📚 A-Level CIE Statistics: Case Study in Action | A-Level CIE 统计：案例分析实战演练
This article walks you through a realistic A-Level CIE Statistics case study, from data collection to advanced inference. Using a coffee‑shop caffeine investigation, we demonstrate how descriptive statistics, probability models, confidence intervals, hypothesis tests, chi‑squared analysis, and regression come together in a single coherent project—exactly as you would tackle in your exams and real‑world research.
本文为你逐步剖析一个真实的 A-Level CIE 统计案例，从数据收集到高级推断。通过咖啡店咖啡因含量的调查，我们展示描述统计、概率模型、置信区间、假设检验、卡方分析以及回归分析如何在一个完整的项目中协同作用——正如你在考试和真实研究中需要面对的一样。
1. Background and Data Collection | 背景与数据收集
A popular coffee chain claims its large latte contains an average of 150 mg of caffeine. To verify this, a food scientist randomly selects 40 large lattes from different branches across the city and measures the caffeine content (mg) in each cup. Additional variables are recorded for each drink: preparation temperature (°C) and customer satisfaction (satisfied / not satisfied). A separate experiment compares the traditional barista method with a modern automated method using two independent samples of 20 cups each.
一家受欢迎的咖啡连锁店声称其大杯拿铁咖啡平均含有 150 毫克咖啡因。为了验证这一声称，一位食品科学家从城市不同门店随机选取 40 杯大杯拿铁，测量每杯咖啡因含量（毫克）。每杯饮品还记录了额外变量：制作温度（°C）和顾客满意度（满意 / 不满意）。另一个独立实验比较了传统咖啡师方法与现代自动化方法，各取 20 杯作为两个独立样本。
Collecting data in this structured way ensures we can explore both one‑sample and two‑sample scenarios, perform association tests, and model relationships. Good statistical practice starts with a clear sampling strategy and well‑defined variables. In CIE Statistics exams, you are often given summary statistics and asked to choose the correct inferential tool—real case studies bridge theory and application.
以这种结构化的方式收集数据，确保我们既能探索单样本和双样本情形，又能进行关联检验和关系建模。良好的统计实践始于清晰的抽样策略和定义明确的变量。在 CIE 统计考试中，你经常会拿到汇总统计量并被要求选择正确的推断工具——真实案例正好架起理论与应用的桥梁。
2. Descriptive Statistics and Visualisation | 描述统计与可视化
For the 40‑cup sample, the sample mean caffeine content is x̄ = 154.2 mg and the sample standard deviation is s = 9.8 mg. The five‑number summary (minimum, Q₁, median, Q₃, maximum) is 133, 148, 154, 160, 179 mg. A box plot reveals a roughly symmetric distribution with one mild outlier at 179 mg. A stem‑and‑leaf diagram confirms the data are unimodal and approximately bell‑shaped, supporting a normal model for further analysis.
对于 40 杯样本，样本平均咖啡因含量为 x̄ = 154.2 毫克，样本标准差为 s = 9.8 毫克。五数概括（最小值、第一四分位数、中位数、第三四分位数、最大值）为 133, 148, 154, 160, 179 毫克。箱线图显示分布大致对称，在 179 毫克处有一个温和的异常值。茎叶图证实数据是单峰且近似钟形，支持后续分析使用正态模型。
Visual summaries are the first step in any statistical inquiry. On your CIE paper, you might be asked to interpret box plots, identify skewness, or justify why a normal approximation is reasonable. Always comment on the shape, centre, spread, and any unusual observations. In this case, the median (154 mg) is close to the mean, and the interquartile range is 12 mg, indicating moderate variability.
可视化汇总是任何统计调查的第一步。在 CIE 试卷上，你可能需要解释箱线图、识别偏态，或说明为何正态近似是合理的。一定要评述形状、中心、离散程度和任何异常观测值。在这个案例中，中位数（154 毫克）接近均值，四分位距为 12 毫克，表明中等变异性。
3. Probability Distribution: Using the Normal Model | 概率分布：正态模型的应用
Given the large sample size (n = 40) and the approximately symmetric histogram, we can model the caffeine content of a randomly selected large latte using a normal distribution with μ = 154.2 mg and σ = 9.8 mg. For a single observation, the probability that a latte contains more than 170 mg of caffeine is P(X > 170). Standardising gives z = (170 − 154.2) / 9.8 = 1.612. From the standard normal table, P(Z < 1.612) ≈ 0.9463, so P(Z > 1.612) = 1 − 0.9463 = 0.0537.
鉴于样本量较大（n = 40）且直方图近似对称，我们可以用 μ = 154.2 毫克、σ = 9.8 毫克的正态分布对随机选取的大杯拿铁的咖啡因含量进行建模。对于单个观测值，一杯拿铁咖啡因含量超过 170 毫克的概率为 P(X > 170)。标准化后得 z = (170 − 154.2) / 9.8 = 1.612。查标准正态表，P(Z < 1.612) ≈ 0.9463，因此 P(Z > 1.612) = 1 − 0.9463 = 0.0537。
In CIE Statistics, you are expected to use the normal distribution after verifying conditions such as a reasonably symmetric population or large sample sizes. Here we treat the sample mean and standard deviation as estimates of the population parameters to illustrate the approach. A common exam question asks you to find the probability of an event using given μ and σ, and to interpret the result in context.
在 CIE 统计中，你需要在验证诸如总体大致对称或样本量较大等条件后使用正态分布。这里我们将样本均值和标准差作为总体参数的估计值，以演示方法。常见的考题是要求利用给定的 μ 和 σ 找出某事件的概率，并在上下文中解释结果。
z = (x − μ) / σ
4. Confidence Interval for the Mean | 均值的置信区间
Since the population standard deviation is unknown, we construct a 95 % confidence interval for the true mean caffeine content μ using the t‑distribution with 39 degrees of freedom. The standard error is s / √n = 9.8 / √40 ≈ 1.549. The critical t‑value for 39 df and a 95 % confidence level is t* ≈ 2.023. The interval is x̄ ± t* × SE = 154.2 ± 2.023 × 1.549 = (151.07, 157.33) mg.
由于总体标准差未知，我们利用自由度为 39 的 t 分布构建真实平均咖啡因含量 μ 的 95 % 置信区间。标准误为 s / √n = 9.8 / √40 ≈ 1.549。39 自由度下 95 % 置信水平的临界 t 值为 t* ≈ 2.023。区间为 x̄ ± t* × SE = 154.2 ± 2.023 × 1.549 = (151.07, 157.33) 毫克。
A confidence interval provides a range of plausible values for μ based on the sample. Since the entire interval is above 150 mg, it suggests the true mean might indeed be slightly greater than the claimed 150 mg. In CIE exams, you may be asked to interpret a confidence interval or to calculate it from summary statistics, always stating the confidence level and the method used.
置信区间根据样本给出了 μ 的合理取值范围。由于整个区间都在 150 毫克以上，表明真实均值确实可能略高于声称的 150 毫克。在 CIE 考试中，你可能会被要求解释置信区间，或根据汇总统计量计算区间，并始终注明置信水平和所使用的方法。
x̄ ± t* × (s / √n)
5. One‑Sample Hypothesis Test: Coffee Claim | 单样本假设检验：咖啡因声称
We now formally test the chain’s claim using a two‑tailed hypothesis test at the 5 % significance level. Let H₀: μ = 150 mg and H₁: μ ≠ 150 mg. The test statistic is t = (154.2 − 150) / (9.8 / √40) ≈ 2.711. With 39 degrees of freedom, the two‑tailed critical value is ±2.023. Since 2.711 > 2.023, we reject H₀. There is sufficient evidence to conclude the true mean caffeine content differs from 150 mg.
我们现在在 5 % 显著性水平下使用双尾假设检验正式检验该连锁店的声称。设 H₀: μ = 150 毫克，H₁: μ ≠ 150 毫克。检验统计量为 t = (154.2 − 150) / (9.8 / √40) ≈ 2.711。自由度为 39 时，双尾临界值为 ±2.023。由于 2.711 > 2.023，我们拒绝 H₀。有充分证据表明真实平均咖啡因含量与 150 毫克存在差异。
The p‑value for t = 2.711 with 39 df is approximately 0.010, which is less than 0.05, confirming the result. Always check assumptions: the sample is random, the data are approximately normal (supported by the box plot), and observations are independent. CIE questions typically require you to state hypotheses, calculate the test statistic, compare it to a critical value or p‑value, and write a contextual conclusion.
自由度为 39 下 t = 2.711 的 p 值约为 0.010，小于 0.05，印证了结果。务必检查假设：样本是随机的，数据近似正态（箱线图支持），且观测值相互独立。CIE 试题通常要求陈述假设、计算检验统计量、将其与临界值或 p 值比较，并写下具有上下文意义的结论。
t = (x̄ − μ₀) / (s / √n)
6. Two‑Sample t‑Test: Comparing Barista Methods | 双样本 t 检验：比较制作方法
The scientist also tests whether the traditional barista method and the automated method produce different mean caffeine levels. Sample statistics: traditional (n₁ = 20): x̄₁ = 156.5 mg, s₁ = 8.7 mg; automated (n₂ = 20): x̄₂ = 151.9 mg, s₂ = 10.2 mg. Assuming equal population variances, we first calculate the pooled variance: sₚ² = [(19 × 8.7² + 19 × 10.2²) / 38] ≈ 89.88, so sₚ ≈ 9.48.
科学家还检验了传统咖啡师方法和自动化方法是否产生不同的平均咖啡因含量。样本统计量：传统组（n₁ = 20）：x̄₁ = 156.5 毫克，s₁ = 8.7 毫克；自动化组（n₂ = 20）：x̄₂ = 151.9 毫克，s₂ = 10.2 毫克。假设总体方差相等，先计算合并方差：sₚ² = [(19 × 8.7² + 19 × 10.2²) / 38] ≈ 89.88，因此 sₚ ≈ 9.48。
The test statistic is t = (156.5 − 151.9) / [9.48 × √(1/20 + 1/20)] = 4.6 / (9.48 × 0.3162) ≈ 1.535. Degrees of freedom = 20+20−2 = 38. At the 5 % level, the two‑tailed critical t‑value is approximately 2.024. Since | 1.535 | < 2.024, we fail to reject H₀. There is insufficient evidence to claim a difference in mean caffeine content between the two methods.
检验统计量为 t = (156.5 − 151.9) / [9.48 × √(1/20 + 1/20)] = 4.6 / (9.48 × 0.3162) ≈ 1.535。自由度为 20+20−2 = 38。在 5 % 水平下，双尾临界 t 值约为 2.024。由于 | 1.535 | < 2.024，我们未能拒绝 H₀。没有足够证据表明两种方法的平均咖啡因含量存在差异。
CIE candidates should be comfortable choosing between the pooled two‑sample t‑test (equal variances) and Welch’s test (unequal variances). Here we assumed equal variances based on prior knowledge, but the large overlap in s₁ and s₂ makes this reasonable. When conducting such tests, always interpret the p‑value and relate it back to the practical context.
CIE 考生应能熟练在合并双样本 t 检验（等方差）和 Welch 检验（不等方差）之间做出选择。这里我们基于先验知识假设方差相等，但 s₁ 和 s₂ 的大幅重叠也使其合理。在进行此类检验时，始终要解释 p 值并联系实际情境。
7. Chi‑Squared Test of Independence | 卡方独立性检验
To investigate whether caffeine level is associated with customer satisfaction, we categorise the 40 lattes into ‘High’ (≥155 mg) and ‘Low’ (<155>
为了探究咖啡因水平是否与顾客满意度相关，我们将 40 杯拿铁分为“高”（≥155 毫克）和“低”（<155 毫克），并记录顾客反馈为“满意”或“不满意”。观察到的频数表如下：满意 & 高：15；满意 & 低：13；不满意 & 高：5；不满意 & 低：7。预期频数在无关联假设下计算。例如，高咖啡因且满意的预期频数为 (20×28)/40 = 14。χ² 统计量为 Σ [(O − E)² / E]。
| High | Low | Total | |
| Satisfied | 15 | 13 | 28 |
| Not Satisfied | 5 | 7 | 12 |
| Total | 20 | 20 | 40 |
Calculating χ²: cell (Satisfied, High) gives (15−14)²/14 = 0.071; (Satisfied, Low): (13−14)²/14 = 0.071; (Not Satisfied, High): (5−6)²/6 ≈ 0.167; (Not Satisfied, Low): (7−6)²/6 ≈ 0.167. Sum = 0.476. The degrees of freedom = (2−1)×(2−1) = 1. The critical value at 5 % significance is 3.841. Since 0.476 < 3.841, we fail to reject the null hypothesis of independence. Caffeine level and satisfaction are not significantly associated in this sample.
计算 χ²：单元格（满意，高）得 (15−14)²/14 = 0.071；（满意，低）：(13−14)²/14 = 0.071；（不满意，高）：(5−6)²/6 ≈ 0.167；（不满意，低）：(7−6)²/6 ≈ 0.167。总和为 0.476。自由度 = (2−1)×(2−1) = 1。5 % 显著性水平下的临界值为 3.841。由于 0.476 < 3.841，我们未能拒绝独立性原假设。在该样本中，咖啡因水平与满意度没有显著关联。
Chi‑squared tests appear frequently in CIE Statistics S2. Remember to check that all expected frequencies are at least 5 before applying the test. Write your conclusion clearly in context, stating the significance level. Even if the result is not significant, explaining what that means in real terms demonstrates strong statistical communication.
卡方检验经常出现在 CIE 统计 S2 中。记住在应用检验前检查所有预期频数是否至少为 5。在上下文中清晰地写下结论，说明显著性水平。即使结果不显著，解释其在现实中的含义也能体现强大的统计交流能力。
χ² = Σ (O − E)² / E
8. Correlation and Linear Regression | 相关性与线性回归
The scientist suspects that higher preparation temperature might reduce caffeine content. For the 40 observation pairs (temperature in °C, caffeine in mg), the product moment correlation coefficient is r = −0.78. This indicates a strong negative linear correlation. A hypothesis test for ρ = 0 against ρ ≠ 0 yields a test statistic t = r √(n−2) / √(1−r²) ≈ −0.78 × √38 / √(1−0.6084) ≈ −7.67. With 38 df, the p‑value is extremely small, so we reject H₀ and conclude there is significant linear correlation.
科学家猜测较高的制作温度可能降低咖啡因含量。对于 40 对观测值（温度 °C，咖啡因 mg），积矩相关系数为 r = −0.78。这表明存在很强的负线性相关。对 H₀: ρ = 0 且 H₁: ρ ≠ 0 的检验得到检验统计量 t = r √(n−2) / √(1−r²) ≈ −0.78 × √38 / √(1−0.6084) ≈ −7.67。自由度为 38 时，p 值极小，因此我们拒绝 H₀，并得出存在显著线性相关的结论。
The least squares regression line of caffeine on temperature can be found using the given summary statistics. Assume ∑x = 3200, ∑y = 6168, Sxx = 2400, Sxy = −936. The slope b = Sxy / Sxx = −936 / 2400 = −0.39. The intercept a = ȳ − b x̄ = 154.2 − (−0.39)×80 = 185.4. Thus, the regression equation is: caffeine = 185.4 − 0.39 × temperature. For each 1 °C increase in preparation temperature, the caffeine content is predicted to drop by 0.39 mg on average.
利用给定的汇总统计量可以找到咖啡因对温度的最小二乘回归直线。假设 ∑x = 3200，∑y = 6168，Sxx = 2400，Sxy = −936。斜率 b = Sxy / Sxx = −936 / 2400 = −0.39。截距 a = ȳ − b x̄ = 154.2 − (−0.39)×80 = 185.4。因此回归方程为：咖啡因 = 185.4 − 0.39 × 温度。制作温度每升高 1 °C，咖啡因含量平均预计降低 0.39 毫克。
Regression is a powerful tool for prediction, but CIE questions often test your ability to interpret b, calculate residuals, and discuss limitations, such as extrapolation. Here, it is sensible to use the model only within the observed temperature range (roughly 70–90 °C). Also remember that correlation does not imply causation—other factors may influence caffeine loss.
回归是预测的有力工具，但 CIE 试题经常考察解释 b 的能力、计算残差以及讨论外推等局限性。在此，仅在观测到的温度范围（约 70–90 °C）内使用该模型是合理的。同时记住，相关并不代表因果——其他因素也可能影响咖啡因的损失。
y = a + b x, b = Sxy / Sxx
9. Reporting Findings and Limitations | 报告发现与局限性
A complete statistical report should synthesise all findings: the mean caffeine content appears slightly above the claimed 150 mg, and the one‑sample t‑test supports this difference. There is no significant difference between barista methods, and caffeine level is not associated with satisfaction in the chi‑squared test. However, a strong negative linear relationship exists with preparation temperature. Limitations include the small sample size for the two‑sample comparison, potential confounding variables (e.g., bean origin, extraction time), and the assumption of normality.
一份完整的统计报告应综合所有发现：平均咖啡因含量似乎略高于声称的 150 毫克，且单样本 t 检验支持这一差异。咖啡师方法之间无显著差异，卡方检验中咖啡因水平与满意度无关联。然而，咖啡因含量与制作温度存在很强的负线性关系。局限性包括双样本比较的样本量较小、潜在的混杂变量（如咖啡豆产地、萃取时间）以及正态性假设。
When writing conclusions in CIE exams, always refer back to the context, use non‑technical language for interpretation, and mention any assumptions or possible errors. A good statistician acknowledges uncertainty and suggests further investigation, such as increasing sample sizes or conducting controlled experiments. This holistic approach earns high marks in the assessment objectives involving communication and evaluation.
在 CIE 考试中书写结论时，始终要回到具体情境，用非技术性语言进行解释，并提及任何假设或可能的误差。一位好的统计学家会承认不确定性，并建议进一步研究，如增大样本量或进行对照实验。这种整体性方法能为涉及交流与评价的评估目标赢得高分。
10. Conclusion and Exam Tips | 结论与考试技巧
This case study illustrates how a single research question can connect the entire CIE Statistics syllabus. From drawing box plots to performing t‑tests and chi‑squared tests, each step requires careful choice of method and clear justification. When you encounter an exam problem, read the question carefully to identify the variable type, sample design, and the inference required. Sketch diagrams, write hypotheses, and always state your conclusion in context.
本案例展示了一个研究问题如何将整个 CIE 统计课程大纲连接起来。从绘制箱线图到进行 t 检验和卡方检验，每一步都需要谨慎选择方法并清晰论证。当你遇到考试题目时，仔细阅读以识别变量类型、样本设计和所需的推断。画出示意图，写出假设，并始终在情境中陈述你的结论。
Practise with past paper case studies and ensure you are fluent in both the calculations and the written explanations. Remember to manage your time wisely and check small but critical details: are you using the correct table (z or t)? Did you account for degrees of freedom? Is your conclusion aligned with the rejection region or p‑value? Master these elements, and you will be well prepared for the CIE Statistics exam.
通过历年试卷中的案例分析进行练习，确保对计算和书面解释都驾轻就熟。记住合理分配时间，并检查细小但关键的细节：你是否使用了正确的表格（z 还是 t）？是否考虑了自由度？你的结论是否与拒绝域或 p 值一致？掌握这些要素，你将为 CIE 统计考试做好充分准备。
Published by TutorHao | 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Cancel reply