📚 AS Edexcel Statistics: Cross-disciplinary Integrated Problem Solving | AS Edexcel 统计：跨学科综合题型训练
This article helps you tackle the cross-disciplinary questions that frequently appear in AS Edexcel Statistics. We explore how to apply data summary, probability, binomial models and hypothesis tests in real-world contexts such as medicine, manufacturing and environmental science. By seeing the statistics embedded in authentic problems, you will build the flexible thinking required for top marks on integrated exam questions.
本文帮你攻克 AS Edexcel 统计中常见的跨学科题目。我们将展示如何把数据概括、概率、二项模型和假设检验应用到医学、制造业和环境科学等真实情境中。通过观察统计方法在真实问题中的嵌入方式，你将培养出应对综合考题所需的那份灵活思维。
1. Data Processing in Environmental Science | 环境科学中的数据处理
Environmental agencies monitor air pollution by recording concentrations of a pollutant every hour. A random sample of 10 readings (μg/m³) gave: 42, 48, 51, 47, 55, 44, 49, 53, 46, 50. The first statistical task is to summarise this data using measures of location and spread.
环境机构通过每小时记录一种污染物的浓度来监测空气质量。一个包含10个读数（μg/m³）的随机样本为：42, 48, 51, 47, 55, 44, 49, 53, 46, 50。首要的统计任务是用位置和分散度量概括这批数据。
Calculate the mean: x̄ 42+48+51+47+55+44+49+53+46+50) ÷ 10 = 48.5 μg/m³. Use this to find the standard deviation.
计算均值：x̄ = (42+48+51+47+55+44+49+53+46+50) ÷ 10 = 48.5 μg/m³。以此为基础计算标准差。
-0.5, ... square each, sum them, divide by n-1=9 and square root. The sample variance s² = .25 + 20.25 + 0.25 + 20.25 + 6.25 + 2.25) ÷ 9] = 142.5 ÷ 9 = 15.833..., so s ≈ 3.98 μg/m³.
计算离差：(42-48.5)=-6.5, (48-48.5)=-0.5, ...将每个平方，求和，除以 n-1=9 再开方。样本方差 s² = [(…)÷9] = 142.5 ÷ 9 = 15.833…，故 s ≈ 3.98 μg/m³。
Interpretation: The mean concentration is close to a safety threshold of 50 μg/m³, but with a standard deviation of about 4, many hours could exceed it. This demands further probability modelling.
解读：平均浓度接近安全阈值 50 μg/m³，但标准差约为 4，许多小时的读数可能超标。这要求进一步进行概率建模。
When data is presented in a grouped frequency table from a larger survey, use midpoints and formulas adapted for grouped data. This technique appears in ecology when measuring tree heights or soil pH levels across different plots.
当数据以分组频数表形式来自更大规模调查时，要使用组中值和适用于分组数据的公式。这种技巧在生态学中测量不同样方的树高或土壤 pH 值时会出现。
2. Probability and Medical Risk Communication | 概率与医疗风险沟通
A diagnostic test for a virus has a sensitivity of 95% (true positive rate) and a specificity of 90% (true negative rate). The prevalence of the virus in a certain region is 2%. A patient tests positive and wants to know the probability they actually have the disease. This is a classic cross-disciplinary problem linking probability with public health.
某种病毒的诊断测试灵敏度（真阳性率）为 95%，特异度（真阴性率）为 90%。某地区该病毒流行率为 2%。一位患者检测呈阳性，想知道他确实患病的概率。这是一个经典的跨学科问题，将概率与公共卫生联系起来。
Define events: D = has disease, + = positive test. P(D)=0.02, P(+|D)=0.95, P(+|not D)=0.10. Use Bayes' theorem or a tree diagram.
定义事件：D = 患病，+ = 检测阳性。P(D)=0.02，P(+|D)=0.95，P(+|not D)=0.10。使用贝叶斯定理或树状图。
P(+) = P(D)×P(+|D) + P(not D)×P(+|not D) = 0.02×0.95 + 0.98×0.10 = 0.019 + 0.098 = 0.117. Then P(D|+) = (0.019)/(0.117) ≈ 0.162, or 16.2%.
P(+) = 0.02×0.95 + 0.98×0.10 = 0.019 + 0.098 = 0.117。则 P(D|+) = 0.019/0.117 ≈ 0.162，即 16.2%。
Interpretation: Even with high sensitivity, the low prevalence means that only about 1 in 6 positive results are true positives. This demonstrates why statistical literacy matters when communicating medical risks to patients.
解读：即便灵敏度高，低流行率意味着约每 6 个阳性结果中只有 1 个是真阳性。这说明了向患者沟通医疗风险时统计素养的重要性。
AS students often meet such probability questions framed as 'test accuracy'. Practice drawing a two-way table or tree with labelled branches to organise the information systematically.
AS 学生经常遇到这种被称为“测试准确性”的概率问题。练习绘制双向表格或标注分支的树状图，以系统地整理信息。
3. Binomial Distribution and Quality Sampling in Manufacturing | 二项分布与制造业质量抽样
A factory produces LED bulbs, and historically 5% are defective. A quality inspector randomly selects 20 bulbs and counts how many are faulty. The binomial model X ~ B(20, 0.05) describes the number of defectives, because each trial is independent with constant probability.
某工厂生产 LED 灯泡，以往有 5% 存在缺陷。质检员随机抽取 20 个灯泡并统计故障数量。二项模型 X ~ B(20, 0.05) 描述了缺陷品数量，因为每次试验独立且概率恒定。
should be accepted. Calculate P(X ≤ 2) = P(X=0)+P(X=1)+P(X=2). Use formula: P(X=k) = ⁿCₖ pᵏ (1-p)ⁿ⁻ᵏ.
跨学科应用：检验员想知道发现不超过 2 个故障灯泡的概率，这有助于决定是否应当接收该批次。计算 P(X ≤ 2) = P(X=0)+P(X=1)+P(X=2)。使用公式：P(X=k) = ⁿCₖ pᵏ (1-p)ⁿ⁻ᵏ。
| k | ²⁰Cₖ | pᵏ(1-p)²⁰⁻ᵏ | Probability |
|---|---|---|---|
| 0 | 1 | 1 × (0.95)²⁰ ≈ 0.3585 | 0.3585 |
| 1 | 20 | 20 × 0.05 × (0.95)¹⁹ ≈ 0.3774 | 0.3774 |
| 2 | 190 | 190×0.05²×0.95¹⁸ ≈ 0.1887 | 0.1887 |
Total P(X ≤ 2) ≈ 0.3585+0.3774+0.1887 = 0.9246. So there is a 92.5% chance of observing 2 or fewer defectives when the process is at 5% defective. This calculation supports statistical process control in engineering.
P(X ≤ 2) ≈ 0.9246。 因此，当缺陷率为 5% 时，观察到不超过 2 个缺陷品的概率为 92.5%。这个计算为工程中的统计过程控制提供支持。
4. Hypothesis Testing and Drug Trials | 假设检验与药物试验
In a medical trial, a new drug claims to reduce the proportion of patients experiencing side effects from the current rate of 40%. Out of 25 patients, only 5 reported side effects. Test at the 5% significance level whether there is evidence that the drug reduces side effects.
在一项医学试验中，一种新药声称能降低患者出现副作用的比例，原来的比例为 40%。在 25 名患者中，只有 5 人报告了副作用。以 5% 显著性水平检验是否有证据表明该药物降低了副作用发生率。
Define X = number with side effects; under null hypothesis X ~ B(25, 0.4). H₀: p = 0.4, H₁: p < 0.4 (one-tailed). Observed x = 5.
定义 X = 出现副作用的患者数；原假设下 X ~ B(25, 0.4)。H₀: p = 0.4，H₁: p < 0.4（单侧）。观测值 x = 5。
formula. P(X=0)= (0.6)²⁵ ≈ very small; using binomial CDF: P(X ≤ 5) ≈ 0.029 (from tables).
使用二项分布表或累积公式计算 P(X ≤ 5)。P(X ≤ 5) 约为 0.029。
Since 0.029 < 0.05, the result is significant. Reject H₀. There is sufficient evidence at the 5% level to suggest the drug reduces the side-effect rate.
因为 0.029 < 0.05，结果显著。拒绝 H₀。在 5% 显著性水平下有足够证据表明该药物降低了副作用率。
Always state conclusions in context: 'The probability of observing 5 or fewer side effects if the drug were ineffective is 2.9%, so we have reason to believe the drug is effective.' Cross-disciplinary tasks often demand clear, non-technical communication of the test outcome.
始终在情境中陈述结论：“如果药物无效，观察到 5 个或更少副作用病例的概率为 2.9%，因此我们有理由相信该药物有效。”跨学科任务经常要求对检验结果进行清晰、非技术性的传达。
5. Data Representation and Business Decision Making | 数据表示与商业决策
A retail company records weekly sales (£000s) over a quarter. Data visualisation helps managers identify trends. In AS statistics, you must be able to construct and interpret cumulative frequency diagrams, box plots and histograms in a business context.
某零售公司记录了一个季度的每周销售额（千英镑）。数据可视化帮助管理者识别趋势。在 AS 统计中，你必须能够在商业情境下制作并解读累积频数图、箱线图和直方图。
= 25.25, median = 27.5, Q₃ = 30.75, maximum = 33. From this, draw a box plot that reveals a slight positive skew and no outliers. Businesses use such diagrams to compare performance across quarters.
例子：12 周的销售额：23, 27, 26, 30, 31, 25, 28, 33, 29, 24, 32, 26。重要的是给出五数概括：最小值 = 23，Q₁ = 25.25，中位数 = 27.5，Q₃ = 30.75，最大值 = 33。据此绘制箱线图，可看出轻微正偏且无异常值。企业用这类图来比较不同季度的表现。
uency ÷ class width. This ensures area represents frequency, a concept that integrates mathematics with practical data handling.
当数据连续时，不等组距的直方图很常见。记住：频数密度 = 频数 ÷ 组距。这样确保面积代表频数，这一概念将数学与实际数据处理相结合。
6. Correlation and Social Science Research | 相关性与社会科学研究
ting a positive linear correlation. The product moment correlation coefficient (PMCC) for this data was calculated as r = 0.82.
社会学家研究每周学习时间与考试成绩之间的关系。15 个学生的散点图呈现出一种正线性相关的模式。计算出该数据的积矩相关系数（PMCC）为 r = 0.82。
on: r = 0.82 indicates a strong positive correlation. However, it does not imply causation — luence both. AS exam questions often ask you to comment on the reliability of a claim that 'more study causes higher marks'.
解读：r = 0.82 表明强正相关。但它并不暗示因果关系——这是跨学科研究中一个典型的警示。其他变量，如先前知识或社会经济因素，可能同时影响两者。AS 考题常常要求你就“更多学习导致更高分数”这一说法的可靠性发表评论。
You might also need to explain the effect of outliers. Removing a data point where a student studied very little but scored highly could increase r further, as it deviates from the general trend. Always support such remarks with the concept of correlation vs. causation.
你可能还需要解释异常值的影响。若移除某个学习时间很少却得分很高的学生数据点，r 可能会进一步增大，因为它偏离了总体趋势。务必用相关与因果关系的概念支撑这些评论。
7. Sampling Methods and Ecological Surveys | 抽样方法与生态调查
Ecologists estimate the population of a rare plant in a large forest. They divide the area into 50 equally sized quadrats and use simple random sampling to select 8 quadrats. This is an application of random sampling techniques taught in AS.
生态学家估算一片大森林中一种稀有植物的种群数量。他们将区域划分为 50 个等大小的样方，并用简单随机抽样选择 8 个样方。这是 AS 所教随机抽样技术的一种应用。
Why random? It minimises bias and allows probability theory to calculate confidence in estimates. In practice, they might use stratified sampling by habitat type. In an exam question, you would need to describe the advantages of one method over another and compare the likely accuracy in different terrains.
为什么随机？它最小化偏差，并能用概率论计算估计的信度。在实践中，他们可能会按栖息地类型进行分层抽样。在考试题中，你需要描述一种方法相对于另一种的优势，并比较在不同地形中可能的准确性。
A sampling frame may be a map with numbered quadrats. When sampling is complete, the mean plant count per quadrat can be scaled up to produce a population estimate. The SE of the mean can then reflect survey precision — linking sampling distributions to ecology.
抽样框可能是一张带有编号样方的地图。抽样完成后，每样方平均植物数量可按比例扩大以生成种群估计值。均值的标准误可以反映调查精度——从而将抽样分布与生态学联系起来。
8. Integrated Problem: Combining Multiple Statistical Concepts | 综合题型：整合多个统计概念
The strongest AS exam questions weave together two or more topics. For instance: a study on a health intervention records before-and-after data, uses a binomial hypothesis test for success rates, and then asks for a critical evaluation of the data collection method.
最强劲的 AS 考题会将两三个主题编织在一起。例如：一项健康干预研究记录了干预前后数据，使用二项假设检验检验成功率，然后要求对数据收集方法进行批判性评估。
Example scenario: A traffic safety campaign aims to reduce the proportion of speeding drivers at a junction from a previous figure of 30%. After the campaign, a random sample of 40 drivers showed 7 speeding. Test at 5% significance. Additionally, the sample was only taken on weekdays — discuss limitations.
示例情境：某交通安全宣传活动旨在将一个路口超速司机的比例从此前的 30% 降下来。活动后，一个包含 40 名司机的随机样本显示 7 人超速。以 5% 显著性进行检验。此外，该样本只包含工作日——讨论其局限性。
Perform test: X ~ 05, not significant. Do not reject H₀; insufficient evidence that campaign lowered speeding.
执行检验：X ~ B(40, 0.3), H₁: p < 0.3。观测值 7。P(X ≤ 7) ≈ 0.072（查表）。0.072 > 0.05，不显著。不拒绝 H₀；没有足够证据表明宣传活动降低了超速率。
Critique sampling: Weekday data may not represent weekend behaviour; the sample is not fully representative. Thus, the test conclusion may be invalid due to a biased sample.
批评抽样：工作日数据可能不代表周末行为；样本不完全具有代表性。因此，由于样本存在偏差，检验结论可能无效。
Such integrated problems mirror real-world statistical thinking, where methodology, probability and context always intertwine.
这类综合问题反映了现实世界中的统计思维，方法、概率和情境总是交织在一起。
9. Using Large Data Sets in Cross-disciplinary Contexts | 在大规模数据集中跨学科应用
Edexcel AS specifications often reference a large data set (LDS) related to a specific theme, such as weather or transport. You need to extract subsets, calculate summary statistics, and compare distributions. For instance, comparing daily mean temperatures in a coastal town versus an inland city over several months.
Edexcel AS 考试大纲常会引用一个与特定主题（如天气或交通）相关的大规模数据集。你需要提取子集、计算概括统计量并比较分布。例如，比较一个沿海小镇与内陆城市数月间的日均气温。
Identify variables: 'daily mean temperature' is quantitative continuous. 'Month' is categorical. You might group by season and construct comparative box plots.
识别变量：“日均温度”为定量连续变量。“月份”为分类变量。你可以按季节分组并构建比较箱线图。
Measure of central tendency: The mean temperature in the coastal town might be lower but with a smaller spread, indicating a maritime influence.
集中趋势度量：沿海小镇的平均温度可能更低，但离散度较小，体现了海洋的影响。
These tasks integrate data handling, graphical representation, and interpretation — core skills for any cross-disciplinary application.
这些任务整合了数据处理、图形表示和解读——任何跨学科应用的核心技能。
10. Common Pitfalls in Cross-disciplinary Questions | 跨学科题目中的常见陷阱
Students often lose marks by forgetting to switch from mathematical symbols to contextual conclusions. After a hypothesis test, always write: 'There is [in]sufficient evidence at the X% significance level to suggest that...' Linking technical output to the real-world problem is vital.
学生常因忘记从数学符号切换到情境结论而失分。在假设检验之后，一定要写：“在 X% 显著性水平下，有／没有足够证据表明……”。将技术输出与现实问题联系起来至关重要。
Another pitfall: misinterpreting p-values. A p-value of 0.08 does not mean the null hypothesis is true with 92% probability. It is the probability of obtaining a result at least as extreme as the observed, assuming H₀ is true.
另一个陷阱：误读 p 值。p 值为 0.08 并不意味着原假设有 92% 概率为真。它是在假设 H₀ 为真的前提下，获得至少与观测结果一样极端的概率。
When dealing with correlation, never write 'proves causation'. Use phrases like 'suggests an association' or 'is consistent with a positive relationship'.
处理相关性时，永远不要写“证明因果关系”。使用诸如“表明一种关联”或“与正向关系一致”的表述。
By consciously practising these contextual interpretations, you turn statistical calculations into meaningful, cross-disciplinary analysis — precisely what examiners reward.
通过有意识地练习这些情境化解说，你将把统计计算转变为有意义的跨学科分析——这正是考官所青睐的。
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