📚 Year 11 CAIE Statistics: Interdisciplinary Comprehensive Problem Training | CAIE 11年级统计：跨学科综合题型训练
In CAIE IGCSE Statistics, interdisciplinary questions bridge mathematical techniques and real-world contexts, from predicting genetic traits to analysing economic trends. Mastering these problems requires not only statistical knowledge but also the ability to interpret scenarios from biology, business, geography, physics, and beyond. This comprehensive training will equip you with strategies and examples to excel.
在 CAIE IGCSE 统计中，跨学科题目将数学方法与现实情境联系起来，从预测遗传特征到分析经济趋势。掌握这些问题不仅需要统计知识，还需要解读生物、商业、地理、物理等领域情境的能力。本综合训练将为你提供策略和示例，助你取得优异成绩。
1. Why Interdisciplinary Questions Dominate the Papers | 为什么跨学科题目主导试卷
Cambridge designs the Statistics syllabus so that almost every question is wrapped inside a context. A question about the binomial distribution may be dressed as a genetics problem; a question about correlation may arrive as a study of rainfall and crop yield. The examiner is testing two things at once: whether you can extract the statistical model from the story, and whether you can communicate the answer back in the language of the story.
剑桥在设计统计考纲时，几乎把每道题都包在真实情境里。一道二项分布题可能伪装成遗传学问题；一道相关性题可能以降雨量与农作物产量研究的形式出现。考官同时在考查两件事：你能否从故事中提取出统计模型，以及你能否用故事本身的语言把答案讲回去。
The key habit is a two-stage translation. Stage one: read the scenario and underline every number, every unit and every relational word such as "per", "more likely than" or "on average". Stage two: rewrite the scenario in abstract symbols before you touch the calculator.
关键习惯是"两段式翻译"。第一段：读情境，划出每一个数字、每一个单位，以及"每""比……更可能""平均而言"这类关系词。第二段：在按计算器之前，先把情境改写成抽象符号。
2. The Core Toolkit You Must Have Ready | 上场前必须备好的核心工具
Before attempting any interdisciplinary question, make sure the following techniques are automatic. They are the vocabulary of every context the examiner can invent, so fluency here buys you time and marks under pressure.
在尝试任何跨学科题目之前，请确保下列技巧已成为条件反射。它们是考官能编出的任何情境的"通用词汇"，熟练于此可以在压力下为你换来时间与分数。
Measures of location and spread: mean, median, mode, range, interquartile range, standard deviation.
位置与离散程度的度量：平均数、中位数、众数、极差、四分位距、标准差。
Graphical display: scatter diagrams, box-and-whisker plots, histograms, cumulative frequency curves, time series graphs.
图形展示：散点图、箱线图、直方图、累积频率曲线、时间序列折线图。
Probability rules: addition, multiplication, tree diagrams, conditional probability, independence.
概率法则：加法、乘法、树形图、条件概率、独立性。
Distributions: the binomial model for counting successes, and the normal model for continuous measurements.
分布：用于计数成功次数的二项模型，以及用于连续测量值的正态模型。
| Context | Typical statistical tool | 常见统计工具 |
| Genetics / Biology | Binomial probability, tree diagrams | 二项概率、树形图 |
| Business / Economics | Index numbers, moving averages, percentage change | 指数、移动平均、百分比变化 |
| Geography / Climate | Scatter diagrams, correlation, regression line | 散点图、相关性、回归直线 |
| Physics / Engineering | Normal distribution, standard deviation, z-values | 正态分布、标准差、z 值 |
| Sport / Health | Box plots, comparing two distributions | 箱线图、比较两组分布 |
3. Probability in Genetics: Biology Meets Statistics | 遗传学中的概率：生物学与统计学的交汇
Mendelian genetics provides a classic application of probability. Consider a single-gene cross in which the dominant trait appears with probability 0.75. Using the binomial distribution, we can calculate the likelihood of exactly r offspring displaying the dominant trait out of n offspring.
孟德尔遗传学提供了概率的经典应用。考虑一个单基因杂交，显性性状出现的概率为 0.75。利用二项分布，我们可以计算在 n 个后代中恰好有 r 个表现出显性性状的概率。
P(X = r) = nCr × pʳ × q⁽ⁿ⁻ʳ⁾
bility of failure. The factor nCr counts the number of orders in which the r successes can be arranged among n trials. The model assumes a fixed number of trials, a constant probability of success, and independence between trials.
其中 p 是单次试验成功的概率，q = 1 − p 是失败的概率。系数 nCr 统计的是 r 次成功在 n 次试验中出现的排列方式数。该模型假设试验次数固定、成功概率恒定，且各次试验相互独立。
s are grown, find the probability that exactly 5 are tall. Here n = 8, p = 0.75, q = 0.25.
示例：如果种植 8 株豌豆，求恰好 5 株为高茎的概率。这里 n = 8，p = 0.75，q = 0.25。
P(X = 5) = 8C5 × (0.75)⁵ × (0.25)³ = 56 × 0.2373 × 0.015625 ≈ 0.2076
for our example μ = 6 and σ² = 1.5. Always state the distribution formally, for example X ~ 8, 0.75), before substituting numbers; examiners award a mark for correct model statement.
该模型的均值与方差为 μ = np、σ² = npq，因此本例中 μ = 6、σ² = 1.5。代入数字之前，请先正式写出分布，例如 X ~ B(8, 0.75)；考官会为正确的模型表述给分。
4. Beyond the Binomial: Tree Diagrams and Medical Testing | 超越二项分布：树形图与医学检测
Biology also brings conditional probability. A screening test for a disease has a sensitivity (probability of a positive result given the person has the disease) and a false positive rate. A tree diagram splits the population first by disease status, then by test result, and each branch is multiplied.
生物学也会带来条件概率。某疾病的筛查检测有灵敏度（患病者检测呈阳性的概率）与假阳性率。树形图先按是否患病分层，再按检测结果分层，每条路径上的概率相乘。
If 2% f a positive result is 0.02 × 0.95 + 0.98 × 0.08 = 0.019 + 0.0784 = 0.0974. The conditional probability that a positive person actually has the condition is 0.019 ÷ 0.0974 ≈ 0.195.
若某人群中 2% 患有该病，检测的检出率为 0.95，假阳性率为 0.08，则检测呈阳性的概率为 0.02 × 0.95 + 0.98 × 0.08 = 0.019 + 0.0784 = 0.0974。而阳性者真正患病的条件概率为 0.019 ÷ 0.0974 ≈ 0.195。
This counter-intuitive result - a large majority of positives are healthy - is exactly the kind of interpretation mark the examiner is looking for. Always write a sentence in context, not just a decimal.
这个反直觉的结论——大多数阳性者其实健康——正是考官所看重的"解释性得分点"。永远要用情境写一句话，而不只是给出一个小数。
5. Business and Economics: Index Numbers and Time Series | 商业与经济：指数与时间序列
Economics contexts usually involve comparing values across time. An index number expresses a value relative to a base period, which is always set to 100. This makes series with different units directly comparable.
经济情境通常涉及跨时间比较数值。指数表示相对于基期的数值，基期恒定为 100，这使得单位不同的数列可以直接比较。
Index number = (value in current period ÷ value in base period) × 100
150 in the base year 2015, the index is (240 ÷ 150) × 100 = 160, meaning prices rose by 60%. A weighted index multiplies each item' ividing by the total weight, reflecting that some items matter more to a household budget.
若一篮子商品 2025 年花费 240，基年 2015 年为 150，则指数为 (240 ÷ 150) × 100 = 160，即价格上涨了 60%。加权指数则先把各项指数乘以权重再除以权重总和，以反映某些项目在家庭预算中更重要。
Time series questions also ask for a moving average to reveal a trend. A four-point moving average smooths seasonal fluctuation; the difference between the raw value and the trend is the seasonal component, which should be averaged and then adjusted so the seasonal effects sum to zero.
时间序列题还会要求用移动平均揭示趋势。四点移动平均可平滑季节波动；原始值与趋势值之差即季节成分，应求其平均并作调整，使季节效应之和为零。
6. Geography and Climate: Correlation and Regression | 地理与气候：相关与回归
Geography questions present paired data such as monthly rainfall and crop yield, and ask you to plot a scatter diagram, describe the correlation and draw a line of best fit. Describing correlation requires three elements: direction (positive or negative), strength (strong, moderate or weak) and form (linear).
地理题常给出配对数据，如月降雨量与农作物产量，要求绘制散点图、描述相关性并画出最佳拟合直线。描述相关性需要三个要素：方向（正或负）、强度（强、中等或弱）与形态（线性）。
y = a + bx, where b = Σ(x − x̄)(y − ȳ) ÷ Σ(x − x̄)² and a = ȳ − b x̄
The gradient b tells you how much y changes for each unit increase in x, and in context this becomes a sentence such as "each additional 10 mm of rainfall is associated with an average increase of 0.4 tonnes per hectare". Never use the regression line to predict far outside the range of the data, and always note that correlation does not prove causation.
斜率 b 表示 x 每增加一个单位时 y 的变化量，在情境中应写成类似"每多 10 毫米降雨量，平均每公顷增产 0.4 吨"的句子。切勿用回归直线去预测远离数据范围的值，并且始终指出相关不等于因果。
7. Physics and Measurement: The Normal Distribution | 物理与测量：正态分布
Physics and engineering supply continuous measurements - the length of a rod, the resistance of a component - whose repeated readings cluster symmetrically around a true value. The normal distribution models this shape, and standardising converts any normal variable into the standard normal scale.
物理与工程提供连续测量值——杆的长度、元件的电阻——重复读数围绕真值对称聚集。正态分布正是刻画这种形状的模型，而标准化可以把任意正态变量转换为标准正态尺度。
z = (x − μ) ÷ σ
Because the normal curve is symmetric, roughly 68% of values lie within one standard deviation of the mean, 95% within two, and 99.7% a threshold, so sketch the curve, shade the required region and mark z = 0 at the centre.
由于正态曲线对称，约 68% 的数值落在均值 ±1 个标准差内，95% 落在 ±2 个标准差内，99.7% 落在 ±3 个标准差内。考官常问高于或低于某阈值的比例，因此要画曲线、阴影标出所求区域，并在中心标出 z = 0。
| Distance from mean | Approximate percentage | 说明 |
| μ ± 1σ | about 68% | 约 68% |
| μ ± 2σ | about 95% | 约 95% |
| μ ± 3σ | about 99.7% | 约 99.7% |
8. Sports Science and Health: Comparing Two Distributions | 运动科学与健康：比较两组分布
A frequent interdisciplinary task is to compare two groups, for example the resting heart rates of athletes and non-athletes. The correct method is to compare medians for a measure of location and interquartile ranges for a measure of spread, always in the context of the variables involved.
常见的跨学科任务是比较两组数据，例如运动员与非运动员的静息心率。正确方法是先用中位数比较集中趋势，再用四分位距比较离散程度，且始终结合变量本身的情境来表述。
A full-mark comparison looks like this: "The athletes have a lower median resting heart rate (58 bpm compared with 72 bpm), and their interquartile range is smaller (10 bpm compared with 18 bpm), so their heart rates are both lower and more consistent." Notice that both the numbers and the interpretation appear together.
满分答案长这样："运动员的静息心率中位数更低（58 次/分 对 72 次/分），且四分位距更小（10 次/分 对 18 次/分），说明他们的心率既更低也更稳定。"注意数字与解读必须同时出现。
If the question instead asks which group is more consistent, the answer is whichever has the smaller spread - not the smaller mean. Confusing location with spread is one of the most common lost marks in this topic.
若题目改问哪一组更稳定，答案取决于离散程度更小的一组，而不是平均数更小的一组。把集中趋势与离散程度混为一谈，是本考点最常见的失分原因之一。
9. Social Sciences: Sampling, Bias and Questionnaires | 社会科学：抽样、偏差与问卷
Sociology and psychology contexts test your understanding of how data is collected rather than how it is calculated. A sample must be representative of the population if conclusions are to be generalised, and the sampling method determines whether that is plausible.
社会学与心理学情境考查的是你对"数据如何被收集"的理解，而非计算。样本必须能代表总体，结论才可推广；而抽样方法决定了这一点是否成立。
Random sampling: every member of the population has an equal chance of selection, but it may be impractical for large populations.
随机抽样：总体中每个个体被选中的机会均等，但对大规模总体可能不切实际。
Systematic sampling: selecting every kth item from a list; simple but vulnerable to periodic patterns in the list.
系统抽样：从名单中每隔 k 个抽取一个；操作简单，但容易受名单周期性规律影响。
Quota sampling: fixed numbers from each subgroup chosen by the interviewer; fast but subjective.
配额抽样：由访问员按子群体固定数量选取；速度快，但带有主观性。
Questionnaire bias: leading questions, poor wording, or surveying only people who happen to be nearby.
问卷偏差：诱导性问题、措辞不当，或只调查恰好路过的人。
When asked to criticise a questionnaire, name the specific flaw and propose a rewrite. "The question 'Do you agree that the new library is a great asset?' is leading, because it assumes agreement; a neutral version would be 'How would you rate the new library?'"
当被要求批评一份问卷时，要指出具体缺陷并提出改写方案。"你是否同意新图书馆是一大财富？"这一问法带有诱导性，因为它预设了同意；中性版本应改为"你如何评价新图书馆？"
10. Histograms, Cumulative Frequency and Grouped Data | 直方图、累积频率与分组数据
When continuous data is grouped into unequal class widths, a bar chart would mislead, so a histogram is used instead. The height of each bar is the frequency density, not the frequency itself, because the area of the bar must be proportional to the frequency.
当连续数据被分入宽度不等的组距时，条形图会造成误导，因此改用直方图。每根柱的高度是频率密度而非频数，因为柱的面积必须与频数成正比。
Frequency density = frequency ÷ class width
Cumulative frequency curves answer questions about medians, quartiles and percentiles. To read a median, find 50% of the total frequency on the vertical axis, move horizontally to the curve, then drop vertically to the horizontal axis. The interquartile range is the upper quartile minus the lower quartile, and it is more resistant to outliers than the range.
累积频率曲线用于回答中位数、四分位数与百分位数问题。读取中位数的方法是：在纵轴找到总频数的 50%，水平移到曲线上，再垂直落到横轴。四分位距等于上四分位数减下四分位数，它比极差更能抵抗异常值的影响。
In an ecosystem study, for example, the lengths of 200 leaves could be grouped into intervals of 0-20, 20-30 and 30-50 mm. Calculating three frequency densities is a standard opening move, and the modal class is the class with the greatest frequency density - not necessarily the greatest frequency.
例如在生态系统研究中，200 片叶子的长度可分为 0–20、20–30、30–50 毫米等组距。先算出三个频率密度是标准起手式，而众数所在组是频率密度最大的组，不一定是频数最大的组。
11. Exam Technique: Structuring Interdisciplinary Answers | 应试技巧：构建跨学科答案
Interdisciplinary questions reward disciplined presentation. A clear structure makes your reasoning visible to the examiner and protects you when a final number goes wrong, because method marks are awarded independently.
跨学科题目奖励有纪律的书写。清晰的结构能让考官看到你的推理过程，即使最后数值算错也能保住步骤分，因为方法分是独立给出的。
Define the variable and its distribution or model before substituting numbers.
在代入数字之前，先定义变量及其分布或模型。
Write the formula, then substitute, then evaluate - three separate lines.
先写公式，再代入，最后计算——分成三行。
Keep full accuracy in the calculator and round only at the final step, usually to 3 significant figures.
计算器内保持全精度，只在最后一步取近似，通常保留 3 位有效数字。
Finish every part with a sentence in context that interprets the number.
每一小问都以一句结合情境的解读句收尾。
Check units: probabilities are between 0 and 1, percentages between 0 and 100, and standard deviations are never negative.
检查单位：概率在 0 与 1 之间，百分比在 0 与 100 之间，标准差不可能为负。
A useful self-check for any result is the sanity test: does this number make sense in the story I was told? If a probability exceeds 1, or a predicted height is 4 metres, the model or the arithmetic has gone wrong.
对任何结果都可用"合理性检验"：这个数字在题目讲的故事里说得通吗？如果概率超过 1，或预测身高是 4 米，那么模型或运算一定出了问题。
12. A Worked Mixed-Topic Question | 混合考点模拟题精讲
Try this composite problem, which deliberately combines biology, probability and data presentation in the style of a Paper 2 final question.
请尝试下面这道综合题，它刻意把生物学、概率与数据展示结合在一起，模仿 Paper 2 压轴题的风格。
A biologist crosses two pea plants. Each offspring independently has probability 0.75 of being tall. (a) State the distribution of the number of tall plants among 8 offspring. (b) Find the probability that exactly 5 are tall. (c) Find the expected number of tall plants and the standard deviation. (d) State one assumption required by the model.
一位生物学家让两株豌豆杂交。每个后代独立地以 0.75 的概率为高茎。（a）写出 8 个后代中高茎株数的分布；（b）求恰好 5 株为高茎的概率；（c）求高茎株数的期望与标准差；（d）写出该模型所需的一个假设。
Part (a): X ~ ): 8C5 × (0.75)⁵ × (0.25)³ = 56 × 0.2373 × 0.015625 ≈ 0.2076. Part (c): μ = np = 6 and σ = √npq = √1 h offspring is independent of the others, or the probability of a tall offspring stays constant at 0.75 for every plant.
（a）X ~ B(8, 0.75)。（b）8C5 × (0.75)⁵ × (0.25)³ = 56 × 0.2373 × 0.015625 ≈ 0.2076。（c）μ = np = 6，σ = √npq = √1.5 ≈ 1.225。（d）例如：各后代彼此独立，或每一株为高茎的概率恒为 0.75。
