📚 Mathematical Statistics: Experimental Design and Data Analysis Methods | 数学统计：实验设计与数据分析方法
Statistics is the science of collecting, organising, analysing and interpreting data. In this article, we explore the core ideas behind experimental design and data analysis methods found in A-Level Mathematics and Statistics.
统计学是收集、整理、分析并解读数据的科学。在本文中，我们将探讨 A-Level 数学与统计中涉及的实验设计与数据分析方法的核心概念。
1. Types of Studies | 研究类型
Studies can be broadly classified into two categories: observational studies and experiments. In an observational study, the researcher observes subjects and records variables of interest without imposing treatments. In an experiment, the researcher actively applies treatments to subjects to observe the effect.
研究可大致分为两类：观察性研究和实验。在观察性研究中，研究者观察对象并记录感兴趣的变量，而不施加处理；在实验中，研究者主动对对象施加处理，以观察其效果。
Choosing the right sampling method is also essential. Common sampling methods include:
选择合适的抽样方法同样至关重要。常见抽样方法包括：
- Random sampling: every member of the population has an equal chance of being selected.
随机抽样：总体中每个成员被选中的概率相等。 - Stratified sampling: the population is divided into strata, and random samples are taken from each.
分层抽样：将总体分成多个层，再从每层中随机抽取样本。 - Systematic sampling: select every k-th member from a list.
系统抽样：从列表中每隔 k 个抽取一个成员。 - Convenience sampling: choose individuals who are easiest to reach.
方便抽样：选择最容易接触到的个体。
2. Principles of Experimental Design | 实验设计原则
Good experimental design is built on four key principles. These help reduce bias and ensure that conclusions are valid.
良好的实验设计建立在四条关键原则之上，这些原则有助于减少偏差并确保结论有效。
| Principle 原则 | Description 描述 |
|---|---|
| Randomisation 随机化 | Assign subjects to treatments randomly to avoid bias. 随机分配对象到各处理组，以避免偏差。 |
| Replication 重复 | Repeat the experiment on multiple subjects to reduce variation. 在多个对象上重复实验，以减小变异。 |
| Blocking 分组 | Group similar subjects together to account for known sources of variation. 将相似对象分组，以控制已知的变异来源。 |
| Control 控制 | Use a control group or control variables to isolate the treatment effect. 使用对照组或控制变量，以分离处理效应。 |
3. Common Experimental Designs | 常见实验设计
Several standard designs are used in practice. Each has its own strengths and weaknesses.
实际中常用若干标准设计，每种设计都有其优点和不足。
- Completely randomised design: all subjects are allocated randomly to treatments.
完全随机化设计：所有对象被随机分配到各处理组。 - Randomised block design: subjects are first divided into blocks based on a nuisance factor, then randomly assigned within each block.
随机区组设计：首先根据干扰因素将对象分成区组，然后在每个区组内随机分配。 - Factorial design: two or more factors are studied simultaneously at different levels.
析因设计：同时在不同水平上研究两个或多个因素。 - Matched pairs design: subjects are paired based on similar characteristics, and each member receives a different treatment.
配对设计：依据相似特征将对象配对，每对中的两个成员接受不同处理。
4. Data Types and Measurement Scales | 数据类型与测量尺度
Data can be qualitative (categorical) or quantitative (numerical). Quantitative data is further divided into discrete and continuous types.
数据可分为定性（类别）数据和定量（数值）数据。定量数据又可进一步分为离散型和连续型。
| Scale 尺度 | Description 描述 | Example 示例 |
|---|---|---|
| Nominal 名义 | Categories with no order. 无顺序的类别。 |
Colour, gender 颜色、性别 |
| Ordinal 有序 | Categories with a meaningful order. 有意义的顺序类别。 |
Grades, rankings 等级、排名 |
| Interval 间隔 | Numerical with equal intervals but no true zero. 等间隔但无绝对零点的数值。 |
Temperature in °C 摄氏温度 |
| Ratio 比率 | Numerical with a true zero point. 具有绝对零点的数值。 |
Height, mass 身高、质量 |
5. Descriptive Statistics | 描述性统计
Descriptive statistics summarise a data set using measures of central tendency and measures of spread.
描述性统计通过集中趋势测量和离散程度测量来概括数据集。
For a population, the mean and variance are denoted by μ and σ². For a sample, they are denoted by x̄ and s².
对于总体，均值和方差记为 μ 和 σ²；对于样本，记为 x̄ 和 s²。
μ = (∑x) / N or x̄ = (∑x) / n
s² = ∑(x – x̄)² / (n – 1) , σ² = ∑(x – μ)² / N
The sample standard deviation is the square root of the variance, often used because it has the same units as the original data.
样本标准差是方差的平方根，由于与原始数据单位相同，因此经常使用。
6. Data Visualisation | 数据可视化
Graphs reveal patterns and outliers that tables may hide. The choice of graph depends on the data type.
图形能揭示表格中可能隐藏的规律和异常值。图形的选择取决于数据类型。
- Bar charts are suitable for categorical data.
条形图适用于类别数据。 - Histograms display the distribution of continuous data.
直方图展示连续数据的分布。 - Box plots show the median, quartiles and outliers.
箱线图显示出中位数、四分位数和异常值。 - Scatter plots illustrate relationships between two variables.
散点图展示两个变量之间的关系。
When constructing a histogram, the area of each bar is proportional to the frequency. This is different from a bar chart, where height alone matters.
绘制直方图时，每个矩形的面积与频数成正比，这与仅看高度的条形图不同。
7. Probability and Sampling Distributions | 概率与抽样分布
Probability theory underpins statistical inference. If X is a random variable, we may write its probability distribution as P(X = x).
概率论是统计推断的基础。若 X 是随机变量，我们可将其概率分布记为 P(X = x)。
For a discrete random variable, the mean and variance are:
对于离散型随机变量，均值和方差为：
E(X) = ∑ x P(X = x) , Var(X) = ∑ x² P(X = x) – (E(X))²
The Central Limit Theorem states that the sampling distribution of the sample mean x̄ is approximately normal when n is large, regardless of the population shape. This justifies the use of normal distribution methods in many situations.
中心极限定理表明，当 n 较大时，样本均值 x̄ 的抽样分布近似正态，无论总体形态如何。这使我们在许多情况下可以使用正态分布方法。
8. Confidence Intervals | 置信区间
A confidence interval gives an estimated range of values for a population parameter. For a population mean with known standard deviation σ, the confidence interval is:
置信区间给出总体参数的估计范围。当总体标准差 σ 已知时，总体均值的置信区间为：
x̄ ± z × (σ / √n)
Here, z is the critical value from the standard normal distribution: z = 1.645 for 90% confidence, z = 1.960 for 95%, and z = 2.576 for 99%.
其中 z 是标准正态分布的关键值：90% 置信度时为 1.645，95% 时为 1.960，99% 时为 2.576。
When σ is unknown, we use the sample standard deviation s and the t-distribution with n – 1 degrees of freedom.
当 σ 未知时，我们使用样本标准差 s 以及自由度为 n – 1 的 t 分布。
x̄ ± t × (s / √n)
Interpretation: if we repeated the sampling process many times, approximately 95% of such intervals would contain the true mean.
解读：若我们重复抽样多次，大约 95% 的此类区间会包含真实均值。
9. Hypothesis Testing | 假设检验
Hypothesis testing is a formal method for making decisions using data. We start with a null hypothesis H₀ and an alternative hypothesis H₁.
假设检验是一种利用数据做决策的正式方法。我们从零假设 H₀ 和备择假设 H₁ 开始。
- H₀: usually states no effect or no difference, e.g. μ = 0.
H₀：通常表示无效或没有差异，例如 μ = 0。 - H₁: states what we suspect, e.g. μ ≠ 0, μ > 0, or μ < 0.
H₁：表示我们怀疑的情况，例如 μ ≠ 0、μ > 0 或 μ < 0。
The test statistic is compared with a critical value, or we compute a p-value. If the p-value is less than the significance level (often 0.05), we reject H₀.
将检验统计量与临界值比较，或计算 p 值。若 p 值小于显著性水平（通常 0.05），则拒绝 H₀。
Two types of errors are possible: Type I error (rejecting H₀ when it is true) and Type II error (failing to reject H₀ when it is false).
可能发生两类错误：第一类错误（H₀ 为真时拒绝 H₀）和第二类错误（H₀ 为假时未拒绝 H₀）。
10. Correlation and Regression | 相关与回归
Correlation measures the strength and direction of a linear relationship between two quantitative variables. Pearson’s correlation coefficient r lies between -1 and 1.
相关衡量两个定量变量之间线性关系的强度和方向。皮尔逊相关系数 r 介于 -1 和 1 之间。
The least squares regression line is used to predict the value of one variable from another.
最小二乘回归线用于从一个变量预测另一个变量的值。
y = a + bx
b = ∑(x – x̄)(y – ȳ) / ∑(x – x̄)² , a = ȳ – bx̄
Note: correlation does not imply causation. A strong relationship may be due to a confounding variable or pure chance.
注意：相关并不意味因果。强关系可能源于混杂变量或纯属巧合。
11. Common Pitfalls | 常见误区
Students often lose marks in exams by ignoring design issues or misinterpreting results. Here are some common pitfalls:
学生在考试中常因忽略设计问题或错误解读结果而失分。以下是一些常见误区：
- Confusing correlation with causation.
混淆相关与因果关系。 - Using the wrong graph for a data type.
使用与数据类型不符的图形。 - Forgetting to include units in the answer.
忘记在答案中写上单位。 - Applying the normal distribution when n is small and σ is unknown.
当 n 较小且 σ 未知时仍然使用正态分布。 - Ignoring outliers when computing the mean and standard deviation.
计算均值和标准差时忽略异常值。 - Stating that a 95% confidence interval contains the sample mean with 95% probability.
错误地声称 95% 置信区间以 95% 概率包含样本均值。
12. Conclusion | 总结
Experimental design and data analysis are essential skills in A-Level Mathematics. By understanding study types, design principles, descriptive tools, probability and inference, you can analyse data accurately and critically.
实验设计与数据分析是 A-Level 数学中的必备技能。通过理解研究类型、设计原则、描述性工具、概率与推断，你能够准确而批判地分析数据。
Always check the conditions before using a statistical method, and remember that statistics is not just about numbers — it is about making sound decisions under uncertainty.
在使用统计方法前务必检查条件，并记住统计不仅是数字——更是关于在不确定性中做出可靠决策。
Published by TutorHao | Mathematics 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