A-Level CIE Statistics: Interdisciplinary Integrated Practice | A-Level CIE 统计：跨学科综合题型训练

📚 A-Level CIE Statistics: Interdisciplinary Integrated Practice | A-Level CIE 统计：跨学科综合题型训练

Taking A-Level Statistics beyond textbook calculations, this article explores how core statistical methods such as hypothesis testing, regression, and probability distributions are applied across biology, economics, physics, and other disciplines. You will encounter typical interdisciplinary integrated problems, learn to identify the appropriate statistical tool, and practise structured problem-solving – exactly what CIE exam papers increasingly demand.

本文将 A-Level 统计从课本计算延伸至实际应用，探讨假设检验、回归和概率分布等核心方法如何跨越生物学、经济学、物理学等学科。你将面对典型的跨学科综合题型，学会选择合适的统计工具，并进行结构化的解题训练——这正是 CIE 考试日益注重的能力。

1. Biology: Chi-Squared Test for Genetic Ratios | 生物学：遗传比例的卡方检验

In Mendelian genetics, a dihybrid cross predicts a 9:3:3:1 phenotypic ratio. A chi-squared goodness-of-fit test compares observed counts of four phenotypes against the expected frequencies. For a total of 556 offspring, the expected counts would be 312.75, 104.25, 104.25 and 34.75 respectively. The test statistic χ² = Σ (O – E)² / E is computed and compared with a critical value from the chi-squared distribution with 3 degrees of freedom at the 5% significance level. A p-value less than 0.05 suggests the observed ratios significantly deviate from Mendelian expectations, possibly indicating linkage or experimental error.

在孟德尔遗传学中，双因子杂交预测表型比例为 9:3:3:1。卡方适合度检验将四种表型的观测频数与期望频率进行比较。对于总计 556 个后代，期望频数分别为 312.75、104.25、104.25 和 34.75。计算检验统计量 χ² = Σ (O – E)² / E，并将其与自由度为 3、显著性水平 5% 的卡方分布临界值比较。若 p 值小于 0.05，表明观测比值显著偏离孟德尔预期，可能暗示基因连锁或实验误差。

A typical exam question might present observed counts of 315 round-yellow, 108 round-green, 101 wrinkled-yellow and 32 wrinkled-green. Students must state hypotheses, calculate χ², determine the degrees of freedom, and draw a conclusion. A small Yates correction may be required when any expected frequency falls below 5.

典型的考题可能给出观测频数：315 粒圆黄、108 粒圆绿、101 粒皱黄和 32 粒皱绿。学生需要陈述原假设与备择假设，计算 χ² 值，确定自由度，并得出结论。当任一期望频数低于 5 时，可能需要进行耶茨校正。

χ² = Σ (O – E)² / E

Phenotype	Observed (O)	Expected (E)	(O–E)²/E
Round-Yellow	315	312.75	0.016
Round-Green	108	104.25	0.135
Wrinkled-Yellow	101	104.25	0.101
Wrinkled-Green	32	34.75	0.218

Summing the final column gives χ² ≈ 0.470. Since 0.470 < 7.815 (critical value for ν = 3 at 5%), we do not reject the null hypothesis. The observed ratios are consistent with the 9:3:3:1 model.

将最后一列求和得 χ² ≈ 0.470。由于 0.470 < 7.815（ν = 3 且 α = 0.05 的临界值），我们不拒绝原假设，观测比值与 9:3:3:1 模型一致。

2. Economics: Index Numbers and Time Series | 经济学：指数与时间序列

Economic statistics frequently involve constructing a chain of Laspeyres price index or analysing quarterly GDP data. A CIE integrated task may ask you to calculate a weighted index, compute a four-point moving average, and then isolate seasonal variation using the additive model Y = T + S + R.

经济统计常常涉及构建拉氏价格指数链或分析季度 GDP 数据。一道 CIE 综合题可能会要求计算加权指数，求出四点移动平均，并使用加法模型 Y = T + S + R 分离季节变动。

Consider a company’s quarterly revenue (in £1000): Q1 42, Q2 56, Q3 68, Q4 50, Q1 44, Q2 58, Q3 71, Q4 53. The four-point centred moving average smooths the trend. For the first two years, the moving average for Q3 of year 1 is (42+56+68+50)/4 = 54, and for Q4 it is (56+68+50+44)/4 = 54.5, hence the centred value between them is 54.25. Seasonal variation for a quarter is then the actual value minus the centred moving average.

某公司季度收入（单位：千英镑）：Q1 42, Q2 56, Q3 68, Q4 50, Q1 44, Q2 58, Q3 71, Q4 53。四点中心移动平均可平滑趋势。前两年的计算中，第一年 Q3 的移动平均为 (42+56+68+50)/4 = 54，第四季度为 (56+68+50+44)/4 = 54.5，因此二者中心化值为 54.25。某季度的季节变动 = 实际值 – 中心移动平均值。

An index number problem often provides prices and quantities for a base year and a current year. The Laspeyres price index is: (Σ pₙ q₀ / Σ p₀ q₀) × 100. Interdisciplinary links appear when comparing inflation across countries or evaluating real wage growth.

指数问题通常给出基年和当年的价格与数量。拉氏价格指数为 (Σ pₙ q₀ / Σ p₀ q₀) × 100。在比较各国通胀率或评估实际工资增长时，统计与其他学科自然交叉。

3. Physics: Measurement Errors and Normal Distribution | 物理学：测量误差与正态分布

Repeated measurements of a physical quantity, such as the acceleration due to gravity g, are subject to random errors. Students can model the readings as a normal distribution with mean µ and standard deviation σ. A one-sample t-test determines whether the experimental mean differs significantly from the accepted value 9.81 m s⁻².

对重力加速度 g 等物理量的重复测量会带有随机误差。学生可将读数建模为均值 µ、标准差 σ 的正态分布，并使用单样本 t 检验判断实验均值是否与公认值 9.81 m s⁻² 存在显著差异。

Given five readings: 9.79, 9.82, 9.85, 9.77, 9.80. The sample mean is 9.806 and the sample standard deviation s = √[Σ(x – x̄)²/(n–1)] ≈ 0.0311. The test statistic t = (x̄ – 9.81) / (s/√n) = (9.806 – 9.81) / (0.0311/√5) ≈ –0.287. With 4 degrees of freedom, the two-tailed critical value at the 5% level is 2.776. Since |t| < 2.776, we conclude there is no significant difference; systematic errors are likely absent.

假设五次读数为：9.79, 9.82, 9.85, 9.77, 9.80。样本均值 x̄ = 9.806，样本标准差 s = √[Σ(x – x̄)²/(n–1)] ≈ 0.0311。检验统计量 t = (x̄ – 9.81) / (s/√n) = (9.806 – 9.81) / (0.0311/√5) ≈ –0.287。自由度为 4 时，5% 水平下的双尾临界值为 2.776。由于 |t| < 2.776，结论是无显著差异，可能不存在系统误差。

Interdisciplinary challenges may also ask you to calculate a 95% confidence interval for µ using the t-distribution: x̄ ± t_crit × (s/√n). This links directly to laboratory reports in physics.

跨学科挑战还可能要求使用 t 分布计算 µ 的 95% 置信区间：x̄ ± t_crit × (s/√n)，这与物理实验报告直接相关。

4. Geography: River Discharge and Probability Distributions | 地理学：河流流量与概率分布

Hydrologists often model annual maximum river discharge with a log-normal or normal distribution. A CIE question might state that the annual maximum discharge (in m³ s⁻¹) follows N(850, 220²). You are asked to find the probability that the discharge exceeds 1200 m³ s⁻¹ in a given year, or to estimate the recurrence interval of a flood of that magnitude.

水文学家常用对数正态或正态分布模拟年最大河流流量。一道 CIE 题目可能会设定年最大流量（立方米每秒）服从 N(850, 220²)，要求计算某年流量超过 1200 m³ s⁻¹ 的概率，或估算如此规模洪水的重现期。

Standardising: Z = (1200 – 850) / 220 ≈ 1.591. Using the standard normal table, P(Z > 1.591) = 1 – Φ(1.591) ≈ 0.0557. Hence the return period is 1 / 0.0557 ≈ 17.9 years. Decision-makers can use this to design flood defences.

标准化：Z = (1200 – 850) / 220 ≈ 1.591。查标准正态表得 P(Z > 1.591) = 1 – Φ(1.591) ≈ 0.0557，故重现期约为 17.9 年。决策者可据此设计防洪工程。

Geographical data often exhibit skewness; students might need to apply a log transformation to achieve normality before performing a t-test or regression. This reinforces the idea that checking distributional assumptions is vital in applied statistics.

地理数据常呈现偏态；学生可能需要先进行对数变换以获得正态性，然后再进行 t 检验或回归。这强化了应用统计中检查分布假设至关重要的理念。

5. Psychology: Correlation and Regression for Test Scores | 心理学：测试分数的相关与回归

Psychological research frequently examines the relationship between variables such as hours of therapy and anxiety score reduction. Pearson’s product-moment correlation coefficient r quantifies the linear association, while least-squares regression allows prediction. The null hypothesis H₀: ρ = 0 can be tested using t = r √[(n–2)/(1–r²)].

心理学研究经常考察变量间的关系，如治疗时数与焦虑分数降低之间的关系。皮尔逊积矩相关系数 r 量化线性关联强度，而最小二乘回归可用于预测。原假设 H₀: ρ = 0 可通过 t = r √[(n–2)/(1–r²)] 进行检验。

Suppose ten patients provided data, and the computed r = 0.68. The test statistic t = 0.68 × √[(10–2)/(1–0.68²)] = 0.68 × √(8/0.5376) ≈ 0.68 × 3.858 ≈ 2.623. With 8 degrees of freedom, the two-tailed critical value at 1% significance is 3.355, but at 5% it is 2.306. Thus, r is significantly different from 0 at the 5% level, indicating a positive linear relationship.

假设十名患者的数据给出 r = 0.68。检验统计量 t = 0.68 × √[(10–2)/(1–0.68²)] = 0.68 × √(8/0.5376) ≈ 0.68 × 3.858 ≈ 2.623。自由度为 8 时，1% 显著水平的双尾临界值为 3.355，而 5% 水平临界值为 2.306。因此，在 5% 水平下 r 显著异于 0，表明存在正向线性关系。

Regression analysis then provides the equation y = a + b x, where b = r (s_y / s_x) and a = ȳ – b x̄. Predicting a patient’s improvement from therapy length becomes possible, but exam questions frequently ask about the limitations of extrapolation.

回归分析随后给出方程 y = a + b x，其中 b = r (s_y / s_x)，a = ȳ – b x̄。由此可基于治疗时长预测患者的改善程度，但试题常会追问外推的局限性。

6. Business: Sampling and Confidence Intervals for Market Research | 商业：市场调研中的抽样与置信区间

Market researchers survey a sample of customers to estimate the proportion p who prefer a new product design. A simple random sample of 200 customers reveals 120 who prefer the new design. The sample proportion p̂ = 0.6. An approximate 95% confidence interval for p is p̂ ± 1.96 √[p̂(1–p̂)/n].

市场研究人员抽样调查顾客，以估计喜欢新设计产品的比例 p。某简单随机样本 200 名顾客中，120 人偏好新设计，p̂ = 0.6。p 的近似 95% 置信区间为 p̂ ± 1.96 √[p̂(1–p̂)/n]。

Calculation: standard error = √(0.6 × 0.4 / 200) ≈ √0.0012 ≈ 0.03464. Margin of error = 1.96 × 0.03464 ≈ 0.0679. The interval is (0.532, 0.668). The company can be 95% confident that the true proportion lies between 53.2% and 66.8%.

计算：标准误 = √(0.6 × 0.4 / 200) ≈ √0.0012 ≈ 0.03464，误差范围 = 1.96 × 0.03464 ≈ 0.0679。置信区间为 (0.532, 0.668)。公司可以有 95% 的把握认为真实比例介于 53.2% 至 66.8% 之间。

Questions may extend to determining the sample size required to achieve a desired margin of error: n = (1.96² × p̂(1–p̂)) / E². Interdisciplinary integration emerges when linking these intervals to business decisions about product launch or stock management.

试题可能扩展为计算达到期望误差范围所需的样本量：n = (1.96² × p̂(1–p̂)) / E²。当将这些置信区间与产品发布或库存管理等商业决策挂钩时，跨学科综合特性便体现出来。

7. Sports Science: Hypothesis Testing for Performance Data | 体育科学：表现数据的假设检验

Coaches often compare athletes’ times before and after a training programme. A paired samples t-test evaluates whether the mean difference d̄ is significantly different from zero. This test reduces inter-athlete variability because each athlete serves as their own control.

教练常比较训练计划前后运动员的成绩。配对样本 t 检验评估平均差值 d̄ 是否显著异于零。该检验因每位运动员自身作为对照而减少了运动员间的变异。

Suppose eight sprinters ran 100 m times (in seconds) before and after an eight-week strength programme. The differences (after – before) are –0.05, –0.12, –0.03, 0.02, –0.08, –0.10, –0.06, –0.04. A negative mean indicates improvement. The sample mean difference d̄ = –0.0575 and s_d = 0.0462. The test statistic t = d̄ / (s_d/√n) = –0.0575 / (0.0462/√8) ≈ –3.52. With 7 degrees of freedom, the one-tailed p-value at 5% has a critical value of –1.895. Since –3.52 < –1.895, we reject H₀ and conclude the training significantly improved times.

假设八名短跑运动员在八周力量训练前后的百公尺成绩（秒）差值（后减前）为：–0.05, –0.12, –0.03, 0.02, –0.08, –0.10, –0.06, –0.04。负值表示进步。样本均值 d̄ = –0.0575，s_d = 0.0462。检验统计量 t = d̄ / (s_d/√n) = –0.0575 / (0.0462/√8) ≈ –3.52。自由度为 7 时，5% 单尾临界值为 –1.895。由于 –3.52 < –1.895，拒绝 H₀，认为训练显著提高了成绩。

The sports science context requires careful consideration of whether a one-tailed test is justified and whether normality of differences holds. Boxplots and normal probability plots are often used to check assumptions.

体育科学的情景需要仔细考量是否适合使用单尾检验以及差值是否满足正态性。常借助箱线图和正态概率图来检验假定。

8. Environmental Science: Poisson Distribution for Rare Events | 环境科学：稀有事件的泊松分布

Environmental scientists monitor the occurrence of rare species or pollution incidents. The Poisson distribution models the number of events in a fixed interval when events occur independently at a constant rate λ. For example, the number of illegal waste dumping incidents per month might follow Poisson(3.2).

环境科学家监测稀有物种出现次数或污染事件。当事件独立发生且平均发生率 λ 恒定时，泊松分布可模拟固定区间内的事件数。例如，每月非法倾倒废物次数可能服从参数为 3.2 的泊松分布。

The probability of exactly k events is P(X = k) = (e^{–λ} × λ^k) / k!. If λ = 3.2, the probability of exactly 2 incidents in a month is (e^{–3.2} × 3.2²) / 2! ≈ (0.0408 × 10.24) / 2 ≈ 0.209. Environmental agencies use such probabilities to allocate inspection resources or to set alert thresholds.

恰好发生 k 次事件的概率为 P(X = k) = (e^{–λ} × λ^k) / k!。若 λ = 3.2，一个月内恰好发生 2 次事件的概率为 (e^{–3.2} × 3.2²) / 2! ≈ (0.0408 × 10.24) / 2 ≈ 0.209。环保机构利用此类概率来分配检查资源或设定预警阈值。

A chi-squared goodness-of-fit test can compare observed monthly incident counts with Poisson expected frequencies, checking whether the Poisson assumption is reasonable. Such questions strengthen the linkage between discrete probability models and real-world environmental data.

卡方适合度检验可将每月事故观测频数与泊松分布期望频数对比，以检查泊松假定的合理性。这类问题加强了离散概率模型与现实环境数据的联系。

9. Medicine: Two-Sample t-Test for Drug Efficacy | 医学：药物疗效的双样本 t 检验

Clinical trials compare a new drug against a placebo. The two-sample t-test assesses whether the true means of the two groups differ significantly. When population variances are assumed equal, the pooled sample variance is s_p² = [(n₁–1)s₁² + (n₂–1)s₂²] / (n₁ + n₂ – 2), and the test statistic is t = (x̄₁ – x̄₂) / (s_p √(1/n₁ + 1/n₂)).

临床试验将新药与安慰剂进行比较。双样本 t 检验评估两组的真实均值是否存在显著差异。假定总体方差相等时，合并样本方差为 s_p² = [(n₁–1)s₁² + (n₂–1)s₂²] / (n₁ + n₂ – 2)，检验统计量 t = (x̄₁ – x̄₂) / (s_p √(1/n₁ + 1/n₂))。

Suppose 12 patients taking the drug show a mean improvement score of 18.5 with s₁ = 4.2, while 10 placebo patients have mean 13.1 with s₂ = 3.8. Then s_p² = (11×4.2² + 9×3.8²) / 20 = (194.04 + 129.96)/20 = 16.2, s_p = 4.025. The t-value is (18.5 – 13.1) / (4.025 × √(1/12 + 1/10)) = 5.4 / (4.025 × 0.428) ≈ 3.13. With 20 degrees of freedom, the critical t at 1% two-tailed is 2.845. Since 3.13 > 2.845, we reject the null hypothesis of no difference. The drug appears effective.

假设 12 名用药患者改善评分均值 18.5，s₁ = 4.2；10 名安慰剂患者均值 13.1，s₂ = 3.8。计算得 s_p² = (11×4.2² + 9×3.8²) / 20 = 16.2，s_p = 4.025。t 值 = (18.5 – 13.1) / (4.025 × √(1/12 + 1/10)) ≈ 3.13。自由度为 20 时，1% 双尾临界 t 值为 2.845。因 3.13 > 2.845，拒绝无差异原假设，药物似乎有效。

Medical statistics also involves calculating confidence intervals for the difference of means. The 95% CI is (x̄₁ – x̄₂) ± t_crit × s_p √(1/n₁ + 1/n₂). This helps communicate the clinical significance beyond mere statistical significance.

医学统计还涉及均值差异的置信区间。95% 置信区间为 (x̄₁ – x̄₂) ± t_crit × s_p √(1/n₁ + 1/n₂)。这有助于在统计显著性之外传递临床意义。

10. Engineering: Reliability and Exponential Distribution | 工程学：可靠性与指数分布

In reliability engineering, the lifetime of electronic components is often modelled by the exponential distribution with probability density f(x) = λ e^{–λx} for x ≥ 0, where λ is the failure rate. The probability that a component survives beyond time t is the reliability function R(t) = e^{–λt}. The mean time to failure is 1/λ.

在可靠性工程中，电子元件的寿命常采用指数分布建模，其概率密度为 f(x) = λ e^{–λx} (x ≥ 0)，其中 λ 为失效率。元件存活超过时间 t 的概率由可靠度函数 R(t) = e^{–λt} 给出，平均故障时间为 1/λ。

If a microprocessor has a constant failure rate of 0.0004 per hour, the probability it operates without failure for 5000 hours is R(5000) = e^{–0.0004×5000} = e^{–2} ≈ 0.1353. Engineers might use this to determine warranty periods or maintenance schedules.

若微处理器恒定失效率为每小时 0.0004，则其无故障运行 5000 小时的概率为 R(5000) = e^{–0.0004×5000} = e^{–2} ≈ 0.1353。工程师可利用此信息确定质保期限或维修排程。

CIE problems may extend to testing whether observed failure times fit an exponential distribution using a chi-squared goodness-of-fit test, or to comparing the reliability of two different designs with a likelihood ratio test. The mathematical foundations of PDFs and CDFs are thus applied directly to safety-critical systems.

CIE 试题可能延伸至使用卡方拟合优度检验判断观测故障时间是否服从指数分布，或通过似然比检验比较两种设计的可靠性。概率密度函数与累积分布函数的数学基础由此直接应用于安全关键系统。

Published by TutorHao | Statistics Revision Series | aleveler.com

更多咨询请联系16621398022（同微信）

A-Level CIE Statistics: Interdisciplinary Integrated Practice | A-Level CIE 统计：跨学科综合题型训练