📚 Year 12 CCEA Statistics: Report Writing Framework and Model Answer | CCEA 12年级统计：调查报告写作框架与范文
Writing a statistical investigation report for CCEA Year 12 is a skill that combines clear communication with rigorous data analysis. This article walks you through a structured framework for planning, executing, and presenting your investigation, along with a full model answer demonstrating how to apply these steps to a real-world scenario. Whether you are comparing two populations or exploring an association, a well-organised report makes your statistical reasoning shine.
为CCEA 12年级撰写统计调查报告是一项将清晰表达与严谨数据分析相结合的技能。本文将带你走过一个结构化的框架，涵盖规划、执行和展示调查的每个环节，并附上一篇完整的范文，展示如何将这些步骤应用到一个真实场景中。无论你是在比较两个总体还是在探索关联性，一份条理清晰的报告都能让你的统计推理大放异彩。
1. Understanding the CCEA Statistical Investigation | 理解CCEA统计调查任务
The CCEA AS Statistics specification requires you to plan, carry out, and interpret a statistical investigation based on primary or secondary data. Your report must show evidence of framing a hypothesis, selecting and justifying sampling methods, summarising data with appropriate graphics and statistics, performing an inferential test, and critically evaluating the whole process. The final written piece is assessed on both statistical accuracy and the quality of your narrative.
CCEA AS统计课程要求你基于一手或二手数据，规划、执行并解释一项统计调查。你的报告必须体现出以下环节：提出假设、选择并论证抽样方法、用恰当的图表和统计量概括数据、进行推断性检验，并对整个过程进行批判性评估。最终的书面作品既考察统计准确性，也评判叙述的质量。
2. Planning: Defining Hypothesis and Variables | 规划阶段：定义假设与变量
Begin by clearly stating your null hypothesis H₀ and alternative hypothesis H₁. For example, if you are investigating whether screen time affects sleep duration, you might set H₀: μ₁ = μ₂ (the mean sleep duration is the same for light and heavy screen users) against H₁: μ₁ ≠ μ₂. Identify the independent variable (screen-time group) and the dependent variable (hours of sleep). Also specify any control variables, such as age range, to limit confounding influences.
首先要清晰陈述你的原假设 H₀ 和备择假设 H₁。例如，如果你正在调查屏幕时间是否影响睡眠时长，你可以设定 H₀: μ₁ = μ₂（轻度屏幕使用者和重度使用者的平均睡眠时长相同），对应 H₁: μ₁ ≠ μ₂。明确自变量（屏幕时间分组）和因变量（睡眠小时数）。同时指定控制变量，如年龄范围，以限制混杂影响。
3. Sampling Methods and Data Collection | 抽样方法与数据收集
Describe your sampling strategy and justify it. A simple random sample reduces bias but may be impractical; stratified sampling ensures representation of key subgroups, like gender or year group. You must explain how the sample was obtained, its size (n ≥ 30 for each group where possible), and any steps taken to minimise non-response or measurement error. Always include a copy of any questionnaire or recording sheet in the appendix.
描述你的抽样策略并论证其合理性。简单随机抽样可以减少偏差，但可能不便实施；分层抽样能确保关键子群体（如性别或年级）的代表性。你必须说明样本是如何获得的、样本量（每组尽量 n ≥ 30），以及为减少无回答或测量误差所采取的措施。任何问卷或记录表都应附在附录中。
4. Presenting Data: Tables and Graphs | 数据展示：表格与图表
Your report must summarise raw data visually. Use a comparative box plot to display the sleep distributions side by side for the two screen-time groups. A stem-and-leaf diagram or histogram can show shape, while a summary table should give frequencies, means, and standard deviations. Label all axes clearly, include titles, and reference each figure in the text. Never paste a screenshot of a spreadsheet; draw graphs carefully or use approved software.
你的报告必须用可视化方式概括原始数据。使用并列的箱线图来展示两个屏幕时间组的睡眠分布。茎叶图或直方图可以显示分布形态，而汇总表则应给出频数、均值和标准差。所有坐标轴要清晰标注，加上标题，并在正文中引用每一张图。切勿粘贴电子表格的截图；应仔细手绘或使用认可的软件绘图。
5. Calculating Descriptive Statistics | 计算描述性统计量
For each group, compute the sample mean x̄, sample standard deviation s, median, and interquartile range. Show the formulas you use, but you may let technology handle the arithmetic. Present the findings in a well-structured table:
对每一组数据，计算样本均值 x̄、样本标准差 s、中位数和四分位距。展示你使用的公式，但可以借助技术来完成算术。将结果呈现在一个结构良好的表格中：
| Group | n | x̄ (hours) | s (hours) | Median | IQR |
|---|---|---|---|---|---|
| Light screen (<3h) | 40 | 8.2 | 1.1 | 8.0 | 1.5 |
| Heavy screen (≥3h) | 35 | 7.0 | 1.4 | 7.1 | 2.0 |
This table makes it easy to spot that the light screen group appears to sleep longer on average and has less variability.
这样的表格便于发现轻度屏幕组似乎平均睡眠时间更长且变异性更小。
6. Inferential Testing: Choosing the Right Test | 推断检验：选择合适的检验方法
When comparing means of two independent groups, and you do not know the population standard deviations, a two-sample t-test for the difference of means is appropriate. Check the assumptions: data should be approximately normally distributed in each group (or n > 30 justifies by the Central Limit Theorem); the samples must be independent; and the variable must be continuous. If variances are unequal, use Welch’s t-test. Mention the significance level, typically α = 0.05.
当比较两个独立总体的均值，并且总体标准差未知时，适宜使用两样本均值差 t 检验。检查前提条件：每组数据应近似正态分布（或 n > 30 可由中心极限定理保证）；样本必须独立；变量必须连续。若方差不相等，则使用 Welch t 检验。说明显著性水平，通常 α = 0.05。
7. Performing the Hypothesis Test Step-by-Step | 逐步进行假设检验
State the hypotheses clearly again. Using the data above:
再次明确陈述假设。使用以上数据：
H₀: μ_Light = μ_Heavy
H₁: μ_Light ≠ μ_Heavy
Calculate the test statistic using the formula for Welch’s t-test:
使用 Welch t 检验公式计算检验统计量：
t = (x̄₁ – x̄₂) / √(s₁²/n₁ + s₂²/n₂) = (8.2 – 7.0) / √(1.1²/40 + 1.4²/35) ≈ 4.10
The degrees of freedom are computed with the Welch-Satterthwaite equation:
自由度使用 Welch–Satterthwaite 公式计算：
df ≈ [(s₁²/n₁ + s₂²/n₂)²] / [ (s₁²/n₁)²/(n₁-1) + (s₂²/n₂)²/(n₂-1) ] ≈ 64.3 → rounded down to 64.
Using technology, the two-tailed p-value for t = 4.10 with df = 64 is p < 0.001. Since p < 0.05, we reject H₀.
借助技术，t = 4.10、df = 64 的双尾 p 值为 p < 0.001。由于 p < 0.05，我们拒绝 H₀。
8. Interpreting Results and Writing Conclusions | 结果解读与结论撰写
Do not simply write ‘reject H₀’. Explain what the finding means in context. For example: ‘There is very strong evidence of a difference in mean sleep duration between light and heavy screen users. The 95% confidence interval for the difference (Light − Heavy) is (0.62 hours, 1.78 hours), suggesting that light screen users sleep between 0.6 and 1.8 hours more on average.’ Always link back to the original investigation aim.
不要只写一句“拒绝 H₀”。要解释这一发现的实际含义。例如：“有非常强有力的证据表明，轻度与重度屏幕用户的平均睡眠时长存在差异。差异（轻度 − 重度）的 95% 置信区间为（0.62 小时，1.78 小时），这说明轻度屏幕用户平均多睡 0.6 到 1.8 小时。”始终与原始调查目标联系起来。
9. Evaluation: Limitations, Reliability and Validity | 评估：局限性、信度与效度
A strong evaluation goes beyond saying ‘the sample was small’. Discuss whether the sample was representative, if measurement instruments (e.g., self-reported sleep diaries) could introduce bias, and whether the assumption of normality was reasonable. Comment on internal validity – could other variables, such as caffeine intake, explain the difference? Suggest concrete improvements, like using a larger stratified random sample or a sleep tracker for objective measurement.
高质量的评估不会仅仅说“样本较小”。要讨论样本是否具有代表性，测量工具（如自述睡眠日记）是否会引入偏差，正态性假设是否合理。对内效度做出评论——其他变量（如咖啡因摄入）能否解释观察到的差异？提出具体的改进措施，比如使用更大的分层随机样本，或利用睡眠追踪仪进行客观测量。
10. Model Report Structure and Marking Criteria | 范文结构模板与评分标准
Organise your report under clear headings. The table below maps each section to what CCEA examiners look for, with dual-language guidance.
将你的报告组织在清晰的标题下。下表将每一部分对应到CCEA考官关注的内容，并提供双语指导。
| Section (English) | What to Include | 部分 (中文) | 应包含内容 |
|---|---|---|---|
| Introduction | Context, aim, hypotheses, variables. | 引言 | 背景、目的、假设、变量。 |
| Methodology | Sampling design, data collection tool, ethical considerations. | 方法 | 抽样设计、数据收集工具、伦理考量。 |
| Data Presentation | Tables, suitable graphs, summary statistics. | 数据呈现 | 表格、恰当的图形、汇总统计量。 |
| Inferential Analysis | Test chosen, assumptions checked, calculations, p-value/CI. | 推断分析 | 所选检验、前提检查、计算、p值/CI。 |
| Conclusion | Decision on hypothesis, contextual interpretation. | 结论 | 对假设的判断、情境化解读。 |
| Evaluation | Limitations, reliability, validity, suggestions for improvement. | 评估 | 局限、信度、效度、改进建议。 |
Following this framework ensures your report addresses every assessment objective, including the crucial evaluation section that many candidates overlook.
遵循这一框架可确保你的报告涵盖每一个评估目标，包括许多考生容易忽略的关键评估部分。
11. Model Answer: Screen Time vs Sleep – A Two-Sample t-Test Investigation | 范文：屏幕时间与睡眠——双样本t检验调查
Abstract: This investigation examines whether daily screen time affects the nightly sleep duration of Year 12 students. Two groups were compared: light users (<3 hours) and heavy users (≥3 hours). A Welch two-sample t-test revealed a significant difference (p < 0.001), with light users sleeping on average 1.2 hours longer. Possible confounding factors and recommendations for further research are discussed.
摘要：本调查考察每日屏幕时间是否影响12年级学生的夜间睡眠时长。将轻度用户（<3小时）和重度用户（≥3小时）两组进行比较。Welch双样本t检验显示存在显著差异（p < 0.001），轻度用户平均多睡1.2小时。文中讨论了可能的混杂因素及后续研究建议。
Introduction: Students often trade sleep for screen time, but is the relationship statistically clear? I aimed to determine if the population mean sleep of light screen users differs from that of heavy screen users. The null hypothesis states no difference; the alternative states a difference exists. The independent variable is screen-time category; the dependent variable is self-reported hours of sleep.
引言：学生们常以睡眠为代价换取屏幕时间，但这种关系在统计上是否明确？我旨在确定轻度屏幕用户的总体睡眠均值是否与重度用户不同。原假设是二者无差异，备择假设是存在差异。自变量为屏幕时间类别，因变量为自报睡眠小时数。
Methodology: I used stratified random sampling to select 75 Year 12 students from a school of 600, ensuring proportional representation by gender. Participants completed a three-day sleep and screen diary. Screen time was averaged and split at 3 hours. Ethical protocols included informed consent and anonymity.
方法：我采用分层随机抽样，从一所600人的学校中选出75名12年级学生，确保性别比例相当。参与者完成了一份三日睡眠与屏幕日记。将屏幕时间取均值，并以3小时为界分组。伦理程序包括知情同意和匿名处理。
Results: Box plots showed the light screen group had a higher median and smaller spread. Summary statistics were as in Section 5. The data satisfied the independence and approximate normality assumptions (normal probability plots were roughly linear). Welch’s t-test gave t = 4.10, df ≈ 64, p < 0.001. The 95% CI for the difference (Light − Heavy) was (0.62, 1.78) hours.
结果：箱线图显示轻度屏幕组的中位数更高、分布更集中。汇总统计量见第5节。数据满足独立性和近似正态性假设（正态概率图基本呈线性）。Welch t检验得到 t = 4.10，df ≈ 64，p < 0.001。差异（轻度 − 重度）的95%置信区间为（0.62, 1.78）小时。
Conclusion: The p-value is far below 0.05, so we reject H₀. There is very strong evidence that light screen users sleep longer. This does not prove causation; the design is observational. However, the effect size is substantial and warrants attention.
结论：p值远小于0.05，因此我们拒绝H₀。有非常强有力的证据表明轻度屏幕用户睡眠更久。这并不能证明因果关系，因为本研究为观察性设计。但效应量较大，值得关注。
Evaluation: The sample was limited to one school, affecting generalisability. Self-reported data may under- or overestimate actual sleep. Moreover, no data on bedtime procrastination or academic workload were collected. Future studies should use activity trackers and control for stress levels. Despite these issues, the investigation was carefully conducted, and the large t-value suggests the result is robust.
评估：样本仅来自一所学校，影响了推广性。自报数据可能低估或高估实际睡眠。此外，未收集就寝拖延或学业负担的数据。未来研究应使用活动追踪器并控制压力水平。尽管存在这些问题，但调查实施严谨，且较大的t值表明结果是稳健的。
12. Common Pitfalls and Examiner Tips | 常见错误与考官建议
Avoid presenting a report that reads like a maths exercise. Do not skip the real-world interpretation of your p-value and confidence interval. Many students lose marks by including inappropriate graphs (e.g., a pie chart for continuous data) or forgetting to label axes. Proofread for clarity: your conclusion should answer the initial question in a way a non-statistician can understand. Finally, always tie your evaluation back to how limitations might have affected the validity of your conclusions, and propose feasible follow-up studies.
避免将报告写得像数学练习题。不要跳过对p值和置信区间的现实意义解读。许多学生因使用不恰当的图表（例如对连续数据用饼图）或忘记标注坐标轴而失分。检查表达的清晰度：你的结论应当以一种非统计专业人士也能理解的方式回答最初的问题。最后，始终将评估与局限性如何影响结论的效度联系起来，并提出切实可行的后续研究方案。
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