📚 Statistical Sampling and Fieldwork Investigation Skills | 统计抽样与实地调查技能
In Edexcel A-Level Mathematics, statistical fieldwork investigation is not about collecting data in a geography field trip; it is the set of skills used to design, carry out, and interpret a data-based study. From choosing a sample to testing a hypothesis, each step must be justified in context.
在 Edexcel A-Level 数学中，统计实地调查并不是地理实地考察中的数据收集，而是设计、实施和解释基于数据的研究的一套技能。从选择样本到检验假设，每一步都必须结合实际情况加以说明。
1. Population, Sample and Census | 总体、样本与普查
The population is the entire set of individuals or items that we want to investigate. A census collects data from every member of the population, while a sample selects only a subset of it.
总体是我们想研究的全部个体或项目的集合。普查收集总体中每个成员的数据，而样本只选择其中的一个子集。
A parameter is a numerical value calculated from the whole population, such as the population mean μ. A statistic is the corresponding value calculated from a sample, such as the sample mean x̄.
参数是从整个总体计算出的数值，例如总体均值 μ。统计量是从样本中计算出的对应数值，例如样本均值 x̄。
Using a sample saves time and cost, but it introduces sampling variability and uncertainty. Edexcel exam questions often ask you to compare a census with a sample in terms of accuracy, cost, and practicality.
使用样本可以节省时间和成本，但会引入抽样变异性和不确定性。Edexcel 考题经常要求从准确性、成本和可行性方面比较普查与样本。
2. Sampling Methods: Simple Random Sampling | 抽样方法：简单随机抽样
A simple random sample is one in which every member of the population has an equal chance of being selected. This can be achieved using random number tables or a calculator’s random number function.
简单随机样本是指总体中每个成员被选中的机会都相等的样本。这可以通过随机数表或计算器的随机数功能来实现。
One advantage is that it removes selection bias and is easy to understand. However, it requires a complete sampling frame, and it may be impractical for very large or geographically spread populations.
其优点是可以消除选择偏差且易于理解。然而，它需要一个完整的抽样框，对于非常大或地理分布很广的总体可能不切实际。
In a fieldwork-style investigation, simple random sampling works well when you can list every participant or item, but it does not guarantee that subgroups are represented proportionally.
在实地调查式研究中，当你能够列出每个参与者或项目时，简单随机抽样效果很好，但它不能保证各个子群体按比例得到代表。
3. Stratified and Systematic Sampling | 分层抽样与系统抽样
Stratified sampling divides the population into distinct groups, called strata, such as age bands or income categories. A random sample is then taken from each stratum in proportion to its size in the population.
分层抽样将总体划分为不同的组，称为层，例如年龄段或收入类别。然后按各层在总体中的比例从每层中随机抽样。
This method ensures that each subgroup is represented, which improves the precision of estimates for the whole population. The main drawback is that it requires clear information about the strata and can be more complex to organise.
这种方法确保每个子群体都有代表，从而提高了对总体估计的精度。主要缺点是它需要关于各层的清晰信息，并且组织起来可能更复杂。
Systematic sampling selects every kth member after a random starting point. For example, you might choose every 10th person on a register.
系统抽样在随机起点之后每隔 k 个成员选择一个。例如，你可以选择名册上每隔 10 个人选一个。
It is quick and convenient, but it can introduce bias if the list has a hidden pattern that matches the sampling interval. You must always check that the ordering is not periodic.
它快速方便，但如果列表存在与抽样间隔相匹配的隐藏模式，则可能引入偏差。你必须始终检查排序是否具有周期性。
4. Quota, Cluster and Opportunity Sampling | 配额抽样、整群抽样与机会抽样
Quota sampling sets fixed numbers, or quotas, of individuals from different groups to interview or observe. It is frequently used in market research because it is cheap and does not need a full sampling frame.
配额抽样设定不同群体的固定数量或配额来进行访谈或观察。它经常在市场研究中使用，因为它成本低且不需要完整的抽样框。
However, quota sampling is not random, so it can suffer from interviewer bias and cannot be used to calculate sampling errors in the same way as random methods.
然而，配额抽样不是随机的，因此可能存在访谈者偏差，并且不能像随机方法那样计算抽样误差。
Cluster sampling divides the population into clusters, such as schools or postcode areas, and randomly selects some clusters. All members within the chosen clusters are then included.
整群抽样将总体划分为群，例如学校或邮政编码区域，然后随机选择一些群。被选中群内的所有成员都包括在内。
It is useful when a full list of individuals is unavailable, but it can increase sampling error if clusters are very different from one another.
当没有完整的个体名单时，这种方法很有用，但如果群之间差异很大，它可能会增加抽样误差。
5. Bias, Error and Reliability | 偏差、误差与可靠性
Sampling error is the natural variation that occurs because only part of the population is observed. Non-sampling error includes mistakes in data collection, non-response, and measurement errors.
抽样误差是由于只观察了总体的一部分而产生的自然变异。非抽样误差包括数据收集中的错误、无回应和测量误差。
Bias is a systematic distortion in the results. Common sources include leading questions, voluntary response, undercoverage, and convenience sampling.
偏差是结果中的系统性扭曲。常见来源包括诱导性问题、自愿回应、覆盖不足和便利抽样。
Reliability refers to whether the method would produce similar results if repeated. A reliable investigation should use a clear sampling protocol and minimise subjective judgement where possible.
可靠性指该方法在重复进行时是否能产生相似的结果。可靠的调查应使用明确的抽样程序，并尽可能减少主观判断。
6. Designing a Fieldwork Investigation | 设计实地调查
Before collecting data, you must state a clear aim and, where appropriate, a null hypothesis H₀ and an alternative hypothesis H₁. For example, H₀: there is no correlation between hours of revision and test score.
在收集数据之前，你必须说明明确的目标，并在适当的情况下说明零假设 H₀ 和备择假设 H₁。例如，H₀：复习时间与考试成绩之间没有相关性。
A pilot study is a small trial run that helps identify problems with the questionnaire, instructions, or data collection method before the main survey. It improves validity and reduces wasted effort.
试点研究是小规模的试运行，有助于在主要调查之前发现问卷、说明或数据收集方法中的问题。它能提高有效性并减少无效努力。
Question design is also a key fieldwork skill. Questions should be clear, unbiased, and either open or closed depending on the type of data you need.
问题设计也是一项关键的实地调查技能。问题应当清晰、无偏，并根据所需数据的类型设计为开放式或封闭式。
7. Data Types and Measurement Scales | 数据类型与测量尺度
Quantitative data are numerical and can be discrete or continuous. Discrete data take distinct values, such as the number of siblings, while continuous data can take any value in an interval, such as height or time.
定量数据是数值型的，可以是离散的或连续的。离散数据取不同的值，例如兄弟姐妹的数量；连续数据可以在一个区间内取任意值，例如身高或时间。
Qualitative data are non-numerical and describe categories or attributes, such as eye colour or type of transport. These can be further classified as nominal or ordinal.
定性数据是非数值的，描述类别或属性，例如眼睛颜色或交通方式。这些数据可以进一步分为名义数据和有序数据。
Nominal data have no natural order, such as blood type. Ordinal data have a natural order but unequal intervals, such as satisfaction ratings from ‘very unhappy’ to ‘very happy’.
名义数据没有自然顺序，例如血型。有序数据有自然顺序但间隔不等，例如从 ‘非常不满意’ 到 ‘非常满意’ 的满意度评分。
8. Cleaning and Presenting Data | 数据清洗与展示
Cleaning data means checking for obvious errors, missing values, and outliers before analysis. An outlier is a value that lies well outside the overall pattern and can distort the mean and standard deviation.
数据清洗意味着在分析之前检查明显的错误、缺失值和异常值。异常值是远远偏离整体模式的值，可能会扭曲均值和标准差。
Box plots are useful for showing the median, quartiles, and outliers. Histograms show the shape of a distribution, while cumulative frequency graphs help estimate percentiles and medians.
箱线图用于显示中位数、四分位数和异常值。直方图显示分布的形状，而累积频率图有助于估计百分位数和中位数。
When presenting fieldwork data, always label axes, include units, and choose a chart that matches the data type. This is a common requirement in Edexcel statistics questions.
在展示实地调查数据时，始终标注坐标轴、标明单位，并选择与数据类型匹配的图表。这是 Edexcel 统计题中的常见要求。
9. Correlation and Regression in Context | 相关与回归在实地调查中的应用
Correlation measures the strength and direction of a linear relationship between two variables. The product moment correlation coefficient r always lies between −1 and 1.
相关衡量两个变量之间线性关系的强度和方向。积矩相关系数 r 始终位于 −1 和 1 之间。
A value of r close to 1 indicates strong positive correlation, while r close to −1 indicates strong negative correlation. However, correlation does not imply causation.
r 接近 1 表示强正相关，r 接近 −1 表示强负相关。然而，相关并不意味着因果。
Regression gives the equation of the line of best fit, often written as y = a + bx, where b is the gradient and a is the intercept. This line can be used to make predictions within the range of the data.
回归给出了最佳拟合线的方程，通常写作 y = a + bx，其中 b 是斜率，a 是截距。该直线可用于在数据范围内进行预测。
Extrapolation beyond the observed range is risky because the linear trend may not continue. In fieldwork conclusions, always comment on the reliability of any prediction.
超出观测范围的外推是有风险的，因为线性趋势可能不会持续。在实地调查结论中，始终要对任何预测的可靠性加以评论。
10. Hypothesis Testing | 假设检验
A hypothesis test in A-Level Mathematics is a structured way to decide whether a sample provides enough evidence to reject a null hypothesis H₀ in favour of an alternative H₁.
A-Level 数学中的假设检验是一种结构化方法，用于判断样本是否提供了足够的证据来拒绝零假设 H₀ 而支持备择假设 H₁。
You are usually given a significance level α, such as 0.05 or 0.01. This is the probability of rejecting H₀ when it is actually true, called a Type I error.
通常会给出显著性水平 α，例如 0.05 或 0.01。这是在 H₀ 实际为真时拒绝它的概率，称为第一类错误。
For binomial and normal tests, compare the calculated p-value with α. If the p-value is less than α, there is sufficient evidence to reject H₀; otherwise there is insufficient evidence.
对于二项分布和正态分布检验，将计算出的 p 值与 α 进行比较。如果 p 值小于 α，则有足够证据拒绝 H₀；否则证据不足。
Always write your conclusion in the context of the original problem. Edexcel examiners expect the phrase ‘there is evidence to suggest…’ or ‘there is insufficient evidence to suggest…’.
始终结合原始问题的背景写出结论。Edexcel 考官期望使用 ‘有证据表明……’ 或 ‘证据不足，无法表明……’ 这样的表述。
11. Evaluation and Conclusions | 评价与结论
A strong fieldwork investigation ends with an evaluation of the method, data quality, and validity of the findings. You should discuss limitations and suggest realistic improvements.
一项出色的实地调查以对方法、数据质量和结论有效性的评价收尾。你应讨论局限性并提出切实可行的改进建议。
Common limitations include small sample size, sampling bias, non-response, and confounding variables. Suggesting stratified sampling or a larger sample often addresses these issues.
常见的局限包括样本量小、抽样偏差、无回应和混杂变量。建议使用分层抽样或更大的样本通常可以解决这些问题。
Finally, link your conclusion back to the original aim and hypothesis. State whether the data support the hypothesis, and acknowledge that the result may not apply beyond the sampled population.
最后，将你的结论与最初的目标和假设联系起来。说明数据是否支持该假设，并承认结果可能不适用于样本之外的总体。
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