📚 Mastering Statistical Report Writing: Framework and Model Answer | AS WJEC 统计论文写作框架与范文
A well-structured statistical report is the backbone of high achievement in AS WJEC Statistics. Unlike pure calculation exercises, the report demands a logical flow that connects a research question, data collection, analysis, interpretation, and a clear conclusion. Mastering this framework not only secures marks for communication but also demonstrates a deep understanding of the statistical enquiry cycle. This article walks you through every essential section of a WJEC-style report, from hypothesis formulation to evaluation, and provides a fully annotated model answer to show exactly how examiners expect you to write.
结构良好的统计报告是 AS WJEC 统计课程取得高分的基石。与纯粹的计算练习不同，报告要求有一条清晰的逻辑线，将研究问题、数据收集、分析、解读与明确的结论串联起来。掌握这个框架不仅能获得表达分，更能展现你对统计探究循环的深入理解。本文带你逐一拆解 WJEC 风格报告中的每一个关键环节，从假设提出到评价总结，并附上一篇带详细注释的范文，让你直观看到考官所期望的写作方式。
1. The Statistical Enquiry Cycle in WJEC | 统计探究循环
The WJEC specification is built around the statistical enquiry cycle (SEC): Problem, Plan, Data, Analysis, Conclusion. Every report you write, whether for coursework or an exam question, should mirror this sequence. Start by identifying the problem and framing it as a precise research question. Then plan what data you need and how to collect or simulate it. After gathering the data, process and present it using appropriate diagrams, summary statistics, and probability models. Analyse the results with inferential methods, and finally draw a conclusion that refers back to the original question and acknowledges limitations. This cyclical thinking ensures no step is missed and that your report reads coherently.
WJEC 考纲围绕着统计探究循环（SEC）展开：问题、计划、数据、分析、结论。你写的每一份报告，无论是课程作业还是考试题目，都应该反映这个顺序。首先要识别问题并将其提炼成一个精确的研究问题；然后计划需要什么数据、如何收集或模拟数据；收集数据后，使用合适的图表、汇总统计量和概率模型进行处理与展示；接着用推断方法分析结果；最后得出结论，回顾原问题并承认局限性。这种循环式思维能确保每个步骤都不遗漏，使报告条理清晰。
2. Formulating a Clear Research Question and Hypothesis | 明确研究问题与假设
stion is specific, measurable, and directly testable with the data available. For example, 'Is there a difference between the mean reaction times of males and females?' or 'Does a new fertiliser increase plant growth?' μ₂, while H₁ could be μ₁ ≠ μ₂ (two-tailed) or μ₁ > μ₂ (one-tailed). Always define population parameters clearly and explain why a one-tailed or two-tailed test is appropriate. This initial step sets the direction for all subsequent statistical work and should be written in plain English alongside mathematical notation.
在任何数据处理之前，你必须先提出一个明确的研究问题。好的研究问题是具体的、可衡量的，并且能够用现有数据直接检验。例如，“男性和女性的平均反应时间是否存在差异？”或“新型肥料是否能促进植物生长？”接下来，将问题转化为原假设和备择假设（H₀ 和 H₁）。对于双样本 t 检验，H₀ 可以是 μ₁ = μ₂，而 H₁ 可以是 μ₁ ≠ μ₂（双尾）或 μ₁ > μ₂（单尾）。务必清晰界定总体参数，并解释为何选用单尾或双尾检验。这第一步为后续所有统计工作指明了方向，应该用简洁的英语和数学符号同时表达。
3. Planning and Describing Data Collection | 数据收集的计划与描述
The data section must explain how the sample was obtained. Was it a random sample, stratified sample, or convenience sample? State the sample size and any variables recorded. For an experiment, outline the treatment and control groups, and mention blinding or randomisation if relevant. Describe potential sources of bias—such as selection bias, measurement error, or non-response—and what steps were taken to minimise them. In WJEC reports, you are often given a scenario; you need to identify the sampling method used and comment on its suitability. Even a short description like 'A simple random sample of 50 students was selected using a random number table' shows awareness of good practice. A table of variables with their types (continuous, discrete, categorical) is a useful addition.
数据部分必须说明样本是如何获得的。是简单随机样本、分层抽样还是方便样本？要写明样本量和所有记录的变量。对于实验，要概述处理组和对照组，并在相关时提及盲法或随机化。描述潜在的偏差来源——如选择偏差、测量误差或无应答——以及采取了哪些步骤来减小它们。在 WJEC 报告中，通常会给你一个情境，你需要识别所用抽样方法并评价其适用性。即使像“使用随机数表选取了 50 名学生的简单随机样本”这样简短的描述，也能体现出对良好实践的意识。用一张表格列出变量及其类型（连续、离散、分类）会很有帮助。
| Variable | Type | Unit |
|---|---|---|
| Reaction time | Continuous | Seconds |
| Gender | Categorical (Binary) | Male / Female |
4. Data Presentation and Visualisation | 数据展示与可视化
Charts and graphs are not just decorations—they communicate the shape, centre, and spread of data at a glance. Choose plots that match your data type: histograms or box plots for continuous data, bar charts for categorical data, and scatter diagrams for bivariate relationships. Always label axes clearly, include a title, and add a brief commentary beneath each figure explaining what the reader should notice. For instance, 'The box plot shows that the median reaction time for females is slightly lower, and the interquartile range is smaller, suggesting less variability.' In a WJEC report, hand-drawn or clearly presented computer-generated graphs are acceptable, but you must refer to them in the text. Never include a graph without interpretation.
图表不仅仅是装饰——它们能让人一目了然地看出数据的形状、中心和离散程度。选择与数据类型相匹配的图形：连续数据用直方图或箱线图，分类数据用条形图，双变量关系用散点图。始终清楚地标记坐标轴，加上标题，并在每个图形下方附上简短说明，告诉读者应该注意什么。例如，“箱线图显示女性的反应时间中位数略低，且四分位距更小，表明变异性较小。”在 WJEC 报告中，手绘或清晰呈现的电脑绘图均可接受，但你必须在正文中引用它们。切勿只放图表而不加以解读。
5. Calculating and Interpreting Descriptive Statistics | 计算与解读描述性统计量
Descriptive statistics summarise the key features of your sample. Report measures of central tendency (mean, median, mode) and measures of dispersion (range, interquartile range, standard deviation). State which measures are most appropriate given the shape of the distribution. For skewed data, the median and IQR are better than the mean and standard deviation. When comparing two groups, present a comparison table with sample sizes, means, and standard deviations. Then interpret the values in context: 'The mean reaction time for males is 0.32 s with a standard deviation of 0.08 s, while for females the mean is 0.29 s with a standard deviation of 0.06 s. This suggests that, on average, females reacted faster and were more consistent.' Use these statistics to form a preliminary judgement before moving to formal hypothesis testing.
描述性统计量能概括你样本的关键特征。要报告集中趋势的度量（均值、中位数、众数）和离散程度的度量（极差、四分位距、标准差）。说明根据分布形状，哪些度量最合适。对于偏态数据，中位数和四分位距优于均值和标准差。比较两组时，用一张对比表列出样本量、均值和标准差。然后将数值放在具体情境中解读：“男性的平均反应时间为 0.32 秒，标准差为 0.08 秒；女性的平均反应时间为 0.29 秒，标准差为 0.06 秒。这表明，平均而言女性反应更快且更稳定。”在进行正式的假设检验之前，先利用这些统计量形成初步判断。
6. Choosing the Right Probability Distribution | 选择合适的概率分布
ch with the same probability of success, the binomial distribution B(n, p) is appropriate. Check that the conditions are met: fixed n, constant p, independent trials, and a binary outcome. For continuous symmetric data, the normal distribution N(μ, σ²) is used— Limit Theorem. When applying the normal model, always standardise to the standard normal Z ~ Z = (X − μ)/σ. Write clearly which distribution you are using and justify it. For example, ' n is large (n=50), the sample mean reaction time can be modelled as approximately normal.' This shows evaluative thinking.
许多 WJEC 报告要求你用二项分布或正态分布对数据进行建模。如果情境包含固定次数的独立试验，且每次成功概率相同，则二项分布 B(n, p) 是合适的。要检查条件是否满足：固定的 n、不变的 p、独立试验以及二元结果。对于连续的对称数据，则使用正态分布 N(μ, σ²)——特别是通过中心极限定理引用样本均值的分布时。在应用正态模型时，始终使用 Z = (X − μ)/σ 将其标准化为标准正态 Z ~ N(0,1)。清楚写明你使用的是哪种分布并说明理由。例如，“由于 n 很大 (n=50)，样本均值反应时间可以近似服从正态分布。”这展示了评价性思维。
7. Conducting Hypothesis Tests Step by Step | 逐步进行假设检验
This is the core of statistical inference. Structure your hypothesis test using a consistent layout:
- Step 1: State H₀ and H₁ in symbols and words.
- Step 2: Determine the significance level (usually 5% or given).
- Step 3: State the test statistic and its distribution under H₀, e.g., t = (x̄₁ − x̄₂)/√s²/n₁ + s²/n₂ ~ t with Welch-Satterthwaite degrees of freedom.
- Step 4: Calculate the test statistic and p-value (or compare with critical value).
- Step 5: Make a decision: reject H₀ if p-value < significance level, otherwise do not reject.
- Step 6: Write a conclusion in context. Never just say 'reject H₀'; explain what this means in the original scenario, including evidence strength.
这是统计推断的核心。用一致的布局来组织你的假设检验：
- 步骤 1：用符号和文字陈述 H₀ 和 H₁。
- 步骤 2：确定显著性水平（通常 5% 或题目给定）。
- 步骤 3：陈述检验统计量及其在 H₀ 下的分布，例如 t = (x̄₁ − x̄₂)/√s²/n₁ + s²/n₂ 服从 Welch-Satterthwaite 自由度下的 t 分布。
- 步骤 4：计算检验统计量和 p 值（或与临界值比较）。
- 步骤 5：做出决策：若 p 值 < 显著性水平则拒绝 H₀，否则不拒绝。
- 步骤 6：结合情境写出结论。绝不要只说“拒绝 H₀”；要解释这在实际情境中意味着什么，包括证据的强度。
8. Drawing Conclusions and Evaluating the Investigation | 得出结论与评估调查
A strong conclusion does three things: it answers the original research question directly, it states the statistical outcome (e.g., 'there is sufficient evidence at the 5% level to suggest...'), and it interprets the effect size, not just significance. Avoid overstating results—statistical significance does not imply practical importance. Then evaluate the whole investigation critically: discuss sample size adequacy, possible confounding variables, outliers, measurement precision, and the validity of assumptions (normality, independence). Suggest realistic improvements, such as a larger sample, better randomisation, or more accurate instruments. An evaluative paragraph where you reflect on limitations shows higher-order thinking and is often a differentiator for top marks in WJEC.
一个好的结论要做到三点：直接回答原研究问题，陈述统计结果（如“在 5% 的显著水平下，有足够证据表明……”），并且解释效应量，而不只是显著与否。避免夸大结果——统计显著并不意味着实际重要。然后对整个调查进行批判性评价：讨论样本量的充分性、可能的混杂变量、异常值、测量精度以及假设的有效性（正态性、独立性）。提出切实可行的改进建议，比如更大的样本、更好的随机化或更精确的仪器。一个反省局限性并思考改进的评价段落能展现高阶思维能力，在 WJEC 中常常是拔得头筹的关键。
9. Structuring the Full Report: A Section-by-Section Template | 完整报告的结构：逐节模板
Here is a ready-to-use template for an AS WJEC statistical report:
- Title: Concise and descriptive.
- Introduction: Background, research question, hypotheses.
- Methodology: Sampling strategy, variables, data collection.
- Data Presentation: Tables, graphs, and commentary.
- Descriptive Statistics: Summary measures with interpretation.
- Inferential Analysis: Choice of test, checking assumptions, hypothesis test steps, p-value.
- Conclusion: Answer question, state statistical decision, discuss effect size.
- Evaluation: Limitations, reliability, improvements.
这是一份现成可用的 AS WJEC 统计报告模板：
- 标题：简洁且具有描述性。
- 引言：背景、研究问题、假设。
- 方法：抽样策略、变量、数据收集。
- 数据展示：表格、图形及说明。
- 描述性统计：汇总度量及解读。
- 推断分析：检验选择、条件检查、假设检验步骤、p 值。
- 结论：回答问题，陈述统计决策，讨论效应量。
- 评价：局限性、可靠性、改进建议。
10. Worked Example: Reaction Times by Gender | 范文示例：按性别划分的反应时间
Below is a model answer for a WJEC-style investigation into whether male and female reaction times differ. Note how every section links to the enquiry cycle.
Title: A Comparison of Simple Reaction Times between Male and Female Sixth-Form Students.
Introduction
s and may be influenced by biological and environmental factors. The research question is: 's there a difference in the mean simple reaction times of male and female Year 12 students?' null hypothesis H₀: μ_m = μ_f ; the alternative H₁: μ_m ≠ μ_f , a two-tailed test at the 5% significance level.
Methodology
A stratified random sample of 30 males and 30 females was selected from a sixth-form college register using random numbers within gender strata. Each participant completed a computer-based reaction test three times, and the average (in seconds) was recorded. The variable 'gender' is binary categorical, and 'reaction time' is continuous. All tests were conducted in the same quiet room at the same time of day to reduce environmental variability.
Data Presentation
Figure 1 (not shown) displays side-by-side box plots. The description in the report reads: 'The box plots indicate that the median reaction time for males (0.34 s) is higher than for females (0.29 s). The female IQR is narrower and the minimum is lower, suggesting greater consistency.' A comparative summary table is also provided:
| Group | Sample size n | Mean (s) | Standard deviation (s) |
|---|---|---|---|
| Male | 30 | 0.352 | 0.078 |
| Female | 30 | 0.298 | 0.062 |
Inferential Analysis
Since we are comparing means from two independent samples and population variances are unknown, a two-sample t-test is appropriate. The assumption of approximate normality is reasonable because sample sizes are 30 each, invoking the Central Limit Theorem. The test statistic is calculated as:
t = (0.352 − 0.298) / √0.078²/30 + 0.062²/30 = 0.054 / √0.0002028 + 0.0001281 = 0.054 / √0.0003309 ≈ 0.054 / 0.01819 ≈ 2.97.
The two-tailed p-value from the t-distribution is p ≈ 0.004. Since p < 0.05, we reject H₀.
Conclusion
There is sufficient evidence at the 5% ion times of male and female Year 12 students differ. The sample mean for females is 0.054 seconds faster, which could be meaningful in contexts requiring quick reflexes. However, the effect size (Cohen' derate-to-large practical difference, but this is only an estimate due to the sample size.
Evaluation
The stratified random sampling enhanced representativeness, but the sample was confined to one college, limiting generalisability. The reaction test itself may have introduced measurement error due to computer latency. Outliers were not identified, but one male with a very slow time (suspected distraction) could have inflated the mean. A larger, multi-site sample and more precise timing equipment would improve reliability. The assumption of equal variances was not formally tested, but the Welch approach allowed for unequal variances. Overall, the investigation provides credible evidence of a gender difference in this population.
Note: this is a condensed model; a full report would include all graphs and extended commentary.
以下是一篇 WJEC 风格的范文，调查男性和女性的反应时间是否存在差异。请注意各部分是如何紧扣探究循环的。
标题：男女高中生简单反应时间的比较
引言
反应时间是衡量对刺激反应速度的指标，可能受生物学和环境因素影响。研究问题是：“12 年级男生和女生的平均简单反应时间是否存在差异？”原假设 H₀: μ_男 = μ_女；备择假设 H₁: μ_男 ≠ μ_女，采用双尾检验，显著性水平 5%。
方法
从一所高中学校注册名单中，按性别分层使用随机数选取了 30 名男生和 30 名女生作为分层随机样本。每位参与者在一项计算机反应测试中完成三次测试，取平均值（秒）记录。变量“性别”为二元分类变量，“反应时间”为连续变量。所有测试均在同一安静房间、同一时间段进行，以减少环境变异。
数据展示
图 1（未显示）展示了并排箱线图。报告中的描述为：“箱线图显示，男生反应时间中位数（0.34 秒）高于女生（0.29 秒）。女生的四分位距更窄，最小值更低，表明其表现更为一致。”同时提供了一张对比汇总表：
| 组别 | 样本量 n | 均值（秒） | 标准差（秒） |
|---|---|---|---|
| 男 | 30 | 0.352 | 0.078 |
| 女 | 30 | 0.298 | 0.062 |
推断分析
由于比较两个独立样本的均值且总体方差未知，亦适用双样本 t 检验。因为每组样本量为 30，依中心极限定理，近似正态性的假设是合理的。检验统计量计算如下：
t = (0.352 − 0.298) / √0.078²/30 + 0.062²/30 = 0.054 / √0.0002028 + 0.0001281 = 0.054 / √0.0003309 ≈ 0.054 / 0.01819 ≈ 2.97。
使用 Welch-Satterthwaite 公式，自由度约为 54。根据 t 分布，双尾 p 值约为 0.004。因为 p < 0.05，拒绝 H₀。
结论
在 5% 显著水平下，有足够证据表明 12 年级男生和女生的平均反应时间不同。女生样本均值快 0.054 秒，这在需要快速反应的情境中可能有实际意义。然而，效应量（Cohen's d ≈ 0.77）表明实际差异处于中等到大之间，但这仅是受样本量影响的估计值。
评价
分层随机抽样增强了代表性，但样本仅限于一所学校，限制了推广性。反应测试本身可能因计算机延迟而引入测量误差。未识别出异常值，但一名男生反应极慢（怀疑分心）可能拉高了均值。更大的多地点样本和更精确的计时设备将提高可靠性。等方差假设虽未正式检验，但 Welch 方法允许方差不相等。总体而言，本次调查为该人群中存在性别差异提供了可信的证据。
注：此为精简版范文，完整报告应包含所有图形和扩展说明。
11. Common Mistakes That Cost Marks | 导致失分的常见错误
Even strong statisticians lose marks by neglecting the report structure. Common pitfalls include: confusing the research question with the hypothesis; forgetting to define parameters (μ, p) when stating H₀ and H₁; using pie charts for continuous data; reporting only the p-value without concluding in context; misapplying the Central Limit Theorem without checking sample size; not commenting on the shape of distributions before choosing a mean or median; and offering vague evaluations like 'more data would be better'. Every criticism in your evaluation should be specific and linked to the scenario. Additionally, many candidates fail to distinguish between population parameters and sample statistics, writing μ for a sample mean—carelessness that examiners penalise. Practice spotting these errors in past papers to sharpen your self-review skills.
即便是统计功底扎实的学生，也会因为忽视报告结构而失分。常见误区包括：混淆研究问题与假设；在陈述 H₀ 和 H₁ 时忘记定义参数（μ, p）；用饼图展示连续数据；只报告 p 值而不在情境中下结论；未检查样本量就误用中心极限定理；选择均值或中位数前不评论分布形状；给出模糊的评价，如“更多数据会更好”。评价中的每一条批评都应具体并与情境相关。此外，许多考生分不清总体参数和样本统计量，把样本均值写作 μ——这种粗心会被考官扣分。通过练习历年真题中的错误识别，你能提升自我审查能力。
12. Final Checklist for an Outstanding Report | 优秀报告的最终检查清单
Before submitting or moving on, run through this checklist:
- Is the research question clear and focused?
- Are H₀ and H₁ correctly stated with defined parameters?
- Is the sampling method described and evaluated for bias?
- Are graphs appropriate, labelled, and interpreted?
- Are descriptive statistics correctly chosen and compared?
- Have you justified the choice of probability distribution or test?
- Are hypothesis test steps logically laid out with working shown?
- Is the p-value correctly interpreted in context?
- Does the conclusion answer the question and mention statistical language?
- Is there a critical evaluation with specific strengths and improvements?
- Is the report written in precise, unbiased language without colloquialisms?
在提交或继续前，请逐项核对这份清单：
- 研究问题是否清晰聚焦？
- H₀ 和 H₁ 是否陈述正确，并定义了参数？
- 抽样方法是否描述并评估了偏差？
- 图形是否适当、标注并解读？
- 描述性统计量是否选择正确并进行了比较？
- 你是否为所选概率分布或检验给出了理由？
- 假设检验步骤是否逻辑清晰地展示，并写出了计算过程？
- p 值是否在情境下正确解读？
- 结论是否回答了问题并使用了统计语言？
- 是否有批判性评价，指出具体优点和改进建议？
- 报告是否用精确、无偏见的语言写成，无口语化表达？
Published by TutorHao | AS 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