Pre-U CCEA Statistics Unit Test Mock Paper Breakdown | Pre-U CCEA 统计单元测试模拟卷解析

📚 Pre-U CCEA Statistics Unit Test Mock Paper Breakdown | Pre-U CCEA 统计单元测试模拟卷解析

In the Pre-U CCEA Statistics specification, unit tests are designed to assess your ability to apply statistical concepts to unseen problems under timed conditions. This article walks you through a typical mock paper, highlighting common question types, key skills, and the reasoning expected in high-scoring answers. We cover probability, discrete and continuous distributions, estimation, hypothesis testing, correlation and regression, all presented at the depth demanded by CCEA.

在 Pre-U CCEA 统计课程中，单元测试旨在评估你在限时条件下将统计概念应用于新问题的能力。本文带你解析一份典型的模拟试卷，突出常见题型、关键技能以及高分答案所需的推理过程。我们涵盖概率、离散和连续分布、估计、假设检验、相关与回归等内容，所有讲解都达到 CCEA 要求的深度。

---

1. Probability Rules and Venn Diagrams | 概率规则与维恩图

Many mock papers open with a question combining conditional probability and set notation. A typical problem gives P(A), P(B) and P(A ∩ B), then asks for P(A ∪ B), P(A’ | B), and whether A and B are independent. The key is to write down known values and use the addition rule: P(A ∪ B) = P(A) + P(B) − P(A ∩ B). Independence is checked by testing P(A ∩ B) = P(A) × P(B). Always interpret results in context, for example stating that knowing B occurred changes the probability of A, so the events are not independent.

许多模拟卷以结合条件概率和集合符号的题目开篇。典型题目给出 P(A)、P(B) 和 P(A ∩ B)，然后要求计算 P(A ∪ B)、P(A’ | B)，并判断 A 与 B 是否独立。关键是将已知数值写下来，并利用加法公式：P(A ∪ B) = P(A) + P(B) − P(A ∩ B)。独立性通过检验 P(A ∩ B) = P(A) × P(B) 来判断。务必结合情境解释结果，例如说明知道 B 发生改变了 A 的概率，因此事件不独立。

---

2. Discrete Random Variables and Expectation Algebra | 离散随机变量与期望代数

A standard question presents a probability mass function for a discrete variable X, often incomplete, requiring you to find the missing probability and then calculate E(X), Var(X), and sometimes E(3X − 2) or Var(4 − 2X). Remember that for any constants a and b, E(aX + b) = aE(X) + b and Var(aX + b) = a²Var(X). If you need the variance, computing E(X²) is essential because Var(X) = E(X²) − [E(X)]². Students often lose marks by forgetting to square the coefficient in variance transformations.

有一道标准题会给出离散变量 X 的概率质量函数，通常不完整，要求你找出缺失的概率，然后计算 E(X)、Var(X)，有时还会计算 E(3X − 2) 或 Var(4 − 2X)。记住对于任意常数 a 和 b，有 E(aX + b) = aE(X) + b 以及 Var(aX + b) = a²Var(X)。如果需要计算方差，求出 E(X²) 至关重要，因为 Var(X) = E(X²) − [E(X)]²。学生常因在方差变换时忘记将系数平方而丢分。

---

3. Evaluating Discrete Distributions: Geometric and Negative Binomial | 离散分布建模：几何分布与负二项分布

When a scenario involves repeated independent trials until a success or a fixed number of successes, you must choose between geometric and negative binomial models. The geometric distribution with parameter p has probability function P(X = k) = (1 − p)^k−1p, for k = 1, 2, … Its mean is 1/p. The negative binomial with parameters r and p (r successes needed) has P(X = k) = (k − 1 choose r − 1) p^r (1 − p)^k−r, for k = r, r+1, … In both cases, state the assumptions clearly: independent trials, constant probability, and trials that are identical. A follow‑up part often asks you to calculate E(X) and Var(X) or to comment on whether the model remains appropriate if conditions change.

当情景涉及重复独立试验，直到成功或达到固定成功次数时，你必须在几何分布和负二项分布之间做出选择。参数为 p 的几何分布的概率函数为 P(X = k) = (1 − p)^k−1p，k = 1, 2, …其均值为 1/p。需要 r 次成功的负二项分布参数为 r 和 p，概率函数为 P(X = k) = (k − 1 choose r − 1) p^r (1 − p)^k−r，k = r, r+1, …两种情况下都要清楚陈述假设：试验独立、概率恒定、每次试验条件相同。后续小问常要求计算 E(X) 和 Var(X)，或评论当条件改变时模型是否仍然适用。

---

4. Poisson Distribution: Goodness of Fit and Approximations | 泊松分布：拟合优度与近似

A classic CCEA question provides observed frequencies for numbers of events per interval and a fitted Poisson mean. You may need to complete a goodness-of-fit test using the Poisson distribution. Remember to combine categories so that all expected frequencies are at least 5. The test statistic follows a χ² distribution with degrees of freedom ν = number of combined categories − 1 − number of estimated parameters. Often the mean has been estimated from the data, costing one extra degree of freedom. The null hypothesis is that the data follow a Poisson distribution; a large test statistic leads to rejection.

一道经典的 CCEA 题目给出每间隔事件数量的观测频数和拟合的泊松均值。你可能需要完成一个使用泊松分布的拟合优度检验。记住要合并类别，使得所有期望频数至少为 5。检验统计量服从 χ² 分布，自由度 ν = 合并后的类别数 − 1 − 估计参数的个数。通常均值是从数据估计的，这会多消耗一个自由度。原假设是数据服从泊松分布；统计量很大则拒绝原假设。

---

5. Normal Distribution: Inverse Problems and Sampling Distributions | 正态分布：反向问题与抽样分布

Proficiency in standardising to Z is tested repeatedly. A typical item gives X ~ N(µ, σ²) with known parameters and asks for a value k such that P(X > k) = 0.1. Use Z = (X − µ) / σ, find the z‑value corresponding to a tail probability of 0.1, then solve for k: k = µ + z × σ. When the sample mean is involved, you must use the sampling distribution X̅ ~ N(µ, σ²/n). Questions can also ask for the probability that the sample total exceeds a number: the total T ~ N(nµ, nσ²). Always sketch a bell curve to confirm the required tail.

将变量标准化为 Z 的熟练度会被反复考查。典型题目给出 X ~ N(µ, σ²) 且参数已知，要求求出使得 P(X > k) = 0.1 的 k 值。使用 Z = (X − µ) / σ，找出对应右尾概率 0.1 的 z 值，然后求解 k：k = µ + z × σ。当涉及样本均值时，必须使用抽样分布 X̅ ~ N(µ, σ²/n)。题目也可能要求计算样本总和超过某一数值的概率：总和 T ~ N(nµ, nσ²)。始终画一条钟形曲线以确认所需尾部。

---

6. Confidence Intervals for Mean and Proportion | 均值和比例的置信区间

Constructing a confidence interval requires selecting the correct standard error and critical value. For a population mean with known σ, use zα/2; if σ is unknown and the sample is large, estimate σ with s and still use z. For a small sample from a normal population with unknown σ, use tν,α/2 with ν = n − 1. A proportion p based on a large sample uses the standard error √[p̂(1 − p̂)/n]. The confidence statement has the form estimate ± (critical value) × standard error. In the exam, you may be asked to interpret the interval: we are C% confident that the true parameter lies between the lower and upper bounds.

构建置信区间需要选择正确的标准误和临界值。对于已知 σ 的总体均值，使用 zα/2；如果 σ 未知但样本量大，用 s 估计 σ 并仍使用 z。对于来自正态总体的小样本且 σ 未知，使用 tν,α/2，其中 ν = n − 1。大样本下的比例 p 使用的标准误为 √[p̂(1 − p̂)/n]。置信区间的形式为 点估计 ± (临界值) × 标准误。考试中可能要求你解释区间：我们有 C% 的信心认为真实参数落在下限与上限之间。

---

7. Hypothesis Testing: One‑Sample and Two‑Sample Tests | 假设检验：单样本与双样本检验

A structured hypothesis test includes stating H₀ and H₁ clearly, calculating the test statistic, and comparing it with a critical value or using a p‑value. For a one‑sample z‑test for the mean, the test statistic is z = (x̅ − µ₀) / (σ/√n). When comparing two means from independent normal populations with known variances, use the two‑sample z‑test. If variances are unknown but assumed equal, the pooled variance estimate is s²ₚ = [(n₁−1)s²₁ + (n₂−1)s²₂] / (n₁+n₂−2), and the test statistic follows a t‑distribution with n₁+n₂−2 degrees of freedom. Do not forget the conclusion: reject H₀ or do not reject, linked to the context.

结构化的假设检验包括清晰地陈述 H₀ 和 H₁，计算检验统计量，并将其与临界值比较或使用 p 值。对于均值的单样本 z 检验，检验统计量为 z = (x̅ − µ₀) / (σ/√n)。比较两个来自正态总体且方差已知的独立样本的均值时，使用双样本 z 检验。如果方差未知但假设相等，合并方差估计为 s²ₚ = [(n₁−1)s²₁ + (n₂−1)s²₂] / (n₁+n₂−2)，检验统计量服从自由度为 n₁+n₂−2 的 t 分布。不要忘记给出结论：拒绝或不拒绝 H₀，并与背景联系起来。

---

8. Chi‑Squared Tests for Association | 卡方独立性检验

Contingency table questions require you to compute expected frequencies under the assumption of no association: expected cell frequency = (row total × column total) ÷ grand total. The test statistic is Σ (O − E)² / E, which under H₀ follows a χ² distribution with (r−1)(c−1) degrees of freedom. When testing at a given significance level, compare the computed χ² with the critical value from tables. If the observed value exceeds the critical value, there is evidence of an association. Comment on the nature of the association by comparing observed and expected frequencies.

列联表题目要求你在假设没有关联的情况下计算期望频数：期望频数 = (行合计 × 列合计) ÷ 总计。检验统计量为 Σ (O − E)² / E，在原假设下服从自由度为 (r−1)(c−1) 的 χ² 分布。在给定显著性水平下检验时，将计算出的 χ² 值与表中的临界值比较。如果观测值大于临界值，则有证据表明存在关联。通过比较观测频数和期望频数来评论关联的性质。

---

9. Correlation and Spearman’s Rank Coefficient | 相关性与斯皮尔曼等级相关系数

Pearson’s correlation coefficient r measures the strength of a linear relationship. However, when data are not normally distributed or are ranked, Spearman’s rank coefficient rₛ is more appropriate. Rank the x‑values and y‑values separately, then compute the difference d for each pair, and use rₛ = 1 − (6 Σd²) / [n(n² − 1)]. A hypothesis test for zero correlation can be performed using the test statistic t = rₛ √[(n−2)/(1−rₛ²)] with n−2 degrees of freedom. In the interpretation, note that correlation does not imply causation.

皮尔逊相关系数 r 衡量线性关系的强度。然而，当数据不服从正态分布或是等级数据时，斯皮尔曼等级相关系数 rₛ 更合适。分别对 x 值和 y 值排秩，然后计算每对数据的差 d，并使用 rₛ = 1 − (6 Σd²) / [n(n² − 1)]。可使用检验统计量 t = rₛ √[(n−2)/(1−rₛ²)] 进行零相关的假设检验，其自由度为 n−2。在解释时注意，相关并不意味着因果关系。

---

10. Linear Regression and Residual Analysis | 线性回归与残差分析

Given summary statistics Σx, Σy, Σx², Σy², Σxy and n, you can fit a least‑squares regression line of y on x: ŷ = a + bx. The slope b = Sxy / Sxx, where Sxy = Σxy − (Σx)(Σy)/n and Sxx = Σx² − (Σx)²/n. The intercept a = y̅ − b x̅. A follow‑up often asks you to predict y for a given x; only do this if x lies within the range of the original data. Residual plots help check model assumptions: constant variance (no pattern) and linearity. A quadratic pattern in residuals suggests a non‑linear relationship. Understanding these checks can earn you marks in the interpretation section.

给定汇总统计量 Σx、Σy、Σx²、Σy²、Σxy 和 n，你可以拟合 y 对 x 的最小二乘回归线：ŷ = a + bx。斜率 b = Sxy / Sxx，其中 Sxy = Σxy − (Σx)(Σy)/n，Sxx = Σx² − (Σx)²/n。截距 a = y̅ − b x̅。后续问题常要求对给定的 x 值预测 y；仅当 x 在原数据范围内时才进行预测。残差图有助于检查模型假设：方差齐性（无模式）和线性。若残差呈现二次曲线模式，则表明存在非线性关系。理解这些检查能在解释部分为你赢得分数。

---

11. Combining Random Variables and Moment Generating Functions | 随机变量组合与矩生成函数

For independent normal variables, linear combinations remain normal. For example, if X ~ N(µ₁, σ₁²) and Y ~ N(µ₂, σ₂²) independently, then X − Y ~ N(µ₁ − µ₂, σ₁² + σ₂²). Questions may ask for the distribution of a sum of i.i.d. variables or a scaled total. In the Pre‑U syllabus, moment generating functions (MGFs) are occasionally tested. The MGF of a random variable X is M_X(t) = E(e^tX). If Y = a + bX, then M_Y(t) = e^at M_X(bt). Recognising standard MGFs helps identify distributions of sums of independent variables. Practice writing the derivation in clear steps.

对于独立的正态变量，线性组合仍服从正态分布。例如，若 X ~ N(µ₁, σ₁²) 且 Y ~ N(µ₂, σ₂²) 独立，则 X − Y ~ N(µ₁ − µ₂, σ₁² + σ₂²)。题目可能要求求出独立同分布变量之和的分布或缩放后的总和。在 Pre‑U 考纲中，矩生成函数 (MGF) 偶有考查。随机变量 X 的 MGF 为 M_X(t) = E(e^tX)。如果 Y = a + bX，则 M_Y(t) = e^at M_X(bt)。识别标准 MGF 有助于确定独立变量之和的分布。练习用清晰步骤写出推导过程。

---

12. Common Pitfalls and Final Advice | 常见错误与最后建议

The most frequent errors include using the wrong variance formula for transformed variables, forgetting to adjust degrees of freedom when parameters are estimated, misreading tail probabilities from normal tables, and omitting units in final answers. Before the exam, create a summary card with all standard errors and degrees of freedom for tests. Under timed pressure, read each question twice: once to grasp the context, and once to extract all numerical details. Show your working step by step – even if your final answer is incorrect, method marks can still be earned. Finally, check that your conclusions match the original problem statement.

最常见的错误包括：对变换后的变量使用错误的方差公式、在估计参数时忘记调整自由度、从正态分布表中误读尾概率、以及最终答案遗漏单位。考前制作一张总结卡，罗列所有检验的标准误和自由度。在限时压力下，每道题读两遍：第一遍把握背景，第二遍提取所有数值细节。按步骤展示解答过程——即使最终答案错误，仍可获得方法分。最后，检查你的结论是否与原始问题陈述相符。

Published by TutorHao | Statistics Revision Series | aleveler.com

更多咨询请联系16621398022（同微信）