Populations and Samples | 总体与样本

📚 Populations and Samples | 总体与样本

In statistics, the concepts of populations and samples are fundamental. A population is the complete set of individuals, items, or data that we wish to study. A sample is a subset selected from that population. Distinguishing between the two is essential for designing experiments, collecting data, and making valid inferences about the world around us.

在统计学中，总体与样本是两个基本概念。总体是我们希望研究的全部个体、项目或数据。样本是从总体中选出的一个子集。区分二者对于实验设计、数据收集以及对周围世界做出有效推断至关重要。

1. What is a Population? | 什么是总体？

A population includes every member of a specified group. In A-Level Mathematics, populations can be finite, such as all students in a particular college, or infinite, such as all possible values generated by a continuous measuring process. The key is that the population contains all the items of interest.

总体包含指定群体中的每一个成员。在A-Level数学中，总体可以是有限的，比如某所大学里的所有学生，也可以是无限的，比如连续测量过程产生的所有可能数值。关键在于总体包含了我们感兴趣的全体。

The size of a population is usually denoted by the capital letter N. For example, the population of UK A-Level Mathematics candidates in a given year might be N = 85,000. Identifying the population correctly is the first step in any statistical investigation.

总体的大小通常用大写字母 N 表示。例如，某一年英国A-Level数学考生的总体大小可能是 N = 85,000。在任何统计调查中，正确界定总体是第一步。

A parameter is a numerical value that summarises a characteristic of the population. Common parameters include the population mean (μ), population variance (σ²), and population proportion (p). Since we rarely have access to the whole population, parameters are typically unknown and must be estimated.

参数是概括总体某个特征的数值。常见参数包括总体均值 (μ)、总体方差 (σ²) 和总体比例 (p)。由于我们很少能接触到整个总体，参数通常是未知的，需要被估计。

2. What is a Sample? | 什么是样本？

A sample is a selection of individuals or observations chosen from the population. The number of observations in a sample is called the sample size and is denoted by n. For instance, if we test 50 light bulbs from a factory’s daily output of 10,000 bulbs, n = 50 and N = 10,000.

样本是从总体中选出的部分个体或观测值。样本中观测值的数目称为样本量，用 n 表示。例如，如果从某工厂日产10,000个灯泡中测试50个，那么 n = 50，N = 10,000。

A good sample should be representative, meaning that its characteristics mirror those of the population. If we only sample the largest apples from an orchard, the sample will not represent the typical size – it will be biased. Proper sampling techniques help give each member a known chance of selection.

一个好的样本应当具有代表性，即其特征反映了总体的特征。如果只从果园中选取最大的苹果作为样本，就无法代表典型大小——这会产生偏差。恰当的抽样技术有助于让每个成员都有已知的被选中的机会。

The numerical summary calculated from a sample is called a statistic. Familiar examples are the sample mean (x̄), sample standard deviation (s), and sample proportion (p̂). Statistics are used to make inferences about the corresponding population parameters.

由样本计算出的数字摘要称为统计量。常见的例子有样本均值 (x̄)、样本标准差 (s) 和样本比例 (p̂)。统计量用于对相应的总体参数进行推断。

3. Why Do We Sample? | 为什么需要抽样？

Sampling is essential because it is usually impractical, expensive, or impossible to measure every member of a population. Checking every battery’s lifetime would destroy all batteries; surveying every voter in a country would cost too much. A well-designed sample provides reliable information at a fraction of the cost and effort.

抽样之所以必要，是因为测量总体中的每一个成员往往不切实际、成本太高或根本不可能。检测每一枚电池的寿命会毁掉所有电池；调查全国每一位选民成本过高。一个精心设计的样本能以较低的成本和精力提供可靠信息。

Moreover, even when populations are finite and accessible, sampling can yield quicker results. In business and science, timely decisions often depend on sample evidence. The key is to understand and control the uncertainty introduced by using a sample instead of the census.

而且，即使总体是可及的有限集，抽样也能更快得到结果。在商业和科学领域，及时的决策常常依赖样本证据。关键在于理解并控制因使用样本而非全面调查所带来的不确定性。

4. Sampling Methods: Random Sampling | 抽样方法：随机抽样

Random sampling methods give every element in the population a known, non‑zero chance of being selected. This reduces selection bias and allows probability theory to be applied to the results. The three main types for Edexcel A-Level are simple random, systematic, and stratified sampling.

随机抽样方法让总体中的每个元素都有已知且非零的被选机会。这减少了选择性偏差，并允许对结果应用概率论。Edexcel A-Level中涉及的三类主要随机抽样是简单随机抽样、系统抽样和分层抽样。

Simple random sampling is like drawing names from a hat. Each member has an equal chance of being chosen, and every possible sample of size n is equally likely. This can be implemented using random number generators or tables. It is straightforward but may not capture subgroups well if they are small.

简单随机抽样就像从帽子里抽名字。每个成员被选中的机会均等，所有容量为 n 的样本出现的可能性相同。这可以用随机数生成器或随机数表实现。这种方法直接，但如果总体中的子群太小，可能难以捕捉。

Systematic sampling selects items at regular intervals from an ordered list. From a random start, every kth item is taken, where k = N/n. It is easier to perform than simple random sampling in many settings but can introduce periodicity bias if the list has an underlying pattern.

系统抽样是从有序列表中按固定间隔选取项目。先随机确定起点，然后每隔 k 个取一个，其中 k = N/n。在许多场合下，系统抽样比简单随机抽样更容易执行，但如果列表具有内在周期，则可能引入周期性偏差。

Stratified sampling divides the population into distinct strata (groups) based on a relevant characteristic, such as age or grade. Then a simple random sample is taken from each stratum, often proportionally to stratum size. This ensures representation of all groups and often yields more precise estimates.

分层抽样根据相关特征（如年龄或年级）将总体分成不同的层（群）。然后从每层中进行简单随机抽样，通常按层的大小成比例抽取。这确保了所有群体都能被代表，并常常能得到更精确的估计。

5. Sampling Methods: Non-Random Sampling | 抽样方法：非随机抽样

Non‑random (non‑probability) sampling methods do not give every individual a known chance of selection. These methods are often easier and cheaper but are more prone to bias. In Edexcel specifications, quota sampling and opportunity sampling are commonly considered.

非随机（非概率）抽样方法不给每个个体已知的被选机会。这些方法通常更方便、成本更低，但更容易产生偏差。在Edexcel大纲中，配额抽样和便利抽样是常见的方法。

Quota sampling involves interviewers being given specific quotas (e.g., 30 males aged 18–25, 30 females aged 26–35) and then selecting individuals that fit until the quotas are filled. While it guarantees the sample composition, the selection within quotas is subjective, which can lead to hidden bias.

配额抽样是指访问员获得特定配额（例如，30位18–25岁的男性、30位26–35岁的女性），然后自行选择符合条件的个体直至配额完成。虽然这保证了样本构成，但配额内的选择是主观的，可能导致隐藏的偏差。

Opportunity (convenience) sampling uses people or items that are easiest to reach: shoppers in a particular mall, or the first 20 customers entering a store. It is quick and cheap, but the sample is unlikely to be representative of the wider population, limiting how far conclusions can be generalised.

便利抽样使用最易接触的人群或物品：某一商场的购物者，或进入商店的前20位顾客。这种方法快速且廉价，但样本不太可能代表更广泛的总体，从而限制了结论的推广程度。

6. Parameters and Statistics | 参数与统计量

It is vital to distinguish between population parameters and sample statistics. A parameter is a fixed (though often unknown) number describing the population, while a statistic is a variable whose value depends on the particular sample drawn.

区分总体参数和样本统计量至关重要。参数是描述总体的固定（尽管常常未知）数值，而统计量是一个变量，其值取决于被抽取的具体样本。

Notation helps clarify: population mean is μ, sample mean is x̄; population standard deviation is σ, sample standard deviation is s. In Edexcel A-Level, using the correct notation is part of clear communication. For proportions, the population proportion is p, and the sample proportion is p̂.

符号可以帮助区分：总体均值为 μ，样本均值为 x̄；总体标准差为 σ，样本标准差为 s。在Edexcel A-Level中，使用正确的符号是清晰表达的一部分。对于比例，总体比例为 p，样本比例为 p̂。

Since parameters are usually unknown, statistics serve as point estimators. For example, x̄ is an estimator for μ. The quality of an estimator is judged by its bias and variability. An unbiased estimator has a sampling distribution centred at the parameter value it estimates.

由于参数通常未知，统计量就作为点估计量。例如，x̄ 是 μ 的估计量。估计量的好坏由其偏差和变异性来判断。无偏估计量的抽样分布以它所估计的参数值为中心。

7. Sampling Distributions | 抽样分布

The concept of a sampling distribution is one of the most powerful ideas in statistics. It is the probability distribution of a statistic if we were to take all possible samples of the same size from the population. It tells us how much a statistic, say x̄, is expected to vary from sample to sample.

抽样分布的概念是统计学中最强大的思想之一。如果从总体中抽取所有可能的大小相同的样本，某个统计量的概率分布就是它的抽样分布。它告诉我们一个统计量，比如 x̄，在不同样本之间预期会如何变化。

For the sample mean, under random sampling, its sampling distribution has mean μ (the population mean) and standard deviation σ/√n, known as the standard error. Even if the population distribution is not normal, the shape of the sampling distribution of the mean becomes approximately normal as n increases, thanks to the Central Limit Theorem.

对于样本均值，在随机抽样的情况下，其抽样分布的均值为 μ（总体均值），标准差为 σ/√n，称为标准误。即使总体分布不是正态的，随着 n 的增大，样本均值的抽样分布的形状也会近似正态，这要归功于中心极限定理。

The standard error measures the precision of the statistic: a smaller standard error means the sample mean is a more precise estimator of μ. This explains why larger samples give more reliable estimates.

标准误衡量统计量的精确程度：标准误越小，样本均值作为 μ 的估计量就越精确。这就解释了为什么大样本给出的估计更可靠。

8. Bias and Sampling Error | 偏差与抽样误差

Bias refers to a systematic tendency to over‑ or underestimate a parameter. It is not reduced by simply increasing the sample size. Bias can arise from a flawed sampling frame, non‑response, leading questions, or choosing a non‑random sample.

偏差指的是系统性地高估或低估某个参数的倾向。仅靠增大样本量并不能减少偏差。偏差可能源于有缺陷的抽样框、无回应、诱导性问题或选择了非随机样本。

Sampling error, in contrast, is the natural variation that occurs because we are looking at a sample rather than the whole population. This error is described by the standard error and decreases as n grows. Random sampling allows us to quantify sampling error using probability theory.

相反，抽样误差是由于观察样本而非整个总体而产生的自然变异。这种误差用标准误来描述，并随 n 增大而减小。随机抽样使我们能够使用概率论来量化抽样误差。

To minimise total error, we aim for an unbiased, efficient sampling design with a suitable sample size. Pilot surveys, careful wording, and high response rates help reduce bias, while confidence intervals capture the sampling error.

为了最小化总误差，我们的目标是采用无偏、高效的抽样设计以及合适的样本量。试点调查、措辞严谨和高回应率有助于减少偏差，而置信区间则捕捉了抽样误差。

9. The Central Limit Theorem (CLT) | 中心极限定理

The Central Limit Theorem is a cornerstone of A‑Level Statistics. It states: for independent observations, no matter what distribution the population has, the sampling distribution of the sample mean tends towards a normal distribution as the sample size increases. A common guideline is that n ≥ 30 is sufficient for the normal approximation to hold.

中心极限定理是A-Level统计学的基石。它指出：对于独立观测值，无论总体服从何种分布，随着样本量增加，样本均值的抽样分布趋于正态分布。一个常见的指导原则是，n ≥ 30 时正态近似即可成立。

More precisely, if samples of size n are taken from a population with mean μ and variance σ², the sample mean x̄ is approximately normally distributed with mean μ and variance σ²/n. Symbolically:

x̄ ~ N(μ, σ²/n) approximately

更精确地说，如果从均值为 μ、方差为 σ² 的总体中抽取容量为 n 的样本，样本均值 x̄ 近似服从正态分布，其均值为 μ，方差为 σ²/n。用符号表示为：

x̄ ~ N(μ, σ²/n)（近似）

This theorem justifies normal‑based inference even when the raw data are skewed or non‑normal. It is why we can use z‑tests and confidence intervals for means with large samples. Note: if the population itself is normal, x̄ is exactly normally distributed for any n.

这一定理使我们在原始数据偏态或非正态时依然可以基于正态分布进行推断。这就是为什么在大样本时我们可以使用z检验和均值的置信区间。注意：如果总体本身就是正态分布，那么无论 n 多大，x̄ 都精确地服从正态分布。

10. Confidence Intervals and Inference | 置信区间与推断

A key application linking samples and populations is the confidence interval. It provides an estimated range of plausible values for an unknown parameter. For a population mean μ when σ is known, a 95% confidence interval is given by:

x̄ ± 1.96 × (σ / √n)

连接样本和总体的一个关键应用是置信区间。它为未知参数提供了合理值的估计范围。当 σ 已知时，总体均值 μ 的95%置信区间由下式给出：

x̄ ± 1.96 × (σ / √n)

This formula uses the sample statistic x̄, the population standard deviation σ, and the standard error. The multiplier 1.96 comes from the standard normal distribution and gives the central 95% of the sampling distribution. The interpretation is that if we were to take many samples and compute such intervals, 95% of them would contain the true μ.

该公式使用了样本统计量 x̄、总体标准差 σ 以及标准误。乘数 1.96 源自标准正态分布，给出了抽样分布中央95%的区域。其解释是：如果我们重复抽取许多样本并计算这样的区间，其中95%的区间会包含真实的 μ。

When σ is unknown – which is typical – we estimate it using the sample standard deviation s and use the t-distribution. This shift brings the precision of the interval in line with the extra uncertainty from estimating σ.

当 σ 未知时（这是典型情况），我们用样本标准差 s 来估计它，并使用 t 分布。这一变化将区间的精确度与因估计 σ 而带来的额外不确定性对应起来。

11. Practical Considerations in Sampling | 抽样中的实际考虑

Real‑world surveys and experiments face challenges that go beyond theory. Undercoverage occurs when the sampling frame (a full list from which the sample is drawn) does not match the target population, e.g., using a telephone directory misses those without landlines. This introduces bias.

现实中的调查和实验往往面临理论之外的挑战。当抽样框（用于抽样的完整名单）与目标总体不匹配时，会产生覆盖不足。例如，使用电话号码簿会遗漏那些没有固定电话的人，从而引入偏差。

Non‑response is another serious problem: selected individuals do not provide data. If those who refuse are systematically different from those who respond (e.g., busier people or those with extreme views), the resulting sample may be biased. Follow‑up contact and incentives can help improve response rates.

无回应是另一个严重问题：被选中的个体没有提供数据。如果拒绝者与回应者存在系统差异（例如更忙碌的人或持极端观点的人），那么得到的样本可能带有偏差。跟进联系和激励措施有助于提高回应率。

Questionnaire design and measurement errors also affect accuracy. Questions must be clear, neutral, and unambiguous. Pilot studies – small trial runs of the sampling protocol – help detect problems before the main data collection begins.

问卷设计和测量误差也会影响准确性。问题必须清晰、中立且无歧义。试点研究——抽样协议的小规模试运行——有助于在主要数据收集开始前发现问题。

12. Summary | 总结

The distinction between a population and a sample is the starting point for all statistical work. Parameters describe populations, statistics describe samples, and inference uses statistics to make probability‑based statements about parameters. Understanding sampling methods, sources of bias, and the behaviour of statistics through their sampling distributions empowers students to design studies and interpret results critically.

区分总体与样本是一切统计工作的起点。参数描述总体，统计量描述样本，而推断就是利用统计量对参数做出基于概率的论断。理解抽样方法、偏差来源以及通过抽样分布所揭示的统计量的行为，能使学生有能力设计研究并批判性地解读结果。

The Central Limit Theorem provides the theoretical justification for many procedures encountered in Edexcel A‑Level Mathematics, showing that normality can often be assumed for sample means even when the underlying data are not normally distributed. Together with confidence intervals and awareness of practical challenges, these concepts form a robust foundation for further study in statistics.

中心极限定理为Edexcel A-Level数学中遇到的众多方法提供了理论依据，表明即使原始数据不服从正态分布，样本均值也常常可以假设为正态。结合置信区间以及对实际挑战的认识，这些概念为进一步学习统计学奠定了坚实的基础。

Published by TutorHao | Mathematics Revision Series | aleveler.com

更多咨询请联系16621398022（同微信）

Populations and Samples | 总体与样本