📚 Pre-U OCR Statistics: Quick Reference Handbook of Formulas and Theorems | Pre-U OCR 统计:公式定理速查手册
This revision guide provides a concise overview of the essential formulas and theorems required for the OCR Pre-U Statistics specification. It is designed for quick reference and revision, covering probability, distributions, inference, and association. Each section presents the core results in a clear, accessible format, with paired English and Chinese explanations to support bilingual learners.
本速查手册概括了 OCR Pre-U 统计课程的核心公式与定理,适用于考前快速复习。内容涵盖概率、概率分布、统计推断以及相关与回归分析。每一个主题都以清晰的中英双语对照呈现,帮助双语学习者准确掌握关键考点。
1. Probability Basics | 概率基础
The addition rule for any two events A and B is:
P(A ∪ B) = P(A) + P(B) − P(A ∩ B)
对于任意两个事件 A 和 B,加法法则为:P(A ∪ B) = P(A) + P(B) − P(A ∩ B)。若 A 与 B 互斥,则 P(A ∩ B) = 0。
Conditional probability is defined as:
P(A | B) = P(A ∩ B) / P(B), P(B) > 0
条件概率定义为:P(A | B) = P(A ∩ B) / P(B),其中 P(B) > 0。利用条件概率可得乘法法则:P(A ∩ B) = P(A | B)·P(B) = P(B | A)·P(A)。
Events A and B are independent if and only if:
P(A ∩ B) = P(A)·P(B) or P(A | B) = P(A)
事件 A 与 B 相互独立的充要条件为:P(A ∩ B) = P(A)·P(B) 或 P(A | B) = P(A)。
Bayes’ theorem provides a way to reverse conditional probabilities:
P(A | B) = [P(B | A)·P(A)] / P(B)
贝叶斯定理可用来由 P(B | A) 求 P(A | B):P(A | B) = [P(B | A)·P(A)] / P(B)。
2. Expectation and Variance | 期望与方差
For a discrete random variable X with possible values xᵢ and probabilities pᵢ:
E(X) = μ = Σ xᵢ pᵢ
对于离散随机变量 X,其期望值(均值)为 E(X) = μ = Σ xᵢ pᵢ。
The variance and standard deviation are:
Var(X) = σ² = E[(X − μ)²] = Σ (xᵢ − μ)² pᵢ = E(X²) − [E(X)]²
方差与标准差:Var(X) = σ² = E[(X − μ)²] = Σ (xᵢ − μ)² pᵢ。常用计算式:Var(X) = E(X²) − [E(X)]²。
For any constants a and b:
E(aX + b) = aE(X) + b, Var(aX + b) = a² Var(X)
对于常数 a 与 b,E(aX + b) = aE(X) + b,且 Var(aX + b) = a² Var(X)。
The covariance between two random variables X and Y is:
Cov(X, Y) = E[(X − μₓ)(Y − μᵧ)] = E(XY) − μₓ μᵧ
随机变量 X 与 Y 的协方差为 Cov(X, Y) = E[(X − μₓ)(Y − μᵧ)] = E(XY) − μₓ μᵧ。若 X、Y 独立,则 Cov(X, Y) = 0。
3. Discrete Distributions: Binomial | 离散分布:二项分布
A binomial distribution models the number of successes in n independent trials, each with success probability p. Notation: X ~ B(n, p).
二项分布用于描述 n 次独立试验中成功的次数,每次成功概率为 p,记作 X ~ B(n, p)。
Probability mass function:
P(X = r) = ⁿCᵨ pʳ (1 − p)ⁿ⁻ʳ, r = 0, 1, …, n
概率质量函数:P(X = r) = ⁿCᵨ pʳ (1 − p)ⁿ⁻ʳ,其中 ⁿCᵨ = n! / [r!(n − r)!]。
Mean and variance:
E(X) = np, Var(X) = np(1 − p)
期望与方差:E(X) = np,Var(X) = np(1 − p)。
Conditions: fixed number of trials, independent trials, constant probability p, two outcomes per trial.
使用条件:试验次数固定,各次试验独立,每次成功概率 p 恒定,每次试验只有两种结果。
4. Discrete Distributions: Poisson | 离散分布:泊松分布
The Poisson distribution models the number of events occurring in a fixed interval of time or space. Notation: X ~ Po(λ), where λ is the mean rate.
泊松分布用于描述固定时间或空间内随机事件发生的次数,记作 X ~ Po(λ),参数 λ 为平均发生率。
Probability mass function:
P(X = r) = (e⁻λ λʳ) / r!, r = 0, 1, 2, …
概率质量函数:P(X = r) = e⁻λ λʳ / r!。
Mean and variance:
E(X) = λ, Var(X) = λ
期望与方差:E(X) = λ,Var(X) = λ。
Approximation: For large n and small p with np = λ, B(n, p) can be approximated by Po(λ). Usually valid when n is large, p < 0.1.
近似条件:当 n 很大、p 很小且 np = λ 时,可用 Po(λ) 近似二项分布 B(n, p),通常需 n 大、p < 0.1。
5. Continuous Distributions: Normal | 连续分布:正态分布
The normal distribution is defined by parameters μ (mean) and σ² (variance). Notation: X ~ N(μ, σ²).
正态分布由均值 μ 和方差 σ² 决定,记作 X ~ N(μ, σ²)。
Probability density function:
f(x) = [1 / (σ√(2π))] × e^{−½[(x − μ)/σ]²}, −∞ < x < ∞
概率密度函数为:f(x) = [1 / (σ√(2π))] × e^{−½[(x − μ)/σ]²},定义在整个实数域。
Standardisation: Z = (X − μ) / σ ~ N(0, 1). The standard normal distribution has mean 0, variance 1.
标准化:令 Z = (X − μ) / σ,则 Z ~ N(0, 1),称为标准正态分布。
Key probabilities: about 68% of data within μ ± σ, about 95% within μ ± 2σ, about 99.7% within μ ± 3σ.
重要概率:大约 68% 的数据落在 μ ± σ 内,约 95% 落在 μ ± 2σ 内,约 99.7% 落在 μ ± 3σ 内。
Normal approximation to binomial: For n large, X ~ B(n, p) approximately N(np, np(1 − p)), provided np ≥ 5 and n(1 − p) ≥ 5. Use continuity correction.
二项分布的正态近似:当 n 很大且满足 np ≥ 5 和 n(1 − p) ≥ 5 时,X ~ B(n, p) 近似服从 N(np, np(1 − p));需使用连续性校正。
6. Sampling Distributions | 抽样分布
The sample mean X̅ from a population with mean μ and variance σ² has sampling distribution:
E(X̅) = μ, Var(X̅) = σ² / n, SE = σ / √n
从均值为 μ、方差为 σ² 的总体中抽取大小为 n 的样本,样本均值 X̅ 的抽样分布满足:E(X̅) = μ,Var(X̅) = σ² / n,标准误为 σ / √n。
Central Limit Theorem: For large n (typically n ≥ 30), X̅ is approximately N(μ, σ²/n), regardless of the population distribution. If the population is normal, the result holds for any n.
中心极限定理:对于大样本(通常 n ≥ 30),不论总体分布如何,X̅ 近似服从 N(μ, σ²/n);若总体本身为正态分布,则该结论对任意 n 均成立。
For the sample proportion P̂ = X / n from a binomial population with parameters n and p:
E(P̂) = p, Var(P̂) = p(1 − p) / n
样本比例 P̂ = X / n(来自二项总体)的抽样分布:E(P̂) = p,Var(P̂) = p(1 − p) / n。大样本下,P̂ 近似正态分布 N(p, p(1 − p)/n)。
7. Confidence Intervals | 置信区间
A 100(1 − α)% confidence interval for the population mean μ when σ is known:
X̅ ± z_{α/2} × (σ / √n)
当总体标准差 σ 已知时,均值 μ 的 100(1 − α)% 置信区间为:X̅ ± z_{α/2} × (σ / √n)。
When σ is unknown and the population is normal, use the t-distribution with n − 1 degrees of freedom:
X̅ ± t_{α/2, n−1} × (s / √n)
若 σ 未知且总体服从正态分布,则使用自由度为 n − 1 的 t 分布:X̅ ± t_{α/2, n−1} × (s / √n),其中 s 为样本标准差。
For a population proportion p (large sample):
P̂ ± z_{α/2} × √[P̂(1 − P̂) / n]
总体比例 p 的大样本置信区间:P̂ ± z_{α/2} × √[P̂(1 − P̂) / n]。
Common z-values: z_{0.05} = 1.645 (one-tailed), z_{0.025} = 1.96 (two-tailed 95% CI).
常用 z 值:单侧 α = 0.05 时 z = 1.645;双侧 95% 置信区间用 z = 1.96。
8. Hypothesis Testing for Means | 均值假设检验
Testing for a population mean μ: H₀: μ = μ₀. Test statistic when σ known:
Z = (X̅ − μ₀) / (σ / √n)
检验总体均值 μ = μ₀:若 σ 已知,检验统计量为 Z = (X̅ − μ₀) / (σ / √n)。
When σ unknown, use the t-statistic:
t = (X̅ − μ₀) / (s / √n) with df = n − 1
若 σ 未知,使用 t 统计量:t = (X̅ − μ₀) / (s / √n),自由度 df = n − 1。
Paired t-test for two means (before/after): compute differences dᵢ and test H₀: μ_d = 0 using
t = d̅ / (s_d / √n) with df = n − 1
配对 t 检验(适用于配对的均值比较):计算差值 dᵢ,检验 H₀: μ_d = 0,统计量 t = d̅ / (s_d / √n),自由度为 n − 1。
Two-sample test for independent means (assuming equal variances):
t = (X̅₁ − X̅₂) / [sₚ √(1/n₁ + 1/n₂)], sₚ² = [(n₁−1)s₁² + (n₂−1)s₂²] / (n₁ + n₂ − 2)
独立双样本均值检验(假定方差相等):t = (X̅₁ − X̅₂) / [sₚ √(1/n₁ + 1/n₂)],其中合并方差 sₚ² = [(n₁−1)s₁² + (n₂−1)s₂²] / (n₁ + n₂ − 2),自由度 n₁ + n₂ − 2。
Decision: Compare test statistic to critical value or use p-value. Reject H₀ if |statistic| > critical value or p < α.
决策准则:比较检验统计量与临界值,或采用 p 值法。若 |统计量| > 临界值 或 p < α,则拒绝原假设 H₀。
9. Chi-Squared Tests | 卡方检验
Chi-squared test for goodness of fit:
χ² = Σ [(Oᵢ − Eᵢ)² / Eᵢ]
卡方拟合优度检验:χ² = Σ [(Oᵢ − Eᵢ)² / Eᵢ],其中 Oᵢ 为观测频数,Eᵢ 为期望频数。自由度 = 类别数 − 1 − 估计参数个数。
Test of independence in a contingency table (r × c):
χ² = Σ Σ [(Oᵢⱼ − Eᵢⱼ)² / Eᵢⱼ], Eᵢⱼ = (row total × column total) / grand total
列联表独立性检验(r × c 表):χ² = Σ Σ [(Oᵢⱼ − Eᵢⱼ)² / Eᵢⱼ],期望频数 Eᵢⱼ = (行合计 × 列合计) / 总合计。自由度 = (r − 1)(c − 1)。
Conditions: all expected frequencies should be ≥ 5 (or no more than 20% below 5). If necessary, combine categories.
使用条件:所有期望频数应 ≥ 5,或低于 5 的比例不超过 20%;必要时可合并相邻类别。
10. Correlation and Regression | 相关与回归
Pearson product-moment correlation coefficient:
r = S_{xy} / √(S_{xx} S_{yy})
皮尔逊积差相关系数:r = S_{xy} / √(S_{xx} S_{yy}),其中 S_{xx} = Σ(xᵢ − x̅)²,S_{yy} = Σ(yᵢ − y̅)²,S_{xy} = Σ(xᵢ − x̅)(yᵢ − y̅)。
Spearman’s rank correlation coefficient ρ:
ρ = 1 − [6 Σ dᵢ²] / [n(n² − 1)]
斯皮尔曼等级相关系数 ρ = 1 − [6 Σ dᵢ²] / [n(n² − 1)],其中 dᵢ 为每对观测值的秩次差。
Least squares regression line y on x: y = a + bx, where
b = S_{xy} / S_{xx}, a = y̅ − b x̅
最小二乘回归直线(y 对 x):y = a + bx,其中斜率 b = S_{xy} / S_{xx},截距 a = y̅ − b x̅。
The residual sum of squares is used to estimate the error variance: s² = Σ (yᵢ − a − b xᵢ)² / (n − 2).
残差平方和用于估计误差方差:s² = Σ (yᵢ − a − b xᵢ)² / (n − 2)。
Hypothesis test for ρ = 0 (bivariate normal distribution) uses the test statistic:
t = r √(n − 2) / √(1 − r²) with df = n − 2
检验相关系数是否为 0(二元正态总体)可用 t 统计量:t = r √(n − 2) / √(1 − r²),自由度 n − 2。
11. Non-Parametric Tests | 非参数检验
The Wilcoxon signed-rank test (one sample or paired) may be used when the population is not normal. Compute differences dᵢ, rank |dᵢ|, and sum the ranks of positive differences (W⁺). For large n,
Z = [W⁺ − n(n+1)/4] / √[n(n+1)(2n+1) / 24]
威尔科克森符号秩检验(单样本或配对样本)适用于非正态总体。计算差值 dᵢ,对 |dᵢ| 排序,将正差值的秩求和得 W⁺。大样本下近似正态:Z = [W⁺ − n(n+1)/4] / √[n(n+1)(2n+1) / 24]。
The Mann–Whitney U test compares two independent samples. Rank all data together, sum the ranks for sample 1 (R₁). The U statistic is:
U = n₁n₂ + n₁(n₁+1)/2 − R₁
曼-惠特尼 U 检验用于比较两个独立样本。合并全部数据并排序,将样本 1 的秩和记为 R₁,U = n₁n₂ + n₁(n₁+1)/2 − R₁。较小样本下查表,大样本可使用正态近似。
These non-parametric tests replace the t‑test when normality is in doubt or data are ordinal.
当正态性假设不成立或数据为顺序变量时,可用上述非参数检验替代 t 检验。
Published by TutorHao | Statistics 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