = μ 且 Var(X̄) = σ²/n。中心极限定理指出，对于足够大的样本，无论总体分布如何，X̄ 的分布都近似正态。这为正态置信区间的使用提供了依据。

A confidence interval provides a range of plausible values for an unknown parameter. For a population mean with known variance, the 95% confidence interval is x̄ ± 1.96 × σ/√n. When σ is unknown, the sample standard deviation s is used and the multiplier becomes a t-value from the t-distribution with n–1 degrees of freedom. For a population proportion p, the interval is p̂ ± z × √[p̂(1–p̂)/n], where p̂ is the sample proportion.

置信区间为未知参数提供了一组合理的取值范围。对于已知方差的总体均值，95% 置信区间为 x̄ ± 1.96 × σ/√n。当 σ 未知时，使用样本标准差 s，乘数变为自由度为 n–1 的 t 分布中的 t 值。对于总体比例 p，置信区间为 p̂ ± z × √[p̂(1–p̂)/n]，其中 p̂ 为样本比例。

Choosing the appropriate sample size to achieve a desired margin of error is also a standard skill: solving the margin of error equation for n. Pre-U questions may also explore the difference between two means or two proportions.

选择适当的样本量以达到期望的误差范围也是标准技能：通过求解误差范围方程得到 n。Pre-U 试题还可能涉及两均值之差或两比例之差的推断。

6. Hypothesis Testing | 假设检验

Hypothesis testing is a formal procedure for making decisions about population parameters. The null hypothesis H₀ usually states ‘no effect’ or ‘no difference’, while the alternative hypothesis H₁ is what the investigator suspects. A test statistic is calculated from the sample, and its value is compared to a critical region defined by the significance level α (commonly 5%).

假设检验是对总体参数做出决策的正式过程。原假设 H₀ 通常陈述“无效应”或“无差异”，而备择假设 H₁ 则是研究者的猜想。由样本计算检验统计量，并将其值与由显著性水平 α（通常为 5%）定义的临界区域进行比较。

The p-value is the probability of obtaining a test statistic at least as extreme as the one observed, assuming H₀ is true. If the p-value is less than α, H₀ is rejected. Pre-U candidates must perform tests for:

Binomial probability p: using exact binomial probabilities to find the p-value or critical regions.
Poisson mean λ: similarly using Poisson probabilities.
Population mean (normal, known variance): z-test, test statistic Z = (x̄ – μ₀)/(σ/√n).
Population mean (normal, unknown variance): t-test, test statistic t = (x̄ – μ₀)/(s/√n) with n–1 degrees of freedom.

p 值是在 H₀ 为真的条件下，获得至少如观测值那样极端的检验统计量的概率。若 p 值小于 α，则拒绝 H₀。Pre-U 考生必须掌握以下检验：

二项概率 p：使用精确二项概率求 p 值或临界区域。
泊松均值 λ：类似地使用泊松概率。
总体均值（正态，已知方差）：z 检验，检验统计量 Z = (x̄ – μ₀)/(σ/√n)。
总体均值（正态，未知方差）：t 检验，检验统计量 t = (x̄ – μ₀)/(s/√n)，自由度为 n–1。

One-tailed and two-tailed tests are distinguished by the form of H₁. Candidates should state conclusions clearly in the context of the problem, avoiding absolute acceptance of H₀. Power, Type I and Type II errors may also be examined.

单尾与双尾检验根据 H₁ 的形式加以区分。考生应在问题情境下清晰地陈述结论，避免绝对地接受 H₀。检验功效、第 I 类与第 II 类错误也可能出现在考题中。

7. Correlation and Linear Regression | 相关与线性回归

Correlation measures the strength and direction of a linear relationship between two variables. The product-moment correlation coefficient r is given by r = S_{xy} / √(S_{xx} S_{yy}), where S_{xy} = Σ(x – x̄)(y – ȳ). The value of r lies between –1 and 1. A value close to 1 or –1 indicates a strong linear association, while values near 0 suggest little or no linear correlation.

相关度量了两个变量之间线性关系的强度与方向。积矩相关系数 r 由 r = S_{xy} / √(S_{xx} S_{yy}) 给出，其中 S_{xy} = Σ(x – x̄)(y – ȳ)。r 的值介于 –1 和 1 之间。接近 1 或 –1 的值表明存在强线性关联，而接近 0 的值则表明几乎没有线性相关。

Spearman’s rank correlation coefficient rₛ is a non-parametric alternative, calculated using the differences dᵢ between the ranks of corresponding x and y values: rₛ = 1 – (6 Σdᵢ²) / [n(n²–1)]. It assesses monotonic, not necessarily linear, relationships and is robust to outliers.

斯皮尔曼等级相关系数 rₛ 是一种非参数替代方法，利用对应 x 值和 y 值排名的差值 dᵢ 计算：rₛ = 1 – (6 Σdᵢ²) / [n(n²–1)]。它评估单调关系（不限于线性），且对异常值具有稳健性。

Linear regression models the relationship with the least squares line y = a + bx. The slope is b = S_{xy} / S_{xx} and the intercept is a = ȳ – bx̄. The regression equation can be used for prediction within the range of the data, but extrapolation must be treated with caution. Residuals, variations about the line, and the coefficient of determination R² = r² (in simple regression) are often discussed to assess model fit.

线性回归用最小二乘直线 y = a + bx 对关系建模。斜率为 b = S_{xy} / S_{xx}，截距为 a = ȳ – bx̄。回归方程可用于数据范围内的预测，但外推需谨慎。残差、直线附近的变异性以及决定系数 R² = r²（在简单回归中）常用于评估模型的拟合程度。

8. Chi-Squared Tests | 卡方检验

Chi-squared (χ²) tests are used for categorical data. The test statistic is X² = Σ [(O – E)² / E], where O and E are observed and expected frequencies. Under the null hypothesis, the statistic approximately follows a χ² distribution with an appropriate number of degrees of freedom, provided that all expected frequencies are sufficiently large (typically ≥ 5).

卡方（χ²）检验用于分类数据。检验统计量为 X² = Σ [(O – E)² / E]，其中 O 和 E 分别为观测频数和期望频数。在原假设下，只要所有期望频数足够大（通常 ≥ 5），该统计量近似服从自由度适当的 χ² 分布。

Two main applications are required: goodness-of-fit tests and tests for independence in contingency tables. In a goodness-of-fit test, the null hypothesis specifies a theoretical distribution, and degrees of freedom equal (number of categories – 1 – number of estimated parameters). For an r × c contingency table testing independence, the degrees of freedom are (r–1)(c–1).

需要掌握两种主要应用：拟合优度检验和列联表中的独立性检验。在拟合优度检验中，原假设指定了一个理论分布，自由度等于（类别数 – 1 – 估计的参数个数）。对于 r × c 列联表的独立性检验，自由度为 (r–1)(c–1)。

Candidates should be able to calculate expected frequencies under the null, combine categories if necessary, and interpret p-values or critical values from χ² tables. Yates’s correction is sometimes applied for 2 × 2 tables, though the Pre-U specification may focus on the standard test with due attention to expected cell counts.

考生应能计算原假设下的期望频数，必要时合并类别，并利用 χ² 分布表解释 p 值或临界值。虽然有时会对 2 × 2 表使用耶茨校正，但 Pre-U 考纲可能侧重于标准检验并关注期望格值的大小。

Published by TutorHao | Statistics Revision Series | aleveler.com

更多咨询请联系16621398022（同微信）

Core Knowledge Points for Pre-U CAIE Statistics | Pre-U CAIE 统计：核心知识点梳理