📚 AS CCEA Statistics: Key Points for Experimental/Practical Assessment | AS CCEA 统计：实验/实践考核要点
In AS CCEA Statistics, the ‘Experimental/Practical Assessment’ component focuses on your ability to plan, carry out, and interpret statistical investigations. Whether you are designing a survey, performing a simulation, or collating data from an experiment, you need to demonstrate a clear understanding of good practice. This article brings together the core ideas you must master for any practical or experimental task.
在 AS CCEA 统计课程中，”实验/实践考核”部分着重考察你规划、执行和解释统计调查的能力。无论你是在设计问卷、进行模拟实验还是整理实验数据，都需要展现出对良好实践准则的清晰理解。本文汇集了你在任何实践或实验任务中必须掌握的核心要点。
1. Understanding Variables and Data Types | 理解变量与数据类型
Before any practical work, identify whether your variables are qualitative (categorical) or quantitative (numerical). Quantitative variables can be further split into discrete and continuous. This choice dictates which diagrams and summary statistics are appropriate.
在开展任何实践工作之前，首先要判断变量是定性（分类）变量还是定量（数值）变量。定量变量又可细分为离散变量和连续变量。这一选择决定了后续适合使用哪种图表和概括统计量。
For example, shoe size is discrete numeric, height is continuous, and eye colour is categorical. Mixing up types can lead to meaningless averages or invalid charts.
例如，鞋码是离散数值变量，身高是连续变量，眼睛颜色是分类变量。混淆数据类型可能会导致计算出的平均数毫无意义，或者绘制的图表无效。
In an experimental setting, the response variable is the outcome you measure, and the explanatory variable is the one you manipulate or observe. Always state these clearly before collecting data.
在实验场景中，响应变量是你测量的结果量，解释变量是你操作或观察的变量。在收集数据之前，务必清晰地阐明这两种变量。
2. Designing Effective Questionnaires | 设计有效问卷
Questionnaires are a common practical task. Questions must be clear, unbiased, and easy to answer. Avoid leading questions such as ‘Don’t you agree that recycling is important?’, which encourage a particular response.
问卷调查是一项常见的实践任务。问题必须清晰、无偏见且易于回答。避免提问诸如”难道你不认为回收很重要吗？”之类的诱导性问题，它们会促使受访者给出特定回答。
Use closed questions with tick-box options where possible, as they are simpler to code and analyse. Provide exhaustive, non-overlapping response categories. For instance, age groups should cover all possibilities without gaps.
尽可能使用带有勾选框选项的封闭式问题，因为它们更容易编码和分析。提供详尽且互不重叠的回答类别。例如，年龄分组应覆盖所有可能情况且不留空隙。
Pilot your questionnaire on a small group to spot ambiguous wording or missing options. This refinement stage is essential for any practical assessment write-up.
先在小群体中试测你的问卷，以发现模棱两可的措辞或遗漏的选项。这个改进阶段对任何实践考核报告来说都至关重要。
3. Sampling Techniques in Practice | 实践中的抽样技术
You must be able to describe and justify your sampling method. Simple random sampling gives every member an equal chance, but you need a sampling frame. Stratified sampling ensures representation from each subgroup proportional to its size.
你必须能够描述并论证所采用的抽样方法。简单随机抽样赋予每个成员均等的机会，但需要一个抽样框。分层抽样则确保按比例从每个子群中获得具有代表性的样本。
When a sampling frame is unavailable, systematic sampling (selecting every k-th item) or quota sampling might be used, but you must discuss the risk of bias. For practical activities, always explain how you selected participants or items.
当抽样框不可用时，可能会采用系统抽样（每隔 k 个抽取一个）或定额抽样，但你必须讨论其中的偏差风险。在实践活动中，始终要说明你是如何选择参与者或物品的。
Sample size matters. A larger sample reduces variability, but practical constraints like time and cost must be acknowledged. Simply stating ‘sample size was 30 because it was convenient’ is not enough; link it to the precision needed.
样本量很重要。较大的样本能降低变异性，但必须承认时间和成本等现实约束。仅仅说”因为方便，样本量为 30″是不够的；要将其与所需的精度联系起来。
4. Planning a Statistical Experiment | 规划统计实验
An experiment involves deliberately changing one factor to see its effect on a response. You must design the experiment to compare a treatment group with a control group when possible. This allows you to attribute any difference to the treatment rather than to natural variation.
实验涉及有意地改变一个因素来观察其对响应的影响。在可能的情况下，你必须设计实验，将处理组与对照组进行比较。这样一来，就能将任何差异归因于处理因素，而非自然变异。
Define your null hypothesis and alternative hypothesis before collecting data, even in a practical investigation. For example, ‘H₀: There is no difference in mean growth between the two fertilisers.’
即使是在实践调查中，也要在收集数据前定义原假设和备择假设。例如，”H₀：两种肥料之间的平均生长量没有差异。”
Use blocking where appropriate to control for known sources of variation. If you suspect that age might affect the outcome, divide participants into age blocks and randomise within each block.
在适当的情况下使用区组来控制已知的变异来源。如果你怀疑年龄可能影响结果，可以将参与者分成不同的年龄区组，并在每个区组内随机分配。
5. Randomisation and Control | 随机化与控制
Randomisation is the backbone of a fair experiment. It helps balance out unknown confounding variables. For a practical assessment, clearly describe how you randomised: you might use random number tables, a calculator function, or drawing lots.
随机化是公正实验的基石。它有助于平衡未知的混杂变量。在实践考核中，要清晰地描述你是如何进行随机化的：可以使用随机数表、计算器函数或抽签等方式。
Control groups should receive no treatment or a placebo, keeping all other conditions identical. In observational studies where experiments are not possible, you can still match subjects on key characteristics to mimic control.
对照组应不接受处理或接受安慰剂，同时保持所有其他条件相同。在无法进行实验的观察性研究中，仍然可以就关键特征对受试者进行匹配，以模拟控制效果。
Blinding (single- or double-blind) reduces bias from participants’ or assessors’ expectations. Even in a simple school-based experiment, you can arrange for a colleague to measure outcomes without knowing which group a subject belongs to.
盲法（单盲或双盲）可减少来自参与者或评估者期望的偏差。即使是在简单的校园实验中，你也可以安排一位不知道受试者属于哪一组的同事来测量结果。
6. Collecting Data with Accuracy | 准确收集数据
Accuracy in measurement depends on the instrument, the operator, and the protocol. Record data to a consistent degree of precision. If you measure lengths to the nearest millimetre, all readings should follow that rule.
测量的准确性取决于仪器、操作者和规程。以一致的精度记录数据。如果你以毫米为最小刻度测量长度，那么所有读数都应遵循这一规则。
In any practical write-up, mention potential sources of measurement error, such as reaction time in stopping a stopwatch or parallax error when reading a scale. Distinguish between random errors (which can be reduced by averaging) and systematic errors (which bias all results in one direction).
在任何实践报告中，都要提及潜在的测量误差来源，例如停止秒表时的反应时间，或是读取刻度时的视差。要区分随机误差（可通过取平均值来减小）和系统误差（会使所有结果向一个方向偏倚）。
Use a well-designed data collection sheet or table with clear headings and units. This makes it easier to spot anomalies and saves time when transferring data to software or drawing graphs.
使用设计良好的数据采集表或表格，包括清晰的标题和单位。这样更容易发现异常值，并在将数据转移到软件或绘制图表时节省时间。
7. Presenting Data from Experiments | 展示实验数据
Choose the correct diagram for your data type. Bar charts for categorical data, histograms for continuous data, and scatter diagrams for paired numerical data. Always label axes, include titles, and use a key if necessary.
根据数据类型选择正确的图表。分类数据用条形图，连续数据用直方图，成对数值数据用散点图。务必标注坐标轴、添加标题，并在必要时使用图例。
In a practical assessment, you may be asked to draw a cumulative frequency curve and estimate quartiles. Plot the upper class boundaries against cumulative frequency, join points with a smooth curve, and read off values for the median and interquartile range.
在实践考核中，你可能需要绘制累积频率曲线并估计四分位数。在上类界处对应地标出累积频率点，用平滑曲线连接，然后读出中位数和四分位距的数值。
Box and whisker plots are excellent for comparing distributions from two or more experiments. They show the median, quartiles, and any outliers clearly. Comment on the central tendency, spread, and skewness when interpreting these plots.
箱形图非常适合比较两个或多个实验得出的分布。它们能清晰地展示中位数、四分位数以及任何异常值。在解读这些图时，要对集中趋势、离散程度和偏度进行评论。
8. Calculating and Interpreting Summary Statistics | 计算和解读概括统计量
From a practical data set, you should be able to compute mean, median, mode, range, interquartile range, and standard deviation. For the mean of a sample, use the formula:
x̄ = Σx / n
从实践数据集中，你应该能够计算平均数、中位数、众数、极差、四分位距和标准差。对于样本平均数，使用公式：x̄ = Σx / n。
For standard deviation, AS candidates often use a calculator but must understand the concept:
s = √[ Σ(x – x̄)² / (n – 1) ]
Choose which statistics are most suitable. The median and interquartile range are preferable for skewed data, while the mean and standard deviation are natural companions for symmetric distributions. Always justify your choice in the context of the experiment.
选择最合适的统计量。对于偏斜数据，中位数和四分位距是首选；对于对称分布，平均数与标准差是天然搭档。始终要在实验背景下为你的选择提供理由。
9. Simulation in Statistical Practice | 统计实践中的模拟
When a real experiment is impractical, simulation models can be used. For example, you might use random numbers to simulate the number of defective items in a batch or the behaviour of a queuing system. Describe the model’s assumptions clearly.
当真实实验难以实施时，可以使用模拟模型。例如，你可以用随机数来模拟一批产品中的缺陷品数量或排队系统的行为。需清晰地说明模型的假设。
In a practical assessment, you may be required to perform a simulation with dice, coins, or a spreadsheet. Explain what each digit or outcome represents, and run enough trials to get stable results. A small number of trials gives unreliable estimates.
在实践考核中，你可能需要用骰子、硬币或电子表格进行模拟。解释每个数字或结果所代表的含义，并进行足够多次数的试验以获得稳定的结果。试验次数过少会产生不可靠的估计。
Compare the simulated results with a theoretical probability where possible. For instance, simulate 200 coin tosses and compare the proportion of heads with 0.5. Discuss any discrepancies in terms of random variation.
尽可能将模拟结果与理论概率进行比较。例如，模拟 200 次抛硬币，将正面出现的比例与 0.5 进行比较。从随机变异的角度讨论任何差异。
10. Avoiding Bias in Practical Work | 避免实践工作中的偏差
Bias can creep in at every stage: design, data collection, and analysis. Selection bias occurs when the sample is not representative. Volunteer samples, for example, often attract people with strong opinions.
偏差可能在设计、数据收集和分析的每个阶段悄悄出现。当样本不具代表性时就会发生选择偏差。例如，志愿者样本往往会吸引持有强烈观点的人。
Measurement bias arises from faulty instruments or leading questions. Response bias happens when participants do not answer truthfully, perhaps to please the researcher. Acknowledge these risks and explain how you minimised them.
测量偏差源于有缺陷的仪器或诱导性问题。回答偏差发生在参与者没有如实作答时，或许是为了取悦研究者。要承认这些风险并解释你是如何将其最小化的。
In experimental design, confounding occurs when you cannot separate the effect of the explanatory variable from another factor. Careful randomisation and blocking help to separate these effects, making your conclusions more valid.
在实验设计中，当你无法将解释变量的影响与另一个因素的影响分开时，就会发生混杂。仔细的随机化和区组设计有助于分离这些影响，使你的结论更加有效。
11. Using Technology in Practical Assessments | 在实践考核中使用科技工具
Graphical calculators and spreadsheets are central to modern statistical practice. You should be able to enter data, produce summary statistics, and draw graphs without manual calculation. However, you must still show your understanding of the processes.
图形计算器和电子表格是现代统计实践的核心。你应该能够输入数据、生成概括统计量并绘制图表，而无需手动计算。但是，你仍然必须展示你对这些过程的理解。
For coursework-type tasks, screen shots of your spreadsheet formulas or calculator steps can demonstrate your methodology. Label everything so the examiner can follow your reasoning. A printout of a graph alone is not sufficient; annotate it with key findings.
对于课程作业类型的任务，你的电子表格公式截图或计算器操作步骤可以证明你的方法论。为所有内容添加标签，以便考官能够跟上你的推理。仅仅打印一张图表是不够的；要标注上关键发现。
When using random number generators, specify the seed (if applicable) so that your simulation can be replicated. Replicability is a hallmark of good statistical practice.
使用随机数生成器时，要指定种子（如适用），以便你的模拟可以复制。可复制性是良好统计实践的一个特征。
12. Writing Up and Interpreting Practical Findings | 撰写和解读实践发现
The write-up should tell a story: purpose, method, results, and conclusion. State clearly whether the findings support the original hypothesis and recognise the limitations of the investigation. Avoid claiming that an experiment proves a theory; instead say ‘the evidence suggests’.
报告应该讲述一个完整的故事：目的、方法、结果和结论。要清楚地说明研究发现是否支持最初的假设，并认识到调查的局限性。避免声称实验”证明”了某个理论；而应表述为”证据表明”。
Relate your conclusions back to the context. If you found that a new revision method did not lead to significantly higher test scores, discuss what might have caused this and suggest further experiments with a larger sample or stricter controls.
将你的结论与背景联系起来。如果你发现一种新的复习方法并没有带来显著更高的考试分数，请讨论可能的原因，并建议采用更大样本或更严格控制的进一步实验。
Peer review of your practical plan can reveal hidden flaws before you start. In an exam or assessment, even a description of how you would peer-review an investigation demonstrates a mature understanding of the statistical cycle.
在开始之前对你的实践计划进行同伴评审，可以揭示隐藏的缺陷。在考试或评估中，即使只是描述你将如何对一项调查进行同伴评审，也能展示出你对统计循环的成熟理解。
Published by TutorHao | Statistics 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