📚 6A Sampling | 第6A章 抽样
Sampling is a cornerstone of statistical investigation. In IB Mathematics, mastering different sampling techniques and recognising their potential biases is vital for conducting surveys, interpreting data, and writing a successful Internal Assessment. This article explores the key methods of sampling, from simple random selection to cluster sampling, and highlights how to choose the right approach for a given context.
抽样是统计调查的基石。在IB数学中，掌握不同的抽样技术并识别其潜在的偏差，对于开展调查、解读数据和撰写优秀的内部评估至关重要。本文将探讨从简单随机抽样到整群抽样的各种关键方法，并强调如何为特定情境选择合适的方式。
1. Population and Sample | 总体与样本
A population is the complete collection of individuals, objects, or measurements that we wish to study. For example, all IB diploma candidates in a city, or every orange produced in a particular orchard. Because examining an entire population is often impractical, we select a sample — a subset of the population — to draw conclusions about the whole. The process must be carefully designed so the sample accurately reflects the population's characteristics.
总体是我们希望研究的全部个体、对象或测量值的集合。例如，一个城市中的所有IB文凭考生，或者某个果园产出的所有橙子。由于检查整个总体通常不切实际，我们会选择一个样本——总体的一个子集——来推断整体的特征。抽样过程必须精心设计，以便样本能够准确地反映总体的特性。
The target population is the group we ideally want to study, while the sampled population is the group from which the sample is actually drawn. Discrepancies between these two can introduce coverage error. For instance, using a school's yearbook list misses students who left before the photo was taken, so the sampled population differs from the true student body.
目标总体是我们理想中希望研究的群体，而被抽样总体是实际从中抽取样本的群体。两者之间的差异可能引入覆盖误差。例如，使用学校的年鉴名册会遗漏那些在拍照前已经离开的学生，因此被抽样总体与真实的学生群体并不一致。
2. Simple Random Sampling | 简单随机抽样
Simple random sampling (SRS) gives every possible sample of a given size n an equal chance of being selected from a population of size N. This is the most fundamental unbiased method. It can be implemented by assigning a unique number to each population member and using a random number generator or drawing lots. SRS requires a complete list of the population, known as a sampling frame.
简单随机抽样使得从大小为 N 的总体中抽取的每一个给定容量 n 的样本都有相等的被选中机会。这是最基本的无偏方法。可以通过给总体中的每个成员分配一个唯一编号，然后使用随机数生成器或抽签来实现。简单随机抽样需要一份完整的总体名单，即抽样框。
One key assumption is independence: selecting one individual does not influence the selection of another. When the sampling frame is accurate and the randomisation is properly executed, SRS minimises selection bias. However, it can be expensive and time-consuming when the population is large or geographically dispersed, and it may not guarantee representation of small subgroups.
一个关键假设是独立性：选中一个个体不影响另一个个体的选择。当抽样框准确且随机化过程得到正确执行时，简单随机抽样能够最大限度地减少选择偏差。然而，当总体很大或地理分布广泛时，这种方法可能成本高且耗时，并且未必能保证较小亚群的代表性。
3. Systematic Sampling | 系统抽样
mple of 80, k = 10; we randomly pick a starting number between 1 and 10, say 4, and then select the 4th, 14th, 24th, and so on.
系统抽样是从有序名单中按固定间隔选取个体。从一个随机选择的起点开始，每隔 k 个元素抽取一个，其中 k 是抽样间隔，由 N/n 计算得出（四舍五入取整）。例如，如果一所学校有800名学生，我们需要一个80人的样本，则 k = 10；我们在1到10之间随机选一个起点，比如4，然后依次选取第4、14、24号，依此类推。
This method is easier to administer than SRS, especially when a physical list exists. It can provide good coverage as long as the list does not have a periodic pattern that matches the sampling interval. For example, if houses on a street are arranged with corner houses every 10th position and those are larger, a systematic sample starting at a corner might over-represent larger houses, leading to bias.
这个方法比简单随机抽样更容易操作，尤其在已有实体名单的情况下。只要名单不存在与抽样间隔一致的周期性模式，系统抽样就能提供良好的覆盖。例如，如果街道上的房屋每第10座是街角的较大房屋，而系统抽样的起点恰好落入街角位置，就可能过度代表大房屋，从而导致偏差。
4. Stratified Sampling | 分层抽样
Stratified sampling divides the population into distinct subgroups, called strata, that share common characteristics (e.g., age group, gender, grade level). A separate random sample is then taken from each stratum. Usually, the sample size from each stratum is proportional to the stratum's share of the population. This guarantees that every subgroup is represented, improving the accuracy of estimates for the overall population.
分层抽样将总体划分为具有共同特征的不同子群，称为层（例如年龄组、性别、年级）。然后从每一层中分别抽取一个随机样本。通常，每层的样本量与该层在总体中所占的比例成正比。这确保了每个子群都能被代表，从而提高对整体总体估计的准确性。
For example, in a survey of student opinions on a new cafeteria menu, splitting the student body into grade levels (Year 1, Year 2, etc.) and sampling proportionally ensures that the views of all grades are reflected, even if the number of students per grade differs. Stratification reduces sampling variability if the strata are homogeneous internally but differ from each other.
例如，在关于新食堂菜单的学生意见调查中，将学生群体按年级分层（一年级、二年级等）并按比例抽样，可以确保所有年级的意见都能反映出来，即便各年级的学生人数不同。如果各层内部是同质的但层间差异较大，分层抽样可以减少抽样变异。
5. Cluster Sampling | 整群抽样
Cluster sampling partitions the population into clusters that are each representative of the entire population on a smaller scale. Instead of sampling individuals from all clusters, we randomly select a few clusters and then collect data from every member within those chosen clusters. This is highly cost-effective when the population is naturally divided into groups, such as households, school classes, or hospital wards.
整群抽样是将总体划分成若干个群，每个群在较小规模上都能代表整个总体。我们不是从所有群中抽取个体，而是随机选择几个群，然后从这些被选中的群中收集每一个成员的数据。当总体自然分成组群（如家庭、班级或医院病房）时，这种方法极具成本效益。
Unlike stratified sampling, where we sample from every stratum, cluster sampling only surveys whole clusters. For instance, to estimate the average height of 14-year-old students in a large city, we might randomly select 10 schools out of 100 and measure every 14-year-old in those schools. The main risk is that if the selected clusters are not truly representative, the sample can be biased; however, increasing the number of clusters reduces this risk.
与从每一层都抽样的分层抽样不同，整群抽样只调查完整的群。例如，要估计一个大城市中14岁学生的平均身高，我们可能从100所学校中随机抽取10所，并测量这10所学校中所有14岁学生的身高。主要风险是如果被选中的群并不真正具有代表性，样本可能会产生偏差；然而，增加群的数量可以降低这种风险。
6. Convenience Sampling | 便利抽样
Convenience sampling involves selecting participants who are readily available and easy to reach. Examples include interviewing shoppers at a single mall, surveying your own classmates, or placing a data-collection form outside the school office. While quick and inexpensive, this method is non-probability sampling, meaning not every member of the population has a known chance of being included.
便利抽样涉及选择那些容易接触到的参与者。例如在某个商场采访购物者、调查自己班上的同学，或者把数据收集表放在学校办公室外面。尽管这种方法快捷且成本低，但它属于非概率抽样，意味着总体中并非每个成员都有已知的入选机会。
The results from convenience samples often suffer from severe bias because the participants may share characteristics that do not mirror the wider population. Nonetheless, convenience samples can be useful in pilot studies or for generating hypotheses, but their findings cannot be safely generalised.
便利抽样的结果往往严重偏离真实情况，因为参与者可能具有某些共同特征，而这些特征不能反映更广泛的总体。尽管如此，便利样本在试点研究或提出假设时可能有用，但其结论不能安全地推广到整体。
7. Voluntary Response Sampling | 自愿响应抽样
Voluntary response sampling, also called self-selection sampling, lets individuals choose whether to participate. Common forms include online polls, phone-in surveys, or feedback forms where people are invited to respond. Because participation is optional, only those with strong opinions or personal interest tend to take part, causing systematic over-representation of extreme views.
自愿响应抽样，也称自选抽样，允许个体自行决定是否参与。常见的形式包括在线投票、电话热线调查或邀请人们填写反馈表。由于参与是自愿的，只有那些抱有强烈意见或个人兴趣的人才会参与，这会导致极端观点在样本中被系统性放大。
This leads to voluntary response bias, making the sample unrepresentative. In IB Mathematics examinations and coursework, it is crucial to identify whether data were collected via voluntary response, and to argue how this affects the reliability of conclusions. For instance, a restaurant that emails a satisfaction survey may receive replies mostly from very satisfied or very dissatisfied customers, while the silent majority is excluded.
这会导致自愿响应偏差，使得样本缺乏代表性。在IB数学考试和课程作业中，识别数据是否通过自愿响应方式收集，并论证这对结论可靠性的影响，是至关重要的。例如，一家餐厅通过电子邮件发放满意度调查，可能只会收到来自非常满意或非常不满意顾客的回复，而沉默的大多数则被排除在外。
8. Sampling Bias | 抽样偏差
Sampling bias arises when the sample collected does not faithfully represent the population. It can occur in many forms: selection bias (certain groups are systematically excluded or under-represented), non-response bias (selected individuals do not respond, and their non-response is related to the variable of interest), and measurement bias (the data collection method consistently overstates or understates values).
抽样偏差发生在所收集的样本不能真实反映总体的情况下。它可以表现为多种形式：选择偏差（某些群体被系统性地排除或代表性不足）、无应答偏差（被选中的个体没有回应，且其不回应与所关注的变量相关）以及测量偏差（数据收集方法系统地高估或低估了数值）。
Recognising bias is a core skill in IB Mathematics. Even a large sample size cannot correct for a biased sampling method. For example, using telephone directories to select participants excludes people without landlines, a selection bias. Likewise, a poorly worded survey question can introduce measurement bias. Always ask: Does every member of the target population have an equal or known chance of being included?
识别偏差是IB数学的一项核心技能。即使样本量很大，也无法纠正有偏差的抽样方法。例如，使用电话簿选择参与者会排除没有固定电话的人，这就是选择偏差。同样，措辞不当的调查问题会引入测量偏差。我们始终要问：目标总体中的每个成员是否都有相等或已知的被选入机会？
9. Sample Size and Error | 样本大小与误差
The size of the sample, denoted n, directly influences the precision of statistical estimates. Larger samples generally produce smaller sampling error — the natural variability that occurs because we examine only a part of the population. One common measure is the standard error of the mean, which decreases as n increases:
样本的大小（记为 n）直接影响统计估计的精确度。较大的样本通常会产生较小的抽样误差——即由于只检查总体的一部分而产生的自然变异性。一种常见的度量是均值的标准误，它随着 n 的增大而减小：
Standard Error (SE) = σ / √n
Here, σ is the population standard deviation. This formula shows that doubling the precision requires quadrupling the sample size. However, practical constraints such as time, cost, and accessibility often limit n. A well-chosen sample of 200 may outperform a biased sample of 2000. In IB coursework, you should justify your chosen sample size based on the desired margin of error and resource limitations.
在此，σ 是总体标准差。这个公式表明，若要将精度提高一倍，样本量需要增加到原来的四倍。然而，时间、成本和可及性等实际限制通常制约着 n 的大小。一个精心选取的200人样本可能优于一个有偏差的2000人样本。在IB课程作业中，你应当根据期望的误差范围和资源限制，为你所选择的样本量提供合理依据。
10. Comparison of Sampling Methods | 抽样方法比较
The table below summarises the main sampling techniques, their implementation, bias potential, and typical scenarios. Choosing the right method involves balancing accuracy, cost, and practical feasibility. Note that non-probability methods (convenience, voluntary response) should be avoided when representativeness is required.
下表总结了主要的抽样技术、它们的实施方式、偏差可能性和典型应用场景。选择正确的方法需要在准确性、成本和实际可行性之间取得平衡。请注意，当需要代表性时，应避免使用非概率方法（便利抽样、自愿响应抽样）。
| Method (方法) | Selection Process (选择过程) | Bias Risk (偏差风险) | Best For (最适合) |
|---|---|---|---|
| Simple Random (简单随机) | Random numbers; equal probability | Low if fully random | Homogeneous, listable populations |
| Systematic (系统) | Every k-th unit from random start | Low to medium (beware patterns) | Ordered lists, production lines |
| Stratified (分层) | Random within each proportional stratum | Very low if strata well defined | Diverse groups, known subgroup sizes |
| Cluster (整群) | Randomly select whole clusters | Medium (if clusters differ) | Geographically spread, natural groups |
| Convenience (便利) | Who is easy to access | High | Pilot studies only |
| Voluntary Response (自愿响应) | Self-selected participants | Very high | Opinion polling with caution |
In examination questions, you may be asked to identify the sampling method used in a scenario and discuss its advantages and limitations. Use this comparison as a quick reference to justify your reasoning with precise statistical terminology.
在考试题目中，你可能会被要求识别某个场景中使用的抽样方法，并讨论其优缺点。使用这个比较表作为快速参考，用准确的统计术语来论证你的推理。
11. Applications in IB Mathematics | IB数学中的应用
Sampling principles appear throughout the IB Mathematics syllabus, especially in the Applications and Interpretation course (Topic 4: Statistics and Probability). Students are expected to design simple sampling strategies for data collection in the Internal Assessment (IA). Whether surveying customer satisfaction, measuring plant growth, or analysing traffic patterns, the IA must detail how the sample was obtained and address potential sources of bias.
抽样原则贯穿整个IB数学大纲，特别是在“应用与解释”课程（主题4：统计与概率）中。学生需要为内部评估的数据收集设计简单的抽样策略。无论是调查顾客满意度、测量植物生长，还是分析交通模式，IA都必须详细说明样本是如何获得的，并讨论潜在的偏差来源。
Typical exam question stylings include: 'Suggest a suitable sampling method for this investigation, giving a reason,' or 'Explain why voluntary response sampling might lead to unreliable conclusions.' A well-structured answer will name the method, describe how it would be implemented, and critique its strengths and weaknesses in that specific context.
典型的考题风格包括：“为这项调查建议一种合适的抽样方法并给出理由”，或“解释为什么自愿响应抽样可能导致不可靠的结论”。一份结构合理的答案应指明方法、描述如何实施，并针对具体情境评述其优缺点。
12. Summary and Key Takeaways | 总结与要点
Sampling is about choosing a representative subset to infer truths about an entire population. The main probability-based methods — simple random, systematic, stratified, and cluster — each have unique strengths and are designed to minimise bias. Non-probability methods like convenience and voluntary response sampling are easy but often yield misleading results. Always assess the sampling frame, randomisation process, and potential biases before trusting any data-driven conclusion.
抽样的核心在于选择一个具有代表性的子集，以推断整个总体的真相。主要的概率抽样方法——简单随机、系统、分层和整群——各有独特优势，旨在最大限度地减少偏差。便利抽样和自愿响应抽样等非概率方法虽然简单，但往往得出误导性的结果。在相信任何基于数据的结论之前，务必评估抽样框、随机化过程以及潜在的偏差。
For IB success, memorise the key definitions, be able to match a scenario to the optimal sampling technique, and practise writing justifications that reference bias, practicality, and sample size. With a solid understanding of these concepts, you'll be well-prepared to tackle statistical questions and produce high-scoring internal assessments.
为了在IB中取得成功，记住关键定义，能够将场景与最优抽样技术相匹配，并练习撰写引用偏差、实用性和样本大小的理由。扎实掌握这些概念后，你将能从容应对统计问题，并撰写出高分的内部评估。
Published by TutorHao | IB Mathematics 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