📚 Core Topics in Pre-U OCR Statistics | Pre-U OCR 统计：核心知识点梳理
Pre-U Statistics under OCR is a demanding yet fascinating subject that bridges pure mathematics with real-world data analysis. A firm grasp of its foundational concepts — from probability and random variables to inference and regression — is essential for success. This revision guide walks you through the ten pivotal topics, pairing clear English explanations with Chinese translations to support bilingual learners. Let’s build your confidence in statistical reasoning and problem-solving.
OCR Pre-U 统计是一门要求严格又引人入胜的学科，它连接了纯数学与现实数据分析。扎实掌握其基础概念——从概率和随机变量到推断和回归——对于取得成功至关重要。本复习指南带你梳理十个关键主题，清晰的英文解释与中文翻译配对，以支持双语学习者。让我们共同建立统计推理和解题的信心。
1. Probability Fundamentals and Bayes’ Theorem | 概率基础与贝叶斯定理
Probability theory forms the backbone of statistics. The sample space, events, and the axioms of probability (non-negativity, unit measure, additivity) provide a rigorous starting point. Conditional probability P(A|B) = P(A ∩ B)/P(B) is central, leading naturally to the law of total probability and Bayes’ theorem: P(A|B) = [P(B|A) P(A)] / P(B). These tools allow us to update beliefs in light of new evidence, a cornerstone of statistical inference.
概率论是统计学的支柱。样本空间、事件以及概率公理（非负性、规范性、可加性）提供了一个严格的起点。条件概率 P(A|B) = P(A ∩ B)/P(B) 是核心，自然引出全概率公式和贝叶斯定理：P(A|B) = [P(B|A) P(A)] / P(B)。这些工具使我们能够根据新证据更新信念，这是统计推断的基石。
Two events A and B are independent if and only if P(A ∩ B) = P(A)P(B). Mutual independence extends this idea to more than two events. Understanding independence is crucial when setting up probability models and later when testing for association in contingency tables.
两个事件 A 和 B 独立当且仅当 P(A ∩ B) = P(A)P(B)。相互独立将这一思想推广到多于两个事件。在建立概率模型以及后续检验列联表中的关联时，理解独立性至关重要。
Tree diagrams and Venn diagrams often help visualise compound events, especially when dealing with sequential experiments like drawing without replacement. Always define your events clearly and check that probabilities sum to one within the correct partitions.
树状图和维恩图有助于直观展示复合事件，尤其在处理不放回抽样等序贯试验时。始终明确定义事件，并确保在正确的划分内概率之和为 1。
2. Discrete Random Variables and Expectation | 离散随机变量与期望
A discrete random variable X takes countable values with a probability mass function (PMF) p(x) = P(X = x). The expectation E(X) = Σ x p(x) measures the centre of the distribution, while variance Var(X) = E[(X – μ)²] = E(X²) – [E(X)]² quantifies spread. The standard deviation σ = √Var(X) returns to the original units.
离散随机变量 X 以概率质量函数 (PMF) p(x) = P(X = x) 取可数值。期望 E(X) = Σ x p(x) 衡量分布的中心，而方差 Var(X) = E[(X – μ)²] = E(X²) – [E(X)]² 量化离散程度。标准差 σ = √Var(X) 恢复到原单位。
For any function g(X), the expected value is E[g(X)] = Σ g(x) p(x). A linear transformation Y = aX + b has mean aE(X) + b and variance a² Var(X). These properties simplify working with coded data and scaling.
对于任意函数 g(X)，期望值为 E[g(X)] = Σ g(x) p(x)。线性变换 Y = aX + b 的均值为 aE(X) + b，方差为 a² Var(X)。这些性质简化了编码数据和尺度变换的处理。
The cumulative distribution function (CDF) F(x) = P(X ≤ x) accumulates probability up to x. For discrete variables, F(x) is a step function. Knowing how to move between PMF and CDF is essential, particularly for finding medians and percentiles.
累积分布函数 (CDF) F(x) = P(X ≤ x) 累积到 x 为止的概率。对于离散变量，F(x) 是阶梯函数。掌握如何在 PMF 和 CDF 之间转换非常重要，尤其在求中位数和百分位数时。
3. Binomial and Poisson Distributions | 二项分布与泊松分布
The binomial distribution models the number of successes in n independent Bernoulli trials, each with success probability p. Its PMF is P(X = r) = ⁿCᵣ pʳ (1 – p)ⁿ⁻ʳ, where r = 0,1,…,n. Mean μ = np, variance σ² = np(1 – p). The distribution is symmetric when p = 0.5, skewed otherwise. Conditions: fixed n, constant p, independence.
二项分布模拟 n 次独立伯努利试验中成功的次数，每次成功概率为 p。其 PMF 为 P(X = r) = ⁿCᵣ pʳ (1 – p)ⁿ⁻ʳ，其中 r = 0,1,…,n。均值 μ = np，方差 σ² = np(1 – p)。p = 0.5 时分布对称，否则偏斜。条件：固定的 n，恒定的 p，独立性。
The Poisson distribution arises as the limit of a binomial when n is large and p is small, giving a constant rate λ = np. PMF: P(X = r) = e⁻λ λʳ / r! for r = 0,1,2,… . Both mean and variance equal λ. It models rare events in a fixed interval of time or space, such as typing errors per page or calls per hour.
泊松分布是二项分布在 n 大 p 小时的极限情况，得到恒定速率 λ = np。PMF：P(X = r) = e⁻λ λʳ / r!，r = 0,1,2,…。均值和方差都等于 λ。它模拟固定时间或空间间隔内的稀有事件，如每页打字错误或每小时呼入电话。
The Poisson distribution can approximate a binomial when n > 50 and np < 5, using λ = np. Conversely, a normal approximation to the Poisson is valid when λ > 15. Continuity corrections improve accuracy when moving from a discrete to a continuous approximating distribution.
当 n > 50 且 np < 5 时，泊松分布可用 λ = np 逼近二项分布。相反，当 λ > 15 时，可用正态分布逼近泊松分布。从离散分布过渡到连续逼近分布时，连续性校正可提高准确性。
4. Continuous Random Variables and Probability Density Functions | 连续随机变量与概率密度函数
A continuous random variable X is described by its probability density function (PDF) f(x), where f(x) ≥ 0 and the total area under the curve is 1. Probabilities are found by integration: P(a ≤ X ≤ b) = ∫ₐᵇ f(x) dx. The cumulative distribution function F(x) = P(X ≤ x) = ∫_{-∞}ˣ f(t) dt.
连续随机变量 X 由其概率密度函数 (PDF) f(x) 描述，满足 f(x) ≥ 0，且曲线下总面积为 1。概率通过积分求得：P(a ≤ X ≤ b) = ∫ₐᵇ f(x) dx。累积分布函数 F(x) = P(X ≤ x) = ∫_{-∞}ˣ f(t) dt。
Expectation and variance are defined analogously: E(X) = ∫ x f(x) dx over the support, and Var(X) = ∫ (x – μ)² f(x) dx = E(X²) – μ². The median m satisfies F(m) = 0.5; the mode is the maximum of f(x). Percentiles can be obtained from the inverse CDF.
期望和方差的定义类似：E(X) = ∫ x f(x) dx （在支撑集上），Var(X) = ∫ (x – μ)² f(x) dx = E(X²) – μ²。中位数 m 满足 F(m) = 0.5；众数是 f(x) 的最大值点。百分位数可从逆 CDF 求得。
Common continuous distributions include the uniform, exponential, and normal. The uniform distribution on [a, b] has PDF f(x) = 1/(b – a). The exponential distribution with parameter λ has PDF f(x) = λ e^{-λx} and models waiting times; its mean is 1/λ and variance 1/λ².
常见的连续分布包括均匀分布、指数分布和正态分布。在 [a, b] 上的均匀分布 PDF 为 f(x) = 1/(b – a)。参数为 λ 的指数分布 PDF 为 f(x) = λ e^{-λx}，用于建模等待时间；其均值为 1/λ，方差为 1/λ²。
5. The Normal Distribution | 正态分布
The normal distribution N(μ, σ²) is the most important continuous distribution in statistics. Its PDF is symmetric, bell-shaped, centred at μ, with points of inflection at μ ± σ. The standard normal Z ~ N(0, 1) is obtained by standardising: Z = (X – μ)/σ. Tables give Φ(z) = P(Z ≤ z), which can be used to find probabilities for any normal variable.
正态分布 N(μ, σ²) 是统计学中最重要的连续分布。其 PDF 对称、钟形，以 μ 为中心，拐点在 μ ± σ。通过标准化 Z = (X – μ)/σ 得到标准正态分布 Z ~ N(0, 1)。标准正态表给出 Φ(z) = P(Z ≤ z)，可用于求任意正态变量的概率。
Key percentages: about 68% of values lie within μ ± σ, 95% within μ ± 2σ, and 99.7% within μ ± 3σ. These empirical rules are useful for quick checks. The normal distribution often arises as an approximation through the Central Limit Theorem, which underpins much of inferential statistics.
关键百分比：约 68% 的值落在 μ ± σ 内，95% 在 μ ± 2σ 内，99.7% 在 μ ± 3σ 内。这些经验法则便于快速检查。正态分布常通过中心极限定理作为逼近出现，这为推断统计的大部分内容奠定了基础。
When using the normal as an approximation to the binomial (np>5, nq>5) or Poisson (λ>15), apply a continuity correction: replace a discrete x with the interval (x-0.5, x+0.5). For example, P(X ≥ 10) becomes P(X > 9.5) under the normal curve.
用正态分布逼近二项分布（np>5, nq>5）或泊松分布（λ>15）时，应用连续性校正：将离散点 x 替换为区间 (x-0.5, x+0.5)。例如，P(X ≥ 10) 在正态曲线下变为 P(X > 9.5)。
6. Sampling and Sampling Distributions | 抽样与抽样分布
Statistical inference relies on samples drawn from a population. A simple random sample of size n ensures every individual has an equal chance of selection, making the sample representative. The sample mean X̄ = ΣXᵢ/n is a random variable whose distribution is the sampling distribution of the mean.
统计推断依赖于从总体中抽取的样本。大小为 n 的简单随机样本确保每个个体有相等的选中机会，使样本具有代表性。样本均值 X̄ = ΣXᵢ/n 是一个随机变量，其分布即是均值的抽样分布。
If X ~ N(μ, σ²), then X̄ ~ N(μ, σ²/n). Even if the population is not normal, the Central Limit Theorem (CLT) states that for sufficiently large n (usually n ≥ 30), X̄ is approximately normal with mean μ and variance σ²/n. This remarkable result justifies many parametric tests even when data are not perfectly normal.
若 X ~ N(μ, σ²)，则 X̄ ~ N(μ, σ²/n)。即使总体不服从正态分布，中心极限定理 (CLT) 指出，当 n 足够大（通常 n ≥ 30）时，X̄ 近似服从均值为 μ、方差为 σ²/n 的正态分布。这一非凡结果使得即使数据不完全正态，许多参数检验依然成立。
The standard error of the mean is σ/√n, which decreases as sample size increases. When σ is unknown, we estimate it with the sample standard deviation s, leading to the t-distribution: (X̄ – μ)/(s/√n) ~ tₙ₋₁. The degrees of freedom ν = n – 1 reflect the penalty for estimating σ.
均值的标准误差为 σ/√n，随着样本量增大而减小。当 σ 未知时，我们用样本标准差 s 估计，从而得到 t 分布：(X̄ – μ)/(s/√n) ~ tₙ₋₁。自由度 ν = n – 1 反映了估计 σ 的代价。
7. Point and Interval Estimation | 点估计与区间估计
A point estimator is a statistic used to estimate a population parameter. Unbiasedness means E(θ̂) = θ. Efficiency refers to having a small variance. The sample mean X̄ is an unbiased estimator of μ, while the sample variance s² = Σ(Xᵢ – X̄)²/(n – 1) is unbiased for σ² (using n – 1 to correct bias).
点估计量是用于估计总体参数的统计量。无偏性意味着 E(θ̂) = θ。有效性指方差较小。样本均值 X̄ 是 μ 的无偏估计量，而样本方差 s² = Σ(Xᵢ – X̄)²/(n – 1) 是 σ² 的无偏估计量（使用 n – 1 校正偏差）。
A confidence interval provides a range of plausible values for a parameter with a specified confidence level, usually 95%. For a normal mean with known σ, the interval is X̄ ± z* × σ/√n, where z* = 1.96 for 95% confidence. With σ unknown, use the t-distribution: X̄ ± t*ₙ₋₁ × s/√n.
置信区间给出参数的可能取值范围，并带有指定的置信水平，通常为 95%。对于已知 σ 的正态均值，区间为 X̄ ± z* × σ/√n，其中 z* = 1.96（95% 置信度）。σ 未知时，使用 t 分布：X̄ ± t*ₙ₋₁ × s/√n。
The width of a confidence interval depends on sample size, variability, and confidence level. Increasing n narrows the interval, improving precision. The interpretation: if we repeatedly sample, 95% of intervals constructed in this way will contain the true μ. It does not mean a 95% probability that the specific interval contains μ (a common misconception).
置信区间的宽度取决于样本量、变异性和置信水平。增大 n 可缩窄区间，提高精度。解释：若重复抽样，以这种方式构造的区间中 95% 会包含真实的 μ。这不意味着特定区间有 95% 的概率包含 μ（常见误解）。
8. Principles of Hypothesis Testing | 假设检验原理
A hypothesis test evaluates evidence against a null hypothesis H₀ in favour of an alternative H₁. The p-value is the probability of observing a test statistic at least as extreme as the one obtained, assuming H₀ is true. A small p-value (typically < 0.05) leads to rejection of H₀ in favour of H₁.
假设检验评估反对零假设 H₀、支持备择假设 H₁ 的证据。p 值是假定 H₀ 为真时，观察到检验统计量至少与实际所得一样极端的概率。小的 p 值（通常 < 0.05）导致拒绝 H₀ 而接受 H₁。
The significance level α is the threshold for rejection (e.g., 0.05). A Type I error occurs when H₀ is true but rejected (probability = α). A Type II error occurs when H₀ is false but not rejected; its probability is β. Power = 1 – β is the probability of correctly rejecting a false H₀. Balancing errors depending on context is crucial.
显著性水平 α 是拒绝的阈值（例如 0.05）。当 H₀ 为真但被拒绝时发生第 I 类错误（概率为 α）。当 H₀ 为假但未被拒绝时发生第 II 类错误，其概率为 β。功效 = 1 – β 是正确拒绝一个假的 H₀ 的概率。根据情境平衡两类错误至关重要。
One-sample z-test for mean (σ known) uses Z = (X̄ – μ₀) / (σ/√n). One-sample t-test (σ unknown) uses tₙ₋₁. Paired t-test compares two related samples by analysing differences. Two-sample t-test compares independent group means, assuming equal or unequal variances. Always check assumptions: normality, independence, and (for two-sample) homogeneity of variance.
单样本 z 检验（σ 已知）使用 Z = (X̄ – μ₀) / (σ/√n)。单样本 t 检验（σ 未知）使用 tₙ₋₁。配对 t 检验通过分析差值比较两个相关样本。双样本 t 检验比较独立组的均值，假定方差相等或不等。始终检查假设：正态性、独立性，以及（对于双样本）方差齐性。
9. Correlation and Linear Regression | 相关性与线性回归
Pearson’s product-moment correlation coefficient r measures the strength and direction of a linear relationship between two variables. r lies in [-1, 1], with r = ±1 indicating perfect linear correlation. The formula: r = Σ (xᵢ – x̄)(yᵢ – ȳ) / √[Σ (xᵢ – x̄)² Σ (yᵢ – ȳ)²]. A hypothesis test for ρ = 0 uses t = r √(n-2) / √(1 – r²) with n-2 df.
皮尔逊积矩相关系数 r 测量两个变量间线性关系的强度和方向。r 的取值范围为 [-1, 1]，r = ±1 表示完全线性相关。公式：r = Σ (xᵢ – x̄)(yᵢ – ȳ) / √[Σ (xᵢ – x̄)² Σ (yᵢ – ȳ)²]。检验 ρ = 0 的假设使用 t = r √(n-2) / √(1 – r²)，自由度为 n-2。
Spearman’s rank correlation ρₛ is a non-parametric alternative that uses ranks and is robust to outliers. It assesses monotonic relationships. Use it when the data are ordinal or when the relationship is not linear.
斯皮尔曼等级相关系数 ρₛ 是一种使用秩次的非参数替代方法，对异常值稳健。它评估单调关系。当数据为次序变量或关系非线性时使用。
Simple linear regression models E(Y|X) = α + βx. The least squares estimates minimise the sum of squared residuals. The slope b = Sxy / Sxx, intercept a = ȳ – b x̄. The coefficient of determination R² = SSR/SST measures the proportion of variance in Y explained by the model. Residual plots help check linearity, constant variance, and normality assumptions.
简单线性回归模型为 E(Y|X) = α + βx。最小二乘估计最小化残差平方和。斜率 b = Sxy / Sxx，截距 a = ȳ – b x̄。决定系数 R² = SSR/SST 衡量模型解释的 Y 的方差比例。残差图有助于检查线性、等方差性和正态性假设。
10. Chi-Squared Tests | 卡方检验
The chi-squared (χ²) distribution arises from summing squares of independent standard normal variables. It is right-skewed with degrees of freedom ν. The χ² test for goodness-of-fit compares observed frequencies Oᵢ with expected frequencies Eᵢ under a hypothesised model. The test statistic X² = Σ (Oᵢ – Eᵢ)² / Eᵢ follows approximately a χ² distribution with df = number of categories – 1 – number of estimated parameters.
卡方 (χ²) 分布源于独立标准正态变量的平方和。它是右偏的，自由度为 ν。拟合优度 χ² 检验比较观测频数 Oᵢ 与假设模型下的期望频数 Eᵢ。检验统计量 X² = Σ (Oᵢ – Eᵢ)² / Eᵢ 近似服从 χ² 分布，自由度 = 类别数 – 1 – 所估计参数个数。
The test of independence (or association) for two categorical variables in a contingency table also uses the same statistic, with expected frequencies computed as E_{ij} = (row i total × column j total) / grand total. The degrees of freedom are (r – 1)(c – 1). All expected frequencies should be at least 5 for the approximation to be valid; if not, combine categories or use Fisher’s exact test.
针对列联表中两个分类变量的独立性（关联性）检验也使用同一统计量，期望频数计算为 E_{ij} = (第 i 行合计 × 第 j 列合计) / 总计。自由度为 (r – 1)(c – 1)。所有期望频数至少为 5 才能使近似有效；若不满足，合并类别或使用费希尔精确检验。
When interpreting a significant result, examine the standardised residuals (O – E)/√E to identify which cells contribute most to the large X². Remember, the χ² test does not imply causation; it only signals an association worth further investigation.
当结果显著时，检查标准化残差 (O – E)/√E 以识别对较大的 X² 贡献最大的单元格。请记住，χ² 检验并不意味因果关系；它仅提示存在值得进一步研究的关联。
Published by TutorHao | Pre-U OCR 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