📚 Year 13 WJEC Statistics: Core Concepts Review | Year 13 WJEC 统计：核心知识点梳理
Year 13 WJEC Statistics builds on foundational knowledge, introducing deeper inferential methods, probability distributions, and statistical tests. A clear understanding of these core topics is essential for success in the A Level examinations, as they form the basis of data-driven decision-making. This article consolidates the key concepts you need to master, from probability models to hypothesis testing and beyond.
Year 13 WJEC 统计学在基础之上引入了更深入的推断方法、概率分布和统计检验。清晰理解这些核心主题是 A Level 考试成功的关键，因为它们构成了数据驱动决策的基础。本文梳理了从概率模型到假设检验等必须掌握的核心知识点。
1. Probability Distributions & Expected Values | 概率分布与期望值
, its probability mass function P(X = x) gives the probability of each specific outcome. The sum of all probabilities must equal 1, a fact used to verify a valid distribution.
概率分布描述了随机变量所有取值的概率分配方式。对于离散型随机变量 X，其概率质量函数 P(X = x) 给出了每个具体结果发生的概率。所有概率之和必须等于 1，这个性质常用于检验分布的有效性。
pread. For a linear function Y = aX + b, we have E(Y) = aE(X) + b and Var(Y) = a²Var(X). These properties are fundamental when working with estimators and test statistics.
期望值 E(X) = Σ x·P(X = x) 代表随机变量的长期平均值。方差为 Var(X) = E(X²) - [E(X)]²，用于衡量离散程度。对于线性函数 Y = aX + b，有 E(Y) = aE(X) + b 和 Var(Y) = a²Var(X)，这些性质在处理估计量和检验统计量时非常基础。
2. Binomial Distribution | 二项分布
e same probability of success p. Its probability function is given by P(X = r) = C(n, r) pʳ (1-p)ⁿ⁻ʳ number of trials, two possible outcomes per trial, constant probability, and independence.
二项分布描述了在 n 次独立试验中成功次数的分布，每次试验成功概率为 p。其概率函数为 P(X = r) = C(n, r) pʳ (1-p)ⁿ⁻ʳ。需要满足：固定试验次数、每次试验只有两种结果、概率恒定以及独立性。
X ~ B(n, p), E(X) = np, Var(X) = np(1-p)
ay require use of the formula, tables, or technology. Always check whether an exact value or a cumulative probability is needed, and remember the complement rule: P(X ≥ r) = 1 - P(X ≤ r-1).
均值 np 和方差 np(1-p) 对于假设检验和正态近似至关重要。在计算概率时，WJEC 题目可能要求使用公式、查表或技术工具。要看清题目要求的是单个值还是累积概率，并记住互补规则：P(X ≥ r) = 1 - P(X ≤ r-1)。
3. Poisson Distribution | 泊松分布
happen independently at a constant mean rate λ. Its probability function is P(X = r) = (λʳ e⁻λ) / r! ndomness, independence, uniform average rate, and zero probability of simultaneous events.
泊松分布用于对固定时间或空间内事件发生次数进行建模，事件独立发生且平均率为 λ。概率函数为 P(X = r) = (λʳ e⁻λ) / r!。关键条件包括随机性、独立性、均匀的平均率以及事件不会同时发生。
X ~ Po(λ), E(X) = λ, Var(X) = λ
Its equal mean and variance provide a useful check when identifying an appropriate model. In WJEC exams, Poisson often appears in contexts like call centre arrivals or defects per metre. For large λ, a normal approximation may be used. When combining independent Poisson variables, the sum also follows a Poisson distribution: X ~ Po(λ₁) + Y ~ Po(λ₂) gives X+Y ~ Po(λ₁+λ₂).
其均值和方差相等这一性质可用于判断模型是否合适。在 WJEC 考试中，泊松分布常出现于呼叫中心来电次数或每米瑕疵数等场景。当 λ 较大时，可以使用正态近似。若将独立的泊松变量相加，和仍服从泊松分布：X ~ Po(λ₁) + Y ~ Po(λ₂) 得到 X+Y ~ Po(λ₁+λ₂)。
4. Normal Distribution & Approximations | 正态分布与近似
, symmetric, and bell-shaped, defined by its mean μ and variance σ². The standard normal Z ~ rdise first. The total area under the curve is 1, and symmetry gives P(Z < -a) = P(Z > a).
正态分布是连续、对称的钟形分布，由均值 μ 和方差 σ² 定义。标准正态分布 Z ~ N(0,1) 通过 Z = (X - μ)/σ 转化而来。使用查表或技术工具求概率；对于任意正态变量都需要先标准化。曲线下总面积为 1，对称性意味着 P(Z < -a) = P(Z > a)。
nuity correction adjusts for the discrete-to-continuous transition. For example, approximating B(n, p) with N(np, np(1-p)), P(X ≤ 23) becomes P(X < 23.5) using normal calculations.
当满足条件时，可将二项或泊松分布近似为正态。二项要求 np 和 n(1-p) 均大于 5；泊松要求 λ > 10。必须进行连续性校正来处理离散到连续的转换。例如，用 N(np, np(1-p)) 近似 B(n, p)，P(X ≤ 23) 应校正为 P(X < 23.5)。
5. Sampling Distributions & Central Limit Theorem | 抽样分布与中心极限定理
limit theorem (CLT) states that for sufficiently large samples (n ≥ 30), the sample mean X̄ follows an approximate normal distribution with mean μ and standard error σ/√n, regardless of the population shape. This underpins many inference procedures.
统计量的抽样分布是指固定样本量下所有可能取值所服从的分布。中心极限定理（CLT）指出，当样本量足够大（n ≥ 30），样本均值 X̄ 近似服从正态分布，均值为 μ，标准误为 σ/√n，而无论总体形状如何。这构成了许多推断方法的基础。
bility to distinguish when to use normal or t-distribution for confidence intervals and hypothesis tests. Remember that the sample proportion also has an approximate normal distribution for large n, using SE = √[p(1-p)/n].
当总体标准差已知时，标准误是精确的；当用 s 估计时，则需使用 t 分布。WJEC 常考查在置信区间和假设检验中何时使用正态分布或 t 分布。记住，样本比例在大样本下也近似为正态分布，标准误为 √[p(1-p)/n]。
6. Confidence Intervals for Mean & Proportion | 均值与比例的置信区间
ble values for an unknown population parameter. For a population mean with known σ, the 95% confidence interval is X̄ ± z × (σ/√n), where z = 1.96. When σ is unknown and estimated by s, use t × (s/√n) with n-1 degrees of freedom. The interpretation is that if we repeated the sampling many times, 95% of such intervals would contain the true mean.
置信区间给出了未知总体参数的可能取值范围。当 σ 已知时，总体均值的 95% 置信区间为 X̄ ± z × (σ/√n)，z = 1.96。当 σ 未知并用 s 估计时，使用 t × (s/√n)，自由度为 n-1。需要正确解读：若重复抽样多次，95% 的此类区间会包含真值。
For a population proportion, the large-sample confidence interval is p̂ ± z × √[p̂(1-p̂ on of the standard error. In each case, verify assumptions such as random sampling and approximate normality.
对于总体比例，大样本下的置信区间为 p̂ ± z × √[p̂(1-p̂)/n]。WJEC 大纲可能还包括两个均值差或两个比例差的置信区间，这时需要仔细计算标准误。每种情况都要验证随机抽样和近似正态性等假设。
7. Hypothesis Testing: One-Sample & Two-Sample | 假设检验：单样本与双样本
Hypothesis testing uses sample data to evaluate a claim about a population parameter. The null hypothesis H₀ represents the status quo, while H₁ is the alternative. A test statistic is computed and compared to a critical value to decide whether to reject H₀. The significance level α sets the probability of a Type I error (rejecting true H₀). A Type II error occurs when failing to reject a false H₀.
假设检验利用样本数据来评价关于总体参数的断言。原假设 H₀ 代表现状，备择假设 H₁ 与之相反。计算检验统计量并与临界值比较，以决定是否拒绝 H₀。显著性水平 α 是犯第一类错误（拒绝真 H₀）的概率；第二类错误指未能拒绝错误的 H₀。
One-sample tests for a mean include the z-test (σ known) and t-test (σ unknown). For a proportion, use a z-test. Two-sample tests compare two populations: for independent means, use a two-sample t-test; for paired data, use a paired t-test. Always check conditions: normality, independence, and equal variances if required by the test.
单样本均值检验包括 z 检验（σ 已知）和 t 检验（σ 未知）。比例检验使用 z 检验。双样本检验比较两个总体：独立样本均值使用双样本 t 检验；配对数据使用配对 t 检验。务必检查条件：正态性、独立性，以及检验所需的方差齐性。
In WJEC, p-values are increasingly used: if p-value < α, reject H₀. Report conclusions in context, using phrases like 'there is sufficient evidence at the 5% level to reject the null hypothesis'. Diagrams can aid understanding of critical regions.
WJEC 考试中越来越多使用 p 值：若 p 值 < α，则拒绝 H₀。在上下文中陈述结论，例如“在 5% 水平下有足够证据拒绝原假设”。图示可帮助理解拒绝域。
8. Correlation & 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 ranges from -1 to +1. A value close to 1 indicates strong positive correlation; near -1 indicates strong negative correlation; near 0 suggests no linear correlation. However, correlation does not imply causation.
相关衡量两变量间线性关系的强度和方向。积矩相关系数 r 取值范围为 -1 到 +1。接近 1 表示强正相关；接近 -1 表示强负相关；接近 0 表明没有线性相关。但相关关系并不意味着因果关系。
efficient of determination, r², shows the proportion of variation in y explained by x. You should be able to interpret a and b, calculate predicted values, and assess reliability. Residual plots help check model assumptions.
线性回归旨在用直线 y = a + bx 对关系建模，其中 b = Sxy / Sxx 由最小二乘法求得。决定系数 r² 表示由 x 解释的 y 的变化比例。应能解释截距 a 与斜率 b、计算预测值并评估可靠性。残差图有助于检查模型假设。
tic. If the correlation is significant, the regression line can be used for predictions within the range of observed data. Extrapolation beyond that range is unreliable.
相关性的假设检验 H₀: ρ = 0 对 H₁: ρ ≠ 0，使用 t 统计量。如果相关性显著，回归线可用于观测数据范围内的预测，超出该范围的外推是不可靠的。
9. Chi-Squared Tests for Goodness of Fit & Independence | 卡方检验：拟合优度与独立性
uencies from a theoretical distribution. The test statistic is Σ (O - E)² / E, with degrees of freedom = number of categories minus number of estimated parameters minus 1. A large value indicates poor fit.
卡方（χ²）检验是用于分类数据的非参数检验。拟合优度检验比较观察频数与理论分布的期望频数。检验统计量为 Σ (O - E)² / E，自由度 = 类别数 - 估计参数个数 - 1。统计量偏大表明拟合不佳。
ed under the assumption of no association: E = (row total × column total) / grand total. Degrees of freedom = (rows - 1)(columns - 1). Ensure expected frequencies are all at least 5; combine categories if necessary.
独立性检验使用列联表评估两个分类变量是否有关联。期望频数基于无关联假设计算：E = (行合计 × 列合计) / 总计。自由度 = (行数 - 1)(列数 - 1)。要确保所有期望频数至少为 5，必要时合并类别。
In WJEC, you must write hypotheses clearly, calculate contributions, and interpret the result. If χ² exceeds the critical value, reject H₀ of a good fit or independence. The test is one-tailed in the upper tail. Always comment on the validity of assumptions.
在 WJEC 考试中，须清晰地写出假设、计算各单元格贡献并解释结果。如果 χ² 超过临界值，则拒绝拟合良好或独立的原假设。该检验是单尾、上侧检验。务必评价假定的有效性。
10. Non-parametric Tests (Spearman & Mann-Whitney) | 非参数检验（斯皮尔曼等级相关与曼-惠特尼检验）
ation distribution and are useful for ordinal data or when normality is violated. Spearman' - (6Σd²) / [n(n² - 1)], where d is the difference in ranks. Test H₀: ρₛ = 0 against H₁: ρₛ ≠ 0 using critical values or p-values.
非参数检验不假定特定的总体分布，适用于顺序数据或正态性不满足的情况。斯皮尔曼等级相关系数（rₛ）衡量两变量间的单调关联强度。将数据分别排秩，使用 rₛ = 1 - (6Σd²)/[n(n² - 1)]，其中 d 为秩次差。检验 H₀: ρₛ = 0 对 H₁: ρₛ ≠ 0，可查临界值或计算 p 值。
late U₁ = R₁ - n₁(n₁+1)/2 (and similarly U₂), using the smaller U. The test examines shift in location; tied ranks require adjustments. It is often used as a non-parametric alternative to the two-sample t-test.
曼-惠特尼 U 检验比较两个独立样本的总体是否不同。合并所有观测并排秩，然后计算 U₁ = R₁ - n₁(n₁+1)/2（类似计算 U₂），取较小的 U 值。该检验考察位置偏移；存在相同秩次时需校正。它常作为双样本 t 检验的非参数替代。
These tests are common in WJEC Statistics, especially when the data are ranks or when distribution-free methods are required. Remember to state hypotheses, calculate correctly, and interpret the result with reference to the significance level.
这些检验在 WJEC 统计学中很常见，尤其当数据为等级或需要自由分布方法时。记住陈述假设、正确计算，并参照显著性水平解释结果。
Published by TutorHao | Statistics Revision Series | aleveler.com
更多咨询请联系16621398022（同微信）
或截图后用「扫一扫 → 相册」选择本图