📚 IGCSE Statistics Practical Assessment Essentials | IGCSE 统计实践考核要点
In Cambridge IGCSE Statistics, the practical component (either Paper 2 or the coursework option) is designed to test your ability to design, conduct and review a real statistical investigation. Success requires a clear grasp of planning, data collection methods, sampling, graphical presentation and critical evaluation. This article walks you through every key stage of the practical assessment, highlighting what examiners look for and how to avoid common mistakes.
在剑桥 IGCSE 统计课程中，实践部分（试卷二或课程作业）旨在考查你设计、实施和回顾真实统计调查的能力。要想取得好成绩，必须清晰掌握规划、数据收集方法、抽样、图表展示和批判性评价。本文将带你遍历实践考核的每一个关键环节，点明考官关注的要点以及如何避免常见错误。
1. Planning a Statistical Investigation | 规划统计调查
Every successful investigation begins with a solid plan. You must first define a clear research question, such as “Does the amount of sleep affect students’ test scores?” Then formulate a null hypothesis (H₀) and an alternative hypothesis (H₁). For example, H₀: There is no relationship between sleep duration and test score. H₁: There is a relationship. Identify your independent variable (sleep duration) and dependent variable (test score), and decide how you will measure them. Also consider feasibility: can you collect the data ethically, within the available time, and with sufficient precision?
每一次成功的调查都始于周密的计划。你必须首先明确研究问题，例如”睡眠时间是否影响学生的考试成绩？”然后设定零假设（H₀）和备择假设（H₁）。例如，H₀：睡眠时长与考试成绩之间没有关系；H₁：两者存在关系。确定自变量（睡眠时长）和因变量（考试成绩），并决定如何测量它们。同时要考虑可行性：你是否能在伦理允许的范围内、在规定时间内以足够的精度收集数据？
It is also wise to anticipate the type of data you will generate — categorical, discrete or continuous — because this will influence your choice of graphs and summary statistics later. A simple planning table outlining variables, measurement tools and sample size will keep your investigation focused.
提前预判你将获得的数据类型（分类数据、离散型数据还是连续型数据）也是明智之举，因为这会影响你之后选择图表和汇总统计量。制作一个简单的规划表，列出变量、测量工具和样本量，能让调查始终围绕核心展开。
2. Designing Effective Questionnaires | 设计有效的问卷
If your investigation uses a survey, the questionnaire must be carefully designed to gather valid data. Use simple, unambiguous language and avoid leading questions. For instance, “How many hours of exercise do you do per week?” is better than “You do a lot of exercise, don’t you?” Include a mix of closed questions (with tick-box options) and, where necessary, open-ended questions for richer detail. Always carry out a pilot study on a small group to check that respondents interpret the questions as you intended.
如果你的调查使用问卷，则必须精心设计以收集有效数据。使用简单、无歧义的语言，避免引导性问题。例如，”你每周锻炼多少小时？”优于”你经常锻炼，对吧？”适当搭配封闭式问题（勾选选项）和必要时使用的开放式问题，以获取更丰富的细节。务必在小群体中先进行预测试，检验受访者是否按照你的意图理解问题。
Response options should be exhaustive and mutually exclusive. For example, age groups like “10-14, 15-19, 20+” leave no gaps. Also, consider anonymity — this encourages honest answers, especially on sensitive topics.
回答选项应穷尽所有可能且互斥。例如，年龄段设为”10–14岁，15–19岁，20岁及以上”就不会留下空白。此外，还要考虑匿名作答——这能鼓励受访者给出诚实回答，尤其是在敏感话题上。
3. Selecting Sampling Methods | 选择抽样方法
Choosing an appropriate sampling method is crucial for obtaining a representative sample. Simple random sampling gives every member of the population an equal chance of selection, which eliminates human bias but can be impractical for large populations. Stratified sampling divides the population into distinct groups (strata) and randomly selects from each in proportion to their size; this ensures all subgroups are represented. Systematic sampling selects every kth item after a random start — it is easier to implement but can introduce hidden periodicity. Quota sampling involves selecting individuals to meet specific quotas, which is quick and cheap but is not random and often leads to bias.
选择合适的抽样方法对于获得代表性样本至关重要。简单随机抽样让人人都有相等的中选机会，消除了人为偏差，但对大规模总体可能不切实际。分层抽样先将总体分成不同的组（层），再按比例从每一层中随机抽取样本，确保所有子群体都有代表。系统抽样在随机起点后每隔 k 个抽取一个样本，便于实施但可能引入隐蔽的周期性。配额抽样按特定配额选取个体，快捷低廉但并非随机，往往导致偏差。
Whichever method you choose, you must justify it and comment on its limitations. In your practical write-up, explain how you attempted to minimise sampling bias and discuss whether your sample size is large enough to draw meaningful conclusions.
无论选择哪种方法，都必须加以论证并评论其局限性。在实践报告里，要说明你如何尽量减小抽样偏差，并讨论样本量是否足够大，能否得出有意义的结论。
4. Data Collection in Practice | 实际数据收集
Data may be collected through experiments, observations, questionnaires or by using secondary sources. Primary data is collected directly for your investigation, giving you control over its accuracy. Secondary data, such as published statistics, can save time but must be checked for reliability. In an experiment, you manipulate the independent variable and measure its effect on the dependent variable, while keeping other conditions the same. In an observational study, you simply record what happens without interference.
数据可通过实验、观察、问卷或使用二手来源进行收集。一手数据是直接为你的调查收集的，你可以掌控其准确性。二手数据，如公开发布的统计数据，能节省时间但必须核查可靠性。在实验中，你操纵自变量并测量其对因变量的影响，同时保持其他条件不变。在观察性研究中，你仅仅记录发生的事情而不加干预。
Always use a data collection sheet designed in advance. Record measurements with appropriate units and consistent precision. If several people are collecting data, brief them to follow the same protocol to reduce observer variability.
务必使用预先设计好的数据记录表。记录测量值时应标出适当单位并保持一致的精度。如果有多人收集数据，要对他们进行简要培训，遵循相同规程，以减少观察者变异性。
5. Controlling Variables in Experiments | 实验中的变量控制
For any experiment-based investigation, controlling variables is essential to establish a causal link. Clearly identify the independent variable (the one you change), the dependent variable (the one you measure) and all control variables (factors kept constant). For example, when testing the effect of light on plant growth, light duration is the independent variable, growth height is the dependent variable, and water, temperature and soil type are control variables.
对于任何基于实验的调查，控制变量是确立因果关系的关键。要清楚地识别自变量（你改变的变量）、因变量（你测量的变量）以及所有控制变量（保持不变的因素）。例如，在测试光照对植物生长的影响时，光照时长是自变量，株高是因变量，而水分、温度和土壤类型属于控制变量。
Use control groups where possible, and replicate trials to reduce the impact of random variation. In your report, explain exactly how you controlled each variable and why any uncontrolled variables might have affected the results.
尽可能设置对照组，并重复实验以减少随机变异的影响。在报告中，准确地说明你如何控制每一个变量，并解释为何某些未受控的变量可能影响了结果。
6. Recording and Organising Data | 记录与整理数据
Raw data should be recorded in a well-structured tally sheet or table. Group discrete or continuous data into frequency tables where necessary, taking care to choose sensible class intervals (usually equal width) and to avoid overlapping boundaries. For example, use interval “10 ≤ x < 20" rather than "10 - 20", which is ambiguous.
原始数据应记录在结构清晰的计数表或表格中。必要时应将离散或连续数据分组到频数表里，注意选择合理的组距（通常宽度相等），并避免边界重叠。例如，使用区间”10 ≤ x < 20"而不是"10 – 20"，后者容易产生歧义。
Always check for outliers — values that are unusually far from the rest — and decide whether they are genuine or the result of recording errors. Document any data-cleaning steps you take, such as correcting an obvious typo or removing an impossible value.
始终检查是否存在异常值——那些与其他数据相距甚远的数值——并判断它们是真实的还是记录错误造成的。记录你采取的每项数据清理步骤，例如更正一个明显的输入错误或剔除一个不可能的数值。
7. Graphical Representation and Charts | 图表与图形展示
Choosing the correct diagram is vital for revealing patterns in your data. Use bar charts for categorical data, pie charts for showing proportions of a whole, histograms or frequency polygons for grouped continuous data, and scatter diagrams to explore relationships between two variables. Box-and-whisker plots are excellent for displaying the median, quartiles and outliers of a dataset.
选择正确的图表对揭示数据模式至关重要。分类数据使用条形图，展示整体中各部分比例使用饼图，分组连续数据使用直方图或频数多边形，探索两个变量间关系使用散点图。盒须图则非常适合展示数据集的中位数、四分位数和异常值。
Every graph must have a clear title, labelled axes with units where appropriate, and a key if needed. Draw graphs in pencil and use a ruler, as neatness contributes to accuracy and is rewarded in exams. In a scatter diagram, you may add a line of best fit to highlight the trend.
每张图都必须有清晰的标题、带单位的坐标轴标签，必要时还要有图例。用铅笔和直尺画图，整洁有利于提高准确性，在考试中也会得到加分。在散点图中，可以添加一条最佳拟合线来突出趋势。
8. Avoiding Bias and Minimising Error | 避免偏差与减少误差
Bias can enter at any stage of an investigation. Sampling bias occurs when some members of the population are more likely to be chosen than others. Non-response bias arises when selected individuals do not reply, and their missing answers differ systematically from those who do respond. Measurement bias results from faulty instruments or poorly worded questions. To reduce bias, use random sampling, write neutral questions, calibrate equipment and follow up on non-responses where feasible.
偏差可以渗入调查的任意阶段。当总体中的某些成员被选中的概率高于其他成员时，就会产生抽样偏差。无回复偏差发生在被选中的个体不回复，而他们的缺失答案与回访者的答案存在系统性差异时。测量偏差则由仪器故障或问题措辞不当引起。为减少偏差，应采用随机抽样，撰写中性问题，校准设备，并在可行时回访未受访者。
Also distinguish between random error (unpredictable fluctuations) and systematic error (consistent distortion). Repeating measurements and taking averages helps to lessen random error, but systematic error must be corrected at the source.
还要区分随机误差（不可预测的波动）和系统误差（一致的扭曲）。重复测量并取平均值有助于减弱随机误差，而系统误差必须从源头改正。
9. Analysing Data and Drawing Conclusions | 分析数据与得出结论
Once data is organised, calculate appropriate summary statistics. For central tendency, use the mean, median or mode depending on the data type and presence of outliers. For spread, compute the range, interquartile range (IQR = Q₃ – Q₁) or standard deviation. In bivariate analysis, draw a scatter diagram and if there is a roughly linear relationship, calculate the correlation coefficient r to measure its strength and direction. Remember, correlation does not imply causation.
数据整理好之后，计算合适的汇总统计量。对于集中趋势，根据数据类型和异常值情况使用平均数、中位数或众数。对于离散程度，计算极差、四分位距（IQR = Q₃ – Q₁）或标准差。在双变量分析中，画出散点图，如果存在大致线性关系，计算相关系数 r 来衡量其强度和方向。请记住，相关并不能推断因果。
Your conclusion must directly address the original hypothesis. State whether you find evidence to support H₁, or whether you fail to reject H₀. Use numerical summaries and charts to justify your reasoning, and avoid over-generalising beyond the sample.
你的结论必须直接回应初始假设。说明你是否找到支持备择假设 H₁ 的证据，还是未能拒绝零假设 H₀。用数字汇总和图表来佐证你的推理，避免将结论过度推广到样本以外。
10. Evaluating the Investigation | 评价调查
A thorough evaluation demonstrates higher-order thinking. Discuss limitations of your method, such as small sample size, possible response bias, or variables you could not control. For example, if you surveyed only Year 11 students, the findings may not apply to all secondary students. Suggest specific improvements: “Next time, I would use stratified sampling to ensure equal representation from each year group.” Also comment on the reliability of your data collection instruments and the validity of your conclusions.
深入评价展示高阶思维。讨论方法的局限性，例如样本量小、可能存在的回答偏差或无法控制的变量。举例而言，如果你只调查了11年级学生，你的发现就未必适用于所有中学生。提出具体的改进建议：”下次我会采用分层抽样，确保每个年级都有均衡的代表。”还要评论数据收集工具的可靠性以及结论的有效性。
If you used secondary data, evaluate its source — is it recent, published by a reputable organisation and free from obvious bias? The evaluation is your chance to show that you understand the investigation’s weaknesses and know how to refine it.
如果你使用了二手数据，评价其来源——是否新近、是否由信誉良好的机构发布、是否没有明显偏差。评价环节是你展示自己理解调查弱点并懂得如何改进的机会。
11. Practical Exam Tips for Statistics | 统计实践考试建议
In a timed practical test, read the entire question before starting. Identify what you are asked to plan, calculate or draw. Show all working clearly — marks are often awarded for method even if the final answer is wrong. When drawing graphs, use a sharp pencil, label everything and choose an appropriate scale that uses at least half the graph paper. Double-check that you have included units and that your charts match the data type.
在限时的实践考试中，动笔前先通读全题。明确考题要求你规划、计算或绘制什么。清晰地展示所有运算步骤——即便最终答案错误，也往往能获得方法分。画图时，使用尖铅笔，标注所有项目，选用合适的比例尺，至少利用坐标纸的一半面积。再次检查是否包含单位，图表是否与数据类型匹配。
Manage your time: do not spend too long perfecting one graph if there are several tasks to complete. If you are asked to comment on a given investigation, use statistical vocabulary (bias, validity, reliability) and structure your answers around the key stages of the statistical enquiry cycle.
管理好时间：如果需要完成好几项任务，不要在某一张图上花费过长时间以求完美。如果被要求评论给定的调查，使用统计学术语（偏差、有效性、可靠性），并围绕统计调查周期的关键阶段组织你的回答。
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