📚 Case Study Practical Exercises | 案例分析实战演练
Welcome to a focused revision guide designed for Year 12 CCEA Statistics students. This article will help you master the art of tackling case study questions by breaking down real-world scenarios into manageable statistical steps. You will learn how to identify the problem, collect and analyse data, apply probability distributions, carry out hypothesis tests, and draw meaningful conclusions. The approach blends theory with practical examples, ensuring you feel confident in applying statistical methods to any context the exam may present.
欢迎阅读这篇为 Year 12 CCEA 统计学生精心准备的专题复习指南。本文将帮助你掌握解决案例分析问题的艺术，通过将真实情境分解为可管理的统计步骤。你将学会如何识别问题、收集和分析数据、应用概率分布、进行假设检验并得出有意义的结论。该方法融合了理论与实践，确保你在任何考试可能出现的背景下都能自信地应用统计方法。
1. What Is a Statistical Case Study? | 什么是统计案例分析？
A statistical case study presents a realistic situation where you are required to apply statistical techniques to investigate a question, test a claim, or inform a decision. In CCEA Year 12 exams, these tasks often involve a short scenario — a business analysing sales, a researcher testing a new medication, or a school evaluating a revision programme. The key is to avoid simply performing calculations; you must interpret results in the context of the original problem and present clear, justified conclusions.
统计案例分析呈现一个现实情境，要求你运用统计技术来调查一个问题、检验一个主张或为决策提供依据。在 CCEA Year 12 考试中，这类任务通常包含一个简短的场景——一家企业分析销售情况、一名研究人员测试新药，或一所学校评估复习计划。关键在于不能仅仅执行计算；你必须结合原始问题解读结果，并给出清晰、有依据的结论。
2. Identifying Objectives and Variables | 确定目标与变量
Every case study starts with a clear objective. Read the scenario and ask: 'What are we trying to find out or prove?' This will lead you to define the response variable (dependent) and one or more explanatory variables (independent). For instance, a coffee shop wants to test whether playing background music increases the average spend per customer. Here, the response is 'spend per customer' and the explanatory variable is 'music condition (on/off)'. Also note whether variables are continuous, discrete, or categorical, as this determines the correct statistical tool.
每一个案例研究都始于清晰的目标。仔细阅读场景并自问：“我们想要发现或证明什么？”这将引导你定义响应变量（因变量）和一个或多个解释变量（自变量）。例如，一家咖啡店想测试播放背景音乐是否会提高顾客的人均消费。这时，响应变量是“顾客人均消费”，解释变量是“音乐状态（开/关）”。同时注意变量是连续的、离散的还是分类的，这将决定你选择哪种正确的统计工具。
3. Planning Data Collection and Sampling | 规划数据收集与抽样
Once the variables are clear, consider how the data will be gathered. Does the case study mention a census or a sample? For samples, you need to identify the sampling method: simple random, stratified, systematic, or opportunity sampling. Evaluate its strengths and potential biases. For example, a stratified sample by age group ensures each segment is represented, reducing bias when comparing spending habits across ages. Also think about sample size — larger samples reduce variability but may not always be feasible. If the data is collected through an experiment, pay attention to control groups, randomisation, and blinding.
一旦变量明确，就要考虑数据如何收集。案例研究中是否提及了普查还是抽样？如果是样本，你需要识别抽样方法：简单随机抽样、分层抽样、系统抽样或机会抽样。评估其优点和潜在偏差。例如，按年龄段分层抽样可确保每个群体都有代表，从而在比较不同年龄段消费习惯时减少偏差。还要考虑样本量——更大的样本能减小变异性，但并非总是可行。如果数据通过实验收集，就要关注对照组、随机化和盲法。
4. Exploratory Data Analysis (EDA) | 探索性数据分析 (EDA)
Before jumping into inference, always summarise the data using descriptive statistics. Calculate measures of central tendency (mean, median, mode) and measures of spread (range, interquartile range, standard deviation). If the data is heavily skewed, the median and IQR are more appropriate than the mean and standard deviation. For bivariate data, compute the correlation coefficient. Present these statistics in a clear table. This step helps you spot outliers, patterns, and unusual features that might influence later analysis.
在进入推断之前，始终用描述性统计方法对数据进行汇总。计算集中趋势的度量（均值、中位数、众数）和离散程度的度量（极差、四分位距、标准差）。如果数据严重偏斜，使用中位数和IQR比均值和标准差更合适。对于双变量数据，计算相关系数。将这些统计量呈现在清晰的表格中。这一步骤帮助你发现可能影响后续分析的异常值、模式和异常特征。
| Statistic | Definition | Typical Use |
|---|---|---|
| Mean (x̄) | Sum of values divided by count | Symmetric, no outliers |
| Median | Middle value when sorted | Skewed data or outliers present |
| Standard deviation (s) | Measures spread around the mean | Used in normal distribution models |
| IQR | Q₃ – Q₁ | Robust spread for skewed data |
表格中：x̄ 表示样本均值，Q₁ 和 Q₃ 为四分位数。选择合适的统计量是有效描述数据的关键。
5. Visualising Data with Appropriate Graphs | 用合适的图表可视化数据
Graphs bring data to life and are often required in case study answers. Choose the chart based on the variable type: histograms for continuous data to show distribution shape, box plots to compare groups and identify outliers, scatter graphs to explore relationships between two continuous variables, and bar charts or pie charts for categorical data. Always label axes clearly and add a title. In an exam, a quick, well-labelled sketch can earn marks even if not perfectly scaled.
图表使数据变得生动，在案例分析答案中也常常需要用到。根据变量类型选择图表：连续数据用直方图展示分布形状，用箱线图比较各组并识别异常值，散点图用于探究两个连续变量之间的关系，分类数据则用条形图或饼图。务必清晰标注坐标轴并添加标题。在考试中，一个快速、标注清晰的草图即使比例不完美也能得分。
6. Probability Concepts and Distributions | 概率概念与分布
success, the binomial distribution B(n, p) is the ideal tool. Check the conditions carefully: each trial must be identical and independent, with only two outcomes. For continuous measurements like weights or test scores, the normal distribution N(μ, σ²) is commonly used, provided the data is roughly symmetric or the sample mean distribution applies via the Central Limit Theorem. You may need to calculate probabilities using tables or standardised z-scores: z = (x − μ)/σ.
许多案例分析涉及概率建模。对于具有固定次数独立试验和恒定成功概率的离散情形，二项分布 B(n, p) 是理想工具。仔细检查条件：每次试验必须相同且独立，只有两种结果。对于重量或考试分数等连续测量值，如果数据大致对称，或通过中心极限定理可应用样本均值分布，则通常使用正态分布 N(μ, σ²)。你可能需要使用表格或标准化 z 分数来计算概率：z = (x − μ)/σ。
7. Building Statistical Models | 构建统计模型
d variance σ². In some case studies, you might fit a least-squares regression line y = a + bx to bivariate data. Write the equation, compute predictions, and calculate residuals to check fit. Remember, the model is a simplification of reality — always discuss its assumptions and limitations in your evaluation.
一旦确定了合适的分布，就要明确模型参数。对于二项分布，清楚说明 n 和 p；对于正态分布，说明均值 μ 和方差 σ²。在某些案例分析中，你可能需要为双变量数据拟合一条最小二乘回归直线 y = a + bx。写出方程，计算预测值，并计算残差以检验拟合优度。请记住，模型是对现实的简化——在评估中一定要讨论其假设和局限性。
8. Inference: Confidence Intervals | 推断：置信区间
stimate population parameters from sample data. For a population mean with known σ, the 95% CI is x̄ ± 1.96 × (σ/√n). When σ is unknown, use the t-distribution with s/√n. For a population proportion p, the interval is p̂ ± z* × √(p̂(1−p̂)/n). In a case study, report the interval in context, e.g., "We are 95% confident that the true average spend increase is between £1.20 and £2.80." This shows you can translate numerical output into business or research language.
置信区间使我们能够根据样本数据估计总体参数。对于已知 σ 的总体均值，95% 置信区间为 x̄ ± 1.96 × (σ/√n)。当 σ 未知时，使用 t 分布，区间为 x̄ ± t* × (s/√n)。对于总体比例 p，置信区间为 p̂ ± z* × √(p̂(1−p̂)/n)。在案例分析中，要结合情境报告区间，例如：“我们有 95% 的把握认为真正的平均消费增长在 £1.20 到 £2.80 之间。”这显示了你能够将数值输出转化为商业或研究语言。
9. Hypothesis Testing: Step-by-Step | 假设检验：逐步讲解
Hypothesis tests are a cornerstone of statistical case studies. Follow a structured procedure:
- State the null hypothesis H₀ and alternative H₁ (one-tailed or two-tailed).
- Select the significance level α (commonly 5% or 1%).
- Calculate the test statistic: for a mean with known σ, use z = (x̄ − μ₀)/(σ/√n); for a binomial proportion, use the normal approximation or exact binomial p-value.
- Compare the test statistic with the critical value, or find the p-value.
- Make a decision: if p-value < α, reject H₀; otherwise, do not reject H₀.
- Write a conclusion in context, avoiding definitive statements like 'prove'. Use 'there is sufficient evidence to suggest...' or 'the result is not significant at the 5% level, so we cannot reject...'.
假设检验是统计案例分析的核心。遵循结构化步骤：
- 陈述原假设 H₀ 和备择假设 H₁（单尾或双尾）。
- 选择显著性水平 α（通常为 5% 或 1%）。
- 计算检验统计量：对于已知 σ 的均值，使用 z = (x̄ − μ₀)/(σ/√n)；对于二项比例，可使用正态近似或精确二项 p 值。
- 将检验统计量与临界值进行比较，或找出 p 值。
- 做出决策：若 p 值 < α，拒绝 H₀；否则，不拒绝 H₀。
- 在情境中写下结论，避免使用“证明”等绝对化词语。应使用“有充分证据表明……”或“在 5% 水平下结果不显著，因此我们不能拒绝……”。
10. Interpreting Statistical Output in Context | 在情境中解读统计输出
Numbers alone are meaningless in a case study; they must be linked back to the real-world question. If a test shows a significant increase in sales after a marketing campaign, discuss whether the increase is practically important, not just statistically significant. Mention potential confounding variables: Was there a seasonal effect? Could other factors have changed? Also, critique the reliability of the data: small sample sizes might lead to wide confidence intervals, and non-random sampling limits generalisability. Asking these questions demonstrates higher-order thinking, which CCEA examiners reward.
在案例分析中，单纯的数字没有意义；它们必须与现实问题联系起来。如果检验显示营销活动后销售额有显著增长，要讨论这种增长是否具有实际重要性，而不仅仅是统计显著性。提及可能存在的混杂变量：是否存在季节性效应？其他因素是否发生了变化？同时，批判性地评估数据的可靠性：样本量小可能导致置信区间较宽，非随机抽样限制了结论的普适性。提出这些问题体现高阶思维，这正是 CCEA 考官所看重的。
11. Walkthrough of a Full Case Study | 完整案例分析贯穿
Let' ore introduces a loyalty card and claims the average monthly spend per member is now above £35. A sample of 50 members yields a mean spend of £37.20 with a standard deviation s = £6.50. The population standard deviation is unknown. We want to test at the 5% significance level whether the true mean has increased above £35.
让我们来演练一个真实的场景：一家书店推出会员卡，声称会员的月均消费现在超过 £35。一个容量为 50 的样本显示，平均消费为 £37.20，标准差 s = £6.50。总体标准差未知。我们想在 5% 显著性水平下检验真实均值是否已增至 £35 以上。
Step 1: Set hypotheses. H₀: μ = 35; H₁: μ > 35 (one-tailed test).
第一步：建立假设。H₀: μ = 35；H₁: μ > 35（单尾检验）。
Step 2: Since σ is unknown, use t-test. The test statistic is t = (x̄ − μ₀)/(s/√n) = (37.20 − 35)/(6.50/√50) ≈ 2.20/0.9192 ≈ 2.394.
第二步：由于 σ 未知，采用 t 检验。检验统计量 t = (x̄ − μ₀)/(s/√n) = (37.20 − 35)/(6.50/√50) ≈ 2.20/0.9192 ≈ 2.394。
Step 3: Degrees of freedom = 49. The critical t-value (one-tailed, 5% tely 1.677. Since 2.394 > 1.677, the result is significant. Alternatively, p-value < 0.01.
第三步：自由度 = 49。查表得单尾 5% 临界 t 值约为 1.677。因为 2.394 > 1.677，结果显著。或者，p 值 < 0.01。
Step 4: Conclusion: There is sufficient evidence at the 5% level to reject H₀ and conclude that the mean monthly spend is greater than £ customers responded. A confidence interval can also be constructed: 37.20 ± 2.010 × (6.50/√50) gives (£35.35, £39.05), supporting an increase.
第四步：结论：在 5% 水平下有足够证据拒绝 H₀，认为月均消费大于 £35。会员卡似乎有效。然而，我们应考虑样本是否随机且具有代表性；也许只有最忠诚的顾客做出了回应。我们还可以构建置信区间：37.20 ± 2.010 × (6.50/√50) 得到 (£35.35, £39.05)，支持有所增长。
This walkthrough illustrates the complete cycle from hypothesis to contextual conclusion, exactly what CCEA case studies demand.
这个案例演示了从假设到情境结论的完整循环，这正是 CCEA 案例分析所要求的。
12. Exam Tips and Practice Strategies | 考试技巧与练习策略
To excel in CCEA statistics case studies, follow these practical tips:
- Read the entire scenario twice before writing anything. Highlight key numbers, variables, and the main question.
- Plan your approach: which statistical technique is most appropriate? Justify your choice briefly.
- Show all working clearly, labeling hypotheses, formulas, and test values. Even if your final answer is wrong, method marks can be earned.
- Always finish with a sentence that links back to the context; never leave a numerical answer floating.
- Practice with past papers under timed conditions. Create your own mini case studies from news articles or everyday situations, and apply the techniques you've learned.
- When revising, produce a summary sheet of key formulas and conditions for each test—binomial, normal, t-test, confidence interval. This builds fluency.
要在 CCEA 统计案例分析中出类拔萃，请遵循这些实用技巧：
- 在动手写任何东西之前，将整个情境读两遍。用高亮标出关键数字、变量和主要问题。
- 规划你的方法：哪种统计技术最合适？简要说明你的选择理由。
- 清晰展示所有演算过程，标注假设、公式和检验值。即使最终答案错误，也能获得方法分数。
- 总是以一句联系上下文的话作结；绝不让数字答案悬空。
- 在限时条件下练习历年真题。根据新闻文章或日常情境自创小案例分析，并应用你所学的技巧。
- 复习时，为每个检验——二项、正态、t 检验、置信区间——制作关键公式和条件摘要表。这将提高熟练度。
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