📚 Year 13 WJEC Statistics: High-Frequency Topics and Common Mistakes Analysis | Year 13 WJEC 统计：高频考点与易错题分析
This article breaks down the most frequently examined topics in Year 13 WJEC Statistics and pulls together the typical errors that can cost you marks. We cover everything from probability distributions to hypothesis testing and regression, with paired English–Chinese explanations that mirror how questions are actually set in WJEC papers. Use this guide to sharpen your revision and walk into the exam with confidence.
本文拆解了 Year 13 WJEC 统计中最常考的主题，并汇总了那些容易让你丢分的典型错误。从概率分布到假设检验，再到回归分析，我们用中英对照的方式还原 WJEC 真题的设问思路。请将这份指南用于精准备考，自信地走进考场。
1. Normal Distribution and Z‑score Mastery | 正态分布与 Z 分数的彻底掌握
ons repeatedly test your ability to standardise a normal variable to the standard normal Z ~ = (X − μ) / σ. A common slip is using the variance σ² instead of the standard deviation σ – always check whether the question gives σ² or σ directly.
WJEC 题目反复考查将正态变量标准化为标准正态 Z ~ N(0, 1) 的能力。你必须会用公式 Z = (X − μ) / σ 正向求概率、反向由概率求 X。常见错误是混淆了方差 σ² 与标准差 σ – 答题前务必确认题目给出的是方差还是标准差。
When finding an unknown mean μ or standard deviation σ given a probability, many students set up the Z‑equation but then forget to handle the sign correctly for lower‑tail probabilities. Draw a quick sketch and label the area – if the given area is less than 0.5, the corresponding Z‑value will be negative; if greater than 0.5, Z is positive.
在给定概率求未知均值 μ 或标准差 σ 时，许多学生能列出 Z 等式，却忽略了根据概率尾端处理正负号。务必画个简图并标出面积 – 若给出面积小于 0.5，对应的 Z 值为负；若大于 0.5，Z 为正。
2. Binomial and Poisson Distributions – Spotting the Conditions | 二项分布与泊松分布 – 识别适用条件
ams, you are expected to state the conditions before applying the formula P(X = r) = ⁿCᵣ pʳ (1 − p)ⁿ⁻ʳ re trials are not independent (e.g. sampling without replacement from a small population).
二项分布 B(n, p) 要求固定次数的独立试验，每次试验仅有两种结果且概率 p 恒定。WJEC 考试要求你在使用公式 P(X = r) = ⁿCᵣ pʳ (1 − p)ⁿ⁻ʳ 之前先陈述这些条件。一个典型错误是将二项分布用于试验并不独立的情形（如从小总体中不放回抽样）。
ppen independently at a constant average rate λ. It is often tested alongside the binomial – ance formula incorrectly when calculating standard deviation: the standard deviation is √λ, not λ.
泊松分布 Po(λ) 对固定区间内独立事件的发生次数建模，事件以恒定平均率 λ 发生。常与二项分布结合考查 – 考官希望你识别何时用 Po(λ)，并牢记其均值与方差都是 λ。一个极常见错误是计算标准差时错误使用泊松方差：标准差是 √λ，而非 λ。
3. Normal Approximations and the Continuity Correction | 正态近似与连续性校正
Both binomial and Poisson distributions can be approximated by a normal distribution under certain conditions (np > 5 and nq > 5 for binomial; λ > 10 for Poisson). WJEC mark schemes are strict about applying the continuity correction npq), a probability such as P(X ≥ 10) must be written as P(X > 9.5) and then standardised.
在满足特定条件时（二项需 np > 5 且 nq > 5；泊松需 λ > 10），可用正态分布近似。WJEC 的评分标准对用正态近似处理离散分布时的连续性校正要求非常严格。例如用 N(np, npq) 近似 B(n, p) 时，P(X ≥ 10) 必须写成 P(X > 9.5) 再标准化。
tween the integer value and the next integer shifts halfway. For ≤, include the upper half (e.g. P(X ≤ 8) → P(X < 8.5)); for ≥, subtract half (e.g. P(X ≥ 8) → P(X > 7.5)).
很多考生因校正方向错误或完全遗漏而丢分。一个记忆诀窍：整数值与下一个整数的边界移动半步。≤ 时取上半步（如 P(X ≤ 8) → P(X < 8.5)）；≥ 时取下半步（如 P(X ≥ 8) → P(X > 7.5)）。
4. Hypothesis Testing Framework and Writing Conclusions | 假设检验框架与结论撰写
Every hypothesis test in WJEC Year 13 must follow a clear structure: state the null hypothesis H₀ and alternative H₁, identify the test statistic and its distribution, calculate the test statistic from the sample, find the critical value or p‑value, compare, and write a conclusion in context. A vague conclusion like 'reject H₀' without linking back to the original claim will not earn full marks.
WJEC Year 13 中任何假设检验都必须遵循清晰结构：陈述原假设 H₀ 与备择假设 H₁，确定检验统计量及其分布，由样本计算检验统计量，查临界值或求出 p 值，进行比较，并在具体语境中写出结论。如果仅仅写“拒绝 H₀”而无原文背景的链接，就无法拿到满分。
Another high‑frequency pitfall is confusing one‑tailed and two‑tailed tests. If the question says 'has increased' or 'is greater than', a one‑tailed test should be set up with H₁: parameter > value. When the wording is 'has changed' or 'differs from', use a two‑tailed test. Always divide the significance level by 2 for two‑tailed critical values; failing to halve α is a common error.
另一个高频易错点是把单尾检验和双尾检验混淆。若题目描述为“是否增加”或“是否大于”，应设立单尾备择假设 H₁: 参数 > 某值。若表达为“是否改变”或“是否有差异”，则用双尾检验。双尾检验的临界值需要将显著性水平 α 除以 2；忘记折半 α 是常见失误。
5. Type I and Type II Errors, and the Power of a Test | 第一类错误、第二类错误与检验功效
WJEC frequently assesses understanding of Type I error (rejecting a true H₀) and Type II error (failing to reject a false H₀). You should be able to link the significance level to the probability of a Type I error and explain how increasing sample size reduces both error probabilities. The phrase 'at the 5% significance level' means the probability of making a Type I error is 0.05.
WJEC 经常考查对第一类错误（拒绝了真实的 H₀）与第二类错误（未能拒绝错误的 H₀）的理解。你需要能将显著性水平与第一类错误的概率关联起来，并解释增加样本量如何同时降低两类错误的概率。“在 5% 显著性水平下”意味着发生第一类错误的概率是 0.05。
size, larger effect size, and larger α all increase power. A common mistake is confusing β (Type II error probability) with power – remember power = 1 − P(Type II error).
检验的功效 (1 − β) 是正确拒绝错误 H₀ 的概率。题目常要求描述影响功效的因素：更大的样本量、更大的效应量、更大的 α 都会提升功效。常见混淆点是将 β（第二类错误概率）与功效本身搞错 – 需牢记功效 = 1 − P(Type II error)。
6. Tests for a Population Mean – When to Use z or t | 总体均值的检验 – 何时用 z 何时用 t
riance σ² is known. If σ² is known (or the sample is large, typically n ≥ 30), you use a z‑test with Z = (x̄ − μ₀) / (σ/√n). WJEC students often apply z when σ is unknown but n is large, relying on the central limit theorem – this is acceptable, but only if the question doesn't give s as an alternative.
检验总体均值时，选择哪种检验统计量取决于总体方差 σ² 是否已知。若 σ² 已知（或样本量很大，通常 n ≥ 30），则使用 z 检验 Z = (x̄ − μ₀) / (σ/√n)。WJEC 学生常在 σ 未知但 n 较大时仍用 z 检验，依据中心极限定理 – 这可行，但前提是题目未提供样本标准差 s。
ion variance is unknown and estimated by the sample standard deviation s, you must use a t‑test with T = (x̄ − μ₀) / (s/√n eedom ν = n − 1. A very frequent error is using the normal critical value instead of the t‑value. Even when n is large, if a t‑distribution is specified or s is the only measure given, stick to t.
当总体方差未知且用样本标准差 s 估计时，必须用 t 检验 T = (x̄ − μ₀) / (s/√n)，自由度为 ν = n − 1。一个极常见错误是使用正态临界值而非 t 临界值。即使 n 很大，只要题目指定 t 分布或仅给 s，就应坚持用 t 检验。
7. Tests and Confidence Intervals for Proportions | 比例的检验与置信区间
p, WJEC tests rely on the normal approximation to the binomial, with test statistic Z = (p̂ e marks when the method is unjustified. Always state the check before performing the test.
对于单个比例 p，WJEC 考试基于正态近似法，检验统计量 Z = (p̂ − p₀) / √[p₀(1 − p₀)/n]。学生有时忘记检查正态近似条件（np₀ > 5, n(1 − p₀) > 5），导致方法不合理而丢分。请务必先陈述条件再进行检验。
Confidence intervals for a proportion are constructed as p̂ ± z* × √[p̂(1 − p̂)/n]. A common misstep is using p₀ in the standard error instead of p̂. For a confidence interval, we estimate the standard error with p̂ e p is unknown. The interval provides a range of plausible values for the true proportion.
比例的置信区间公式为 p̂ ± z* × √[p̂(1 − p̂)/n]。常见失误是在标准误中使用 p₀ 而不是 p̂。在构造置信区间时，由于 p 未知，我们使用 p̂ 来估计标准误。该区间给出了真实比例的可能范围。
8. Chi‑Squared Tests – Goodness of Fit and Independence | 卡方检验 – 拟合优度与独立性检验
Chi‑ eedom ν = number of categories − 1 − number of estimated parameters. A classic error is failing to adjust degrees of freedom when population parameters are estimated from the data (e.g. a normal distribution with mean and variance estimated from the sample loses two extra degrees of freedom).
卡方检验几乎每年都会出现在 WJEC A2 统计试卷中。拟合优度检验的统计量为 χ² = Σ (Oᵢ − Eᵢ)² / Eᵢ，自由度 ν = 类别数 − 1 − 估计参数的个数。经典错误是当总体参数由数据估计时未调整自由度（例如，用样本估计正态分布的均值和方差时，会额外减少两个自由度）。
cies are calculated as (row total × column total) / grand total. WJEC candidates often mis‑ requires all expected frequencies to be at least 5; if not, categories should be combined – this instruction is frequently missed.
对于列联表中的独立性检验，期望频数按（行合计 × 列合计）/ 总计 计算。WJEC 考生常把自由度误算为 行数 × 列数，正确应为 (行数 − 1) × (列数 − 1)。此外，检验要求所有期望频数至少为 5；若不满足，应合并类别 – 此要求经常被忽略。
9. Correlation and Regression – Inference beyond the Line | 相关与回归 – 超越回归线的推断
WJEC expects you to perform a hypothesis test for the product‑moment correlation coefficient ρ using either a t‑ es from a table. When the null hypothesis H₀: ρ = 0 is tested, the test statistic is T = r √n − 2 / √1 − r² ks by using the wrong tail or misreading the Spearman rank correlation table when data are non‑linear.
WJEC 要求你对积差相关系数 ρ 进行假设检验，使用 t 检验或查表找临界值。在检验 H₀: ρ = 0 时，检验统计量为 T = r √n − 2 / √1 − r²，自由度 ν = n − 2。学生们常因用错尾端，或在非线性数据下误用 Spearman 等级相关系数表而失分。
For the least‑ e asked to find a confidence interval for the slope β or test H₀: β = 0. The standard error of b involves the residual variance s² = Σ(y − ŷ)² / (n − 2). A common slip is using n or n − 1 instead of n − 2.
对于最小二乘回归线 y = a + bx，你需要根据汇总统计量计算 a 和 b，并解释斜率 b。在推断问题中，可能会要求求斜率 β 的置信区间或检验 H₀: β = 0。b 的标准误涉及残差方差 s² = Σ(y − ŷ)² / (n − 2)。常见错误是除数误用 n 或 n − 1，正确应为 n − 2。
10. Common Exam Traps and How to Avoid Them | 常见考试陷阱及规避方法
After years of WJEC mark schemes, a pattern of common slips becomes clear. Always check whether a variance or standard deviation is given; misreading this can destroy an entire normal or binomial question. In continuity corrections, forgetting to adjust the boundary by 0.5 is the top single error in approximation questions.
分析多年 WJEC 评分方案后可看出一个清晰模式。务必确认给出的是方差还是标准差；一旦误读，整道正态或二项题目就会崩塌。在连续性校正中，忘记将边界移动 0.5 是近似类题目中最惨重的单点失误。
Writing final answers with appropriate units and to the required precision (usually 3 significant figures for test statistics) is insisted upon. In hypothesis tests, stating 'accept H₀' instead of 'do not reject H₀' is penalised because we never accept the null hypothesis – we only lack evidence to reject it.
最终答案须带上恰当的单位并达到规定精度（检验统计量通常取 3 位有效数字）。在假设检验中，将“不拒绝 H₀”写成“接受 H₀”会被扣分，因为我们从不接受原假设 – 只是缺乏拒绝的证据。
Finally, practice interpreting p‑values in context: 'If H₀ were true, the probability of obtaining a result at least as extreme as the observed one is the p‑value.' This phrase is a frequent request in WJEC explanation questions, and a memorised template can secure two easy marks.
最后，要练习在语境中解释 p 值：“若 H₀ 为真，得到与观测结果相同或更极端结果的概率即为 p 值。”这句话是 WJEC 解释题中的常客，背下模板就能稳拿两分。
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