📚 GCSE Cambridge Statistics: Experimental and Practical Assessment Key Points | GCSE 剑桥统计学：实验与实践考核要点
Mastering the practical side of statistics is essential for success in Cambridge GCSE Statistics. This guide covers the key concepts you must know for planning investigations, collecting data, avoiding bias, and drawing valid conclusions. Whether you are designing a questionnaire, choosing a sampling method, or evaluating an experiment, these points align closely with the exam's emphasis on experimental design and practical data handling.
掌握统计学的实践应用是剑桥 GCSE 统计学考试成功的关键。本文涵盖了你必须了解的规划调查、收集数据、避免偏差以及得出有效结论的核心概念。无论你在设计问卷、选择抽样方法还是评估一项实验，这些要点都与考试对实验设计和实际数据处理能力的考查高度吻合。
1. The Statistical Investigation Cycle | 统计探究周期
The foundation of any practical statistical work is the investigation cycle, often remembered as PPDAC: Problem, Plan, Data, Analysis, Conclusion. You should be able to describe each stage and apply it to real-world scenarios.
任何统计实践工作的基础都是探究周期，通常记为 PPDAC：问题、计划、数据、分析、结论。你应能描述每个阶段并将其应用于真实场景。
The Problem stage involves clearly stating the question or hypothesis you want to investigate. A well-formulated question is specific, measurable, and achievable.
问题阶段需要清晰地陈述你想研究的疑问或假设。一个好的研究问题是具体、可测量且可实现的。
In the Plan stage, you decide how to collect data: what to measure, the target population, sampling method, and how to control for variables. This is the most critical part for practical marks.
在计划阶段，你要决定如何收集数据：测量什么、目标总体是什么、采用何种抽样方法以及如何控制变量。这是实践考核最关键的部分。
Data collection follows the plan, after which you analyse the data using appropriate charts, averages, and spread measures. The final Conclusion must link back to the original problem and discuss limitations.
数据收集按计划执行，之后你用合适的图表、平均数和离散程度指标分析数据。最后的结论必须回扣最初的问题，并讨论局限性。
2. Types of Data | 数据类型
Recognising data types helps decide which display and analysis are appropriate. Data can be categorical (qualitative) or numerical (quantitative), and numerical data can be discrete or continuous.
识别数据类型有助于决定采用哪种图表和分析方法。数据分为类别（定性）数据和数值（定量）数据，数值数据又分离散型和连续型。
Categorical data, such as eye colour or mode of transport, is best summarised using mode and displayed with bar charts or pie charts. Numerical discrete data, like number of siblings, is counted and takes whole-number values.
类别数据（如眼睛颜色或交通方式）最好用众数总结，并用条形图或饼图展示。数值离散型数据（如兄弟姐妹数量）是可数的，取整数值。
Continuous data, such as heights or times, is measured and can take any value in a range. It is often grouped into class intervals for frequency tables and histograms. Understanding this distinction stops you from applying incorrect statistical techniques in the exam.
连续型数据（如身高或时间）是测量得到的，可在一定范围内取任意值。通常将其分组成区间，制作频数表和直方图。理解这一区别能避免你在考试中使用错误的统计方法。
3. Planning a Survey or Experiment | 规划调查或实验
A statistical survey collects data about a population by asking questions, while an experiment actively imposes a treatment to observe an effect. Both require careful planning to produce reliable evidence.
统计调查通过提问收集关于总体的数据，而实验则主动施加某种处理以观察效果。两者都需要周密规划才能提供可靠证据。
When planning a survey, define your target population clearly. For example, 'all Year 11 students in my school' is more precise than 'teenagers'. This affects the sampling frame and generalisability of conclusions.
规划调查时，要清晰地定义目标总体。例如，“我校所有11年级学生”比“青少年”更精确。这会影响到抽样框和结论的普适性。
In an experiment, identify the factor you will change (independent variable) and the outcome you will measure (dependent variable). All other variables must be kept constant to establish a causal link.
在实验中，确定你要改变的因素（自变量）和你要测量的结果（因变量）。所有其他变量必须保持不变，以建立因果关系。
Always list the equipment, materials, and step-by-step procedure you will follow. The exam may ask you to critique a plan or suggest improvements, so look for missing controls or unclear measurement methods.
始终列出你将要使用的设备、材料以及分步操作流程。考试可能要求你评判一个计划或提出改进建议，因此要留意缺失的对照或不清楚的测量方法。
4. Sampling Methods | 抽样方法
Selecting a representative sample is essential to draw valid conclusions about a population. You must be able to describe and evaluate different sampling techniques.
选取一个有代表性的样本对于得出关于总体的有效结论至关重要。你必须能描述并评价不同的抽样技术。
| Method | Description | Advantage / Disadvantage |
|---|---|---|
| Simple Random | Every member has an equal chance of being chosen, e.g. using a random number generator. | Unbiased but may be impractical for large populations. |
| Stratified | Population divided into groups (strata), with a random sample taken from each in proportion to size. | Ensures representation of specific subgroups; requires knowledge of population structure. |
| Systematic | Selecting every nth member from a list after a random start. | Simple to implement, but can introduce bias if the list has a hidden pattern. |
| Quota | Interviewers choose a set number (quota) of individuals with particular characteristics. | Quick and cheap, but is non-random and prone to interviewer bias. |
| Opportunity (Convenience) | Sampling those who are easily available, e.g. people walking past a shop. | Very easy, but highly likely to produce a biased, unrepresentative sample. |
In experimental contexts, random allocation of subjects to treatment groups is crucial to reduce bias. You should also consider sample size: a larger sample generally gives more reliable results but costs more time and money.
在实验情境中，将受试者随机分配到处理组对减少偏差至关重要。你还应考虑样本量：较大样本一般能给出更可靠的结果，但需要更多时间和金钱。
5. Designing Questionnaires | 设计问卷
A poorly designed questionnaire can ruin a survey. Questions must be clear, unbiased, and easy to answer. You need to identify common flaws and write good questions.
糟糕的问卷设计会毁掉调查。问题必须清晰、无偏见且易于回答。你需要识别常见缺陷并撰写好的问题。
Avoid leading questions such as 'Don't you agree that homework is a waste of time?' Instead, ask neutrally: 'What is your opinion on the amount of homework?' Also avoid double-barrelled questions that ask two things at once, e.g. 'Do you enjoy maths and science?'
避免引导性问题，如“你难道不同意家庭作业是浪费时间吗？”。相反，应保持中立：“你对家庭作业的数量有何看法？”也要避免一次问两件事的双重问题，如“你喜欢数学和科学吗？”。
Response options should be exhaustive and mutually exclusive. A question about age should include all possible ranges without overlap. Provide a 'prefer not to say' option where sensitive data is collected.
回答选项应详尽且互斥。关于年龄的问题应包含所有可能的范围且没有重叠。在收集敏感数据时，提供“不愿透露”选项。
Always consider whether questions should be open-ended (free text) or closed (tick boxes). Closed questions are easier to analyse but may limit expression. In GCSE exams, you will often be asked to criticise a given questionnaire and suggest improvements.
始终考虑问题应为开放式（自由文本）还是封闭式（勾选框）。封闭式问题更易于分析，但可能限制表达。在 GCSE 考试中，你常会被要求评判一份给出的问卷并提出改进建议。
6. Data Collection Sheets & Recording | 数据收集表与记录
Organising raw data efficiently from the start saves time and prevents errors. A well-structured data collection sheet or tally chart is a fundamental practical skill.
从一开始就有效地整理原始数据能节省时间并防止错误。结构良好的数据收集表或划记表是一项基本的实践技能。
A tally chart uses tallies (strokes grouped in fives) to record frequencies for categorical or discrete data. When designing one, include clear headings for the variable and the tally column, and leave enough space for large frequencies.
划记表使用划记（每五个为一组）来记录类别或离散数据的频数。设计时，应包含清晰的变量标题和划记列，并为较大的频数留出足够空间。
For measurements, use a results table with columns for the independent variable, dependent variable, and any repeated readings or calculated averages. Always record units and the same precision for all measurements (e.g. all to 0.1 cm).
对于测量数据，使用包含自变量、因变量以及重复读数或计算平均值的多列结果表。务必记录单位，并保持所有测量值精度一致（如都精确到 0.1 cm）。
In an experiment, recording conditions such as temperature or time of day can be part of good practice, showing awareness of control variables. The exam may provide a poorly designed table and ask you to correct it.
在实验中，记录温度或一天中的时刻等条件是一种良好做法，表明你对控制变量有意识。考试可能会提供一个设计不佳的表格让你改正。
7. Pilot Studies & Pre-testing | 试点研究与预测试
A pilot study is a small-scale trial run of a survey or experiment before doing the main data collection. It is a hallmark of practical rigour and often gains marks when mentioned.
试点研究是在主要数据收集之前，对调查或实验进行的小规模试运行。它是实用严谨性的标志，提到这一点通常能得分。
Pilot studies help identify problems with question wording, response options, or data collection sheets. They also give an estimate of the time required and variability, which informs sample size decisions.
试点研究有助于发现问卷措辞、回答选项或数据收集表的问题。它们还能估算所需时间和变异性，为样本量决策提供依据。
After a pilot, you might reword ambiguous questions, adjust the length of the questionnaire, or calibrate measuring instruments. This refinement increases validity and reliability in the final investigation.
试点之后，你可能会修改模糊的问题、调整问卷长度或校准测量仪器。这种改进能提升最终探究的效度和信度。
8. Sources of Bias & Error | 偏差与误差来源
Bias occurs when the data collected systematically over- or underestimates the true value. Recognising and minimising bias is a key objective in practical assessments.
偏差是指收集的数据系统性地高估或低估真实值。识别并最小化偏差是实践考核的一个关键目标。
Selection bias arises when the sample does not reflect the population, e.g. only surveying people in a library to assess reading habits. Measurement bias happens if an instrument is faulty or if a question is leading.
当样本不能反映总体时，就产生了选择偏差，例如只在图书馆调查人们的阅读习惯。如果测量工具有故障或问题具有引导性，则会产生测量偏差。
Non-response bias occurs when a significant proportion of selected individuals do not respond, and their views differ from those who do. You should suggest using follow-ups or incentives to reduce this.
当有相当大比例的被选对象没有回应，且他们的观点与回应者不同时，就产生无回应偏差。你应建议使用跟进或激励来减少这种情况。
Random error is natural variability; it can be reduced by increasing sample size or taking repeated measurements. Systematic error is consistent inaccuracy (e.g. a scale always reading 2 g too high) and must be corrected by calibrating equipment or adjusting methods.
随机误差是自然变异，可通过增加样本量或重复测量来减小。系统误差是一致的不准确（如天平读数总是偏高 2 g），必须通过校准设备或调整方法来纠正。
9. Controlling Variables in Experiments | 实验中的变量控制
For a fair test, only the independent variable should be deliberately changed. All other variables that could affect the outcome must be controlled or kept constant.
要进行公平测试，只能刻意改变自变量。所有其他可能影响结果的变量都必须加以控制或保持不变。
Control variables might include temperature, humidity, time, or material type. Write them explicitly: 'Keep the same volume of water (200 ml) for each trial' or 'Use the same brand of measuring cylinder'.
控制变量可能包括温度、湿度、时间或材料类型。要明确写出它们，例如：“每次试验保持水量相同 (200 ml)” 或 “使用同一品牌的量筒”。
A control group, which receives no treatment or a placebo, provides a baseline. This is vital in experiments testing effectiveness of fertilisers, learning methods, or medical treatments.
不接受处理或接受安慰剂的对照组提供了一个基线。这在测试肥料、学习方法或医学治疗效果的实验中至关重要。
Random allocation of subjects to treatment and control groups helps distribute unknown confounding variables evenly, strengthening causal claims. The examiners expect you to recognise when this is missing.
将受试者随机分配到处理组和对照组有助于均匀分布未知的混杂变量，从而增强因果推断的可信度。考官期望你能识别出缺少这一步骤的情形。
10. Ethical Considerations | 伦理考量
Practical statistical work, especially involving people, must follow ethical guidelines. Consent, confidentiality, and avoidance of harm are core principles.
实践统计工作，尤其是涉及人的，必须遵循伦理准则。知情同意、保密和避免伤害是核心原则。
Participants should give informed consent, knowing the purpose of the study and how their data will be used. For minors, parental consent is necessary. They must also be free to withdraw at any time.
参与者应知情同意，了解研究目的及其数据将如何使用。对于未成年人，需获得父母同意。他们也必须能够随时自由退出。
Anonymise data by using codes instead of names. Never share individual data publicly. In observations, ensure you do not infringe on people's privacy without permission.
使用代码代替姓名来匿名化数据。切勿公开分享个人数据。在观察研究中，确保你不未经允许侵犯他人隐私。
Ethics are not only for human subjects; animal experiments must also minimise distress and follow the 3Rs (Replacement, Reduction, Refinement). Though rare in GCSE Statistics, you may be asked to comment on ethical issues in a given scenario.
伦理不限于人类对象；动物实验也必须尽量减少痛苦，遵循 3R 原则（替代、减少、优化）。虽然在 GCSE 统计学中不常见，但你可能要就给定情景中的伦理问题发表评论。
11. Presenting Data & Choosing Charts | 数据呈现与图表选择
Choosing the correct graphical display is a vital practical skill. The chart must suit the data type and highlight the patterns or comparisons you wish to make.
选择正确的图形展示是一项关键的实践技能。图表必须适合数据类型，并突出你想展示的规律或比较。
For categorical data, use bar charts (gaps between bars) or pie charts (showing proportions). Pictograms can be engaging but must use consistent symbols and a clear key. Avoid using 3D effects that distort proportions.
对于类别数据，使用条形图（条间有间隔）或饼图（显示比例）。象形图可以很吸引人，但必须使用一致的符号和清晰的图例。避免使用会扭曲比例的 3D 效果。
he vertical axis are required, where frequency density = frequency / class width. Do not confuse histograms with bar charts.
对于离散数值数据，垂直线图或条形图适合。对于连续型分组数据，需要使用纵轴为频数密度的直方图，其中频数密度 = 频数 / 组距。不要将直方图与条形图混淆。
Scatter graphs show relationships between two continuous variables. A line of best fit (not necessarily through the origin) helps describe correlation. Line graphs are for time-series data, with time on the horizontal axis.
散点图显示两个连续变量之间的关系。最佳拟合线（不必经过原点）有助于描述相关性。折线图适用于时间序列数据，横轴为时间。
Always label axes with variable names and units, give a title, and use an appropriate scale that spreads data across the graph area. Poor scaling can lose marks in practical questions.
始终用变量名和单位标注坐标轴，给出标题，并使用能使数据分布于整个图形区域的合适刻度。不合理的刻度会在实践题目中失分。
12. Evaluating & Drawing Conclusions | 评估与得出结论
The final stage of the investigation cycle requires you to link findings back to the original question, evaluate the reliability of your data, and discuss limitations.
探究周期的最后阶段要求你将发现与原始问题联系起来，评估数据的可靠性，并讨论局限性。
A conclusion must state whether any pattern or difference found is likely to be genuine, using statistical language. For experiments, relate the conclusion to the hypothesis and refer to the data as evidence.
结论必须说明所发现的模式或差异是否可能是真实的，并使用统计语言。对于实验，将结论与假设联系起来，并引用数据作为证据。
You should comment on outliers – extreme values that do not fit the overall pattern. Decide whether to exclude them with justification (e.g. recording error vs. genuine variability).
你应评论异常值——不符合整体规律的极端值。判断是否排除它们并给出理由（例如是记录错误还是真实变异）。
Discuss sources of uncertainty and suggest how to improve the investigation: perhaps a larger sample, better randomisation, or more precise measurement. A good evaluation demonstrates critical thinking, which distinguishes top-level answers.
讨论不确定性的来源，并提出改进调查的建议：可能是更大的样本、更好的随机化或更精确的测量。好的评估能展现批判性思维，这是高分答案的标志。
Remember that a conclusion is not just a summary; it should interpret the results in context. Avoid overstating – correlational data does not prove causation unless a well-controlled experiment was conducted.
记住，结论不仅仅是总结；它应结合情境解释结果。避免夸大其词——除非进行了严格控制的实验，否则相关性数据并不能证明因果关系。
Published by TutorHao | Statistics 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