📚 A-Level CCEA Statistics: Key Points for the Practical Investigation | CCEA A-Level 统计学：实践调查考核要点
The practical investigation, assessed through Unit A2 2: Statistical Investigations, is a unique component of the CCEA A-Level Statistics course. It requires you to design, carry out, analyse and report on your own statistical enquiry. This independent project is marked by your teacher and externally moderated, contributing 20% of your A2 grade. Mastering the key assessment criteria is essential to achieving a high mark, and this guide will take you through every stage, from planning to final report, ensuring you understand what examiners are looking for.
实践调查通过单元 A2 2：统计调查进行评估，是 CCEA A-Level 统计学课程中一个独特的组成部分。它要求你自己设计、执行、分析并报告一项统计研究。这个独立项目由你的老师打分并接受外部审核，占 A2 成绩的 20%。掌握关键的考核要点对于取得高分至关重要，本指南将带你走完从规划到最终报告的每个阶段，确保你理解考官所关注的内容。
1. Understanding the Assessment Objectives | 理解考核目标
The investigation is assessed against three overarching objectives. You must demonstrate the ability to define a problem and plan a statistical approach (AO1), collect, process and represent data appropriately (AO2), and finally interpret results, draw conclusions and evaluate the whole process (AO3). Each objective carries roughly equal weight, so your work must show strength in design, execution and critical reflection.
调查依据三个总体目标进行评分。你必须展现出定义问题并规划统计方法的能力（AO1），恰当地收集、处理和呈现数据（AO2），以及最终解释结果、得出结论并评估整个过程（AO3）。每个目标权重接近，因此你的工作必须在设计、执行和批判性反思方面都表现出色。
Examiners closely scrutinise the match between your research question, the data collected and the statistical techniques applied. A superficial or disjointed project will lose marks even if the arithmetic is correct. Keep a log of decisions and alterations, as this evidence of thoughtful planning strongly supports AO3.
考官会仔细审查你的研究问题、收集的数据与所用统计技术之间是否匹配。一个浅薄或脱节的项目即便计算正确也会丢分。记录你的决策和改动，因为这些体现深思熟虑的规划证据对 AO3 部分非常有利。
2. Choosing a Manageable and Meaningful Research Question | 选择一个可行而有意义的研究问题
Your project must begin with a clear, focused question that can be investigated using data you can realistically obtain. Avoid vague ideas like ‘Does exercise affect health?’ — instead refine it to ‘Is there a negative correlation between weekly exercise hours and resting heart rate among Year 12 students at my school?’ This gives you a specific population, two measurable variables and a clear direction for analysis.
你的项目必须始于一个清晰、聚焦的问题，该问题应能用你可实际获得的数据进行研究。避免诸如“运动影响健康吗？”这样模糊的想法——将其细化为“我校12年级学生的每周锻炼小时数与静息心率之间是否存在负相关？”这为你明确了总体、两个可测变量和清晰的分析方向。
Choose a topic you genuinely find interesting; this will sustain motivation through the many hours of work. However, do consider practical constraints: access to participants, time needed for data collection and any ethical sensitivities. A project on caffeine intake and reaction time might be feasible using schoolmates, whereas a study on income and happiness could be problematic.
选择一个你真正感兴趣的话题；这能支撑你在长时间工作中的动力。但务必考虑实际限制：接触参与者的途径、数据收集所需的时间及任何伦理敏感性。利用同学来调查咖啡因摄入与反应时间或许可行，而关于收入与幸福感的研究可能困难重重。
3. Planning and Stating Hypotheses | 规划与陈述假设
Once the question is set, translate it into statistical hypotheses. For a correlation you would specify a null hypothesis H₀: ρ = 0 and an alternative H₁: ρ ≠ 0 (or one-tailed if justified). For comparing groups, use clear notation such as H₀: μ₁ = μ₂ vs H₁: μ₁ ≠ μ₂. Your plan should also detail the sampling method, the variables to be recorded, and the instruments or questionnaires to be used.
一旦确定了问题，就将其转化为统计假设。对于相关性，你要明确原假设 H₀: ρ = 0 与备择假设 H₁: ρ ≠ 0（若有理由也可用单尾）。对于组别比较，使用清晰的符号如 H₀: μ₁ = μ₂ 与 H₁: μ₁ ≠ μ₂。你的计划还应详细说明抽样方法、要记录的变量及将使用的工具或问卷。
A detailed planning statement is explicitly assessed. You will need to justify why a random sample, stratified sample or systematic sample is chosen, and how you intend to minimise bias. Pilot testing your data collection instrument on a few friends can reveal ambiguities, and noting this in your report demonstrates excellent AO1 practice.
详细的规划陈述是被明确评估的。你需要解释为什么选择简单随机抽样、分层抽样或系统抽样，以及你打算如何减少偏差。先在几个朋友中试测你的数据收集工具，可以发现模糊之处，在报告中记下这一点能展示出色的 AO1 技能。
4. Collecting Quality Data | 收集高质量数据
Data collection is the heart of AO2. Whether you gather primary data through experiments or surveys, or use secondary data from reliable databases, you must demonstrate rigour. For primary data, describe exactly how measurements were taken — for example, timing a 100 m sprint using the same stopwatch and flat track, or recording heart rate with a consistent protocol.
数据收集是 AO2 的核心。无论你是通过实验或问卷收集一手数据，还是使用可靠数据库的二手数据，都必须表现出严谨性。对于一手数据，要精确描述测量是如何进行的——例如，使用相同的秒表和平面跑道计时 100 米短跑，或按照一致的操作规程记录心率。
Avoid small sample sizes. For a correlation, aim for at least 30 paired observations; for comparing two groups, try to have at least 15 in each. Larger samples improve the power of your tests and make your conclusions more credible. Document any unexpected incidents during collection, as honest reflection enhances the evaluation component.
避免小样本量。对于相关分析，至少要有30组成对观测；对于两组比较，每组至少要有15个样本。更大的样本可以提高检验功效，使你的结论更可靠。记录收集过程中的任何意外事件，因为诚实的反思会增强评估部分。
5. Ensuring Ethical Practice and Data Integrity | 确保伦理规范与数据完整性
Ethical considerations are not optional. You must obtain informed consent from participants (or parents, if under 18), guarantee anonymity and allow them to withdraw at any time. In your report, include a short section explaining how you ensured confidentiality — for instance, by assigning each participant a code number instead of using names.
伦理考量并非可有可无。你必须获取参与者（若未满18岁则需父母）的知情同意，保证匿名并允许他们在任何时候退出。在你的报告中，包含一个简短的部分说明你如何确保保密——例如，为每位参与者分配一个代号而不是使用姓名。
Data integrity also means no fabrication or selective omitting of inconvenient data. If an outlier arises, flag it, but do not delete it without a very good statistical reason. An outlier that is a genuine observation should be included in the analysis, with a sensitivity check shown — perhaps running the test with and without the outlier.
数据完整性还意味着不能捏造数据，也不能选择性删去不利数据。如果出现异常值，要标出，但除非有充分的统计理由不要删除它。一个作为真实观测的异常值应纳入分析，并展示敏感性检验——或许分别在不含和含异常值的情况下运行检验。
6. Descriptive Statistics and Graphical Representation | 描述性统计与图形呈现
Before launching into complex inference, you must summarise your data appropriately. For univariate continuous data, report the mean, standard deviation, median and interquartile range, and present box plots or histograms. For bivariate data, produce a well-labelled scatter diagram and comment on the form, direction and strength of any relationship.
在进入复杂的推断之前，你必须恰当地概括你的数据。对于单变量连续数据，报告均值、标准差、中位数和四分位距，并绘制箱线图或直方图。对于双变量数据，绘制标注清晰的散点图，并评论任何关系的形式、方向和强度。
Every diagram must have a clear title, labelled axes with units, and an appropriate scale. Computer-generated graphs are preferred, but hand-drawn ones can still score full marks if they are neat and accurate. Avoid overcomplicating: for small data sets a pie chart is rarely helpful; choose displays that honestly communicate the data’s story.
每个图形必须有清晰的标题、带单位的坐标轴标签和合适的比例。计算机生成的图形是首选，但如果手绘图整洁且准确仍然可以获得满分。避免过度复杂化：对于小型数据集，饼图很少有帮助；要选择能够如实传达数据信息的展示方式。
7. Choosing the Right Inferential Test | 选择正确的推断检验
Selecting an appropriate hypothesis test is a key discriminator between grades. If you are testing for a linear relationship between two numerical variables, use Pearson’s product-moment correlation coefficient, checking the scatter plot for linearity first. To compare the means of two independent groups, and the data are approximately normal with equal variances, the two-sample t-test is standard.
选择合适的假设检验是区分成绩等级的关键。如果你要检验两个数值变量之间的线性关系，使用 Pearson 积矩相关系数，但首先要检查散点图的线性性。要比较两个独立组的均值，且数据近似正态且方差相等时，双样本 t 检验是标准方法。
When conditions are not met, you may need a non-parametric alternative: Spearman’s rank correlation for monotonic association or the Mann-Whitney U test for comparing medians. You must state clearly why you chose a particular test, and verify its assumptions using outputs such as a normal probability plot or Bartlett’s test for equality of variance.
当条件不满足时，你可能需要非参数替代方法：Spearman 秩相关系数用于单调关联，Mann-Whitney U 检验用于比较中位数。你必须清楚说明为什么选择某个检验，并利用正态概率图或 Bartlett 方差齐性检验的输出来验证其假设。
8. Performing Calculations and Using Technology | 执行计算与使用技术
You are expected to use a calculator or software such as Excel, GeoGebra or JASP to carry out the computations, but you must show evidence of the process. Do not just paste a table of output; explain what each figure (e.g., r, p-value, t-statistic) means. For the Product Moment Correlation Coefficient, show the formula used and illustrate a couple of steps before presenting the final result.
你应使用计算器或软件（如 Excel、GeoGebra 或 JASP）来进行计算，但必须展示过程证据。不要只是粘贴输出表格；要解释每个数值（如 r、p值、t统计量）的含义。对于积矩相关系数，给出所用的公式并在展示最终结果前演示几个运算步骤。
When performing regression analysis, provide the equation of the least squares regression line y = a + bx, clearly stating which variable is the predictor and which is the response. Interpret the slope in context. Also calculate the coefficient of determination R² and comment on the proportion of variation explained by the model.
执行回归分析时，给出最小二乘回归线方程 y = a + bx，清楚说明哪个是预测变量、哪个是响应变量。在背景中解释斜率。还要计算决定系数 R²，并评论模型解释的变异比例。
9. Interpreting P-Values and Drawing Conclusions | 解释 P 值并得出结论
The conclusion must link directly back to the original research question and hypotheses. Compare the p-value to a pre-specified significance level, usually α = 0.05. If p < 0.05, state that there is sufficient evidence to reject H₀ at the 5% level, and then restate this in plain English: 'The data suggest a significant positive correlation between study hours and test scores.'
结论必须直接回应最初的研究问题和假设。将 p 值与预先设定的显著性水平（通常 α = 0.05）进行比较。若 p < 0.05，说明在5%水平下有足够的证据拒绝 H₀，然后用通俗语言重申：“数据表明学习小时数与测验分数之间存在显著正相关。”
Never say ‘prove’ — statistical significance does not prove a hypothesis; it only provides evidence against the null. Also discuss the practical significance: a statistically significant result with a tiny effect size (e.g., r = 0.1) may be scientifically uninteresting. Your judgment here is part of AO3 evaluation.
绝不要说“证明”——统计显著性并不能证明一个假设；它只是提供了反对原假设的证据。还要讨论实际显著性：一个统计显著但效应量很小（例如 r = 0.1）的结果可能没有实际意义。你对此的判断是 AO3 评价的一部分。
10. Evaluating the Investigation and Suggesting Improvements | 评估调查并提出改进建议
A high-level project does not simply end with a conclusion; it includes a thorough evaluation. Identify limitations such as sampling bias, measurement error, small sample size or uncontrolled confounding variables. Explain how each limitation could have affected your results. For example, if you measured resting heart rate in a noisy classroom, the values might be elevated.
一个高水平的项目不仅仅是得出结论，还要包含全面的评估。指出局限性，如抽样偏差、测量误差、小样本量或未控制的混杂变量。解释每项局限如何影响你的结果。例如，如果是在嘈杂的教室测量静息心率，测值可能偏高了。
Concrete suggestions for improvement are essential. If you used convenience sampling, propose a stratified random sample next time. If your questionnaire lacked a question on caffeine consumption before the test, note that as a confounding factor to be measured in a future study. This section is crucial for AO3 marks.
具体的改进建议至关重要。如果你此前使用了便利抽样，下次建议采用分层随机抽样。如果你的问卷缺少关于测试前咖啡因摄入的问题，要指出这是今后研究需要测量的混杂因素。此部分对 AO3 评分极为关键。
11. Structuring the Written Report | 拟定报告结构
The report should have a professional structure: title page, abstract, introduction, methodology, data summary and graphs, inferential analysis, discussion, conclusion, evaluation and references. Use clear headings so that examiners can quickly locate each required component. An abstract of about 150 words, summarising objectives, methods and key findings, is highly recommended.
报告应有一个专业的结构：标题页、摘要、引言、方法、数据汇总与图表、推断分析、讨论、结论、评估和参考文献。使用清晰的标题，以便考官能迅速找到每个必要部分。强烈建议撰写约150字的摘要，总括目标、方法和主要发现。
Keep your writing precise and avoid irrelevant chatter. All graphs and tables should be embedded in the text, not placed in an appendix, unless supplementary raw data is required. Number your figures and tables consecutively and refer to them in the analysis. Use of statistical vocabulary like ‘bivariate normal assumption’ or ‘homoscedasticity’ should be correct and context-appropriate.
保持写作精确，避免无关的闲聊。所有图形和表格都应嵌入正文，而非置于附录中（除非需要补充原始数据）。为你的图表连续编号，并在分析中加以引用。统计术语如“双变量正态假设”或“方差齐性”的使用要正确且符合语境。
12. Common Mistakes to Avoid | 常见错误与规避方法
One frequent mistake is misalignment between the research question and the statistical test. Students sometimes collect categorical data but then attempt a correlation. Map your question to the correct data type and test from the very beginning. Another pitfall is ignoring assumptions — using Pearson correlation without checking for linearity can yield a misleading r.
一个常见错误是研究问题与统计检验不匹配。学生有时收集分类数据，却试图进行相关分析。从一开始就要将问题与正确的数据类型及检验对应起来。另一个误区是忽略假设——未检查线性性就使用 Pearson 相关，可能得出误导性的 r。
Poor time management is also a major risk. Start early, set deadlines for each phase, and build in time for unforeseen delays, such as waiting for questionnaire returns. Finally, do not overlook the review of the statistical process: the evaluative sections are often where the highest grades are won or lost.
时间管理不善也是一大风险。及早开始，为每个阶段设定截止日期，并预留应对意外延误（如等待问卷回收）的时间。最后，不要忽视对统计过程的回顾：评判性部分往往是决定能否获得最高分的关键所在。
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