📚 Year 11 SQA Statistics: Experimental/Practical Assessment Key Points | Year 11 SQA 统计：实验/实践考核要点
In the SQA Statistics course at Year 11 level, the experimental or practical assessment allows you to demonstrate your ability to plan, carry out, analyse and evaluate a statistical investigation. Success hinges not only on correct calculations but on a clear understanding of the entire statistical cycle. This article walks you through the essential elements that examiners will look for, from formulating a hypothesis to presenting a coherent final report.
在 Year 11 阶段的 SQA 统计课程中，实验或实践考核旨在检验你规划、实施、分析和评价一项统计调查的能力。成功的关键不仅在于正确的计算，还在于对整个统计循环的清晰理解。本文将逐一梳理考官所关注的核心要点，从形成假设到呈现一份连贯的最终报告。
1. Understanding the Assessment Objectives | 理解考核目标
The practical assessment is designed to test your ability to apply statistical concepts to a real-world problem. You are expected to identify a clear aim, collect appropriate data, process that data using suitable techniques, and reach a justified conclusion. Marks are awarded for each stage of the investigation, so a logical and well-documented approach is vital.
实践考核旨在测试你将统计概念应用于真实问题的能力。你需要确定清晰的目标，收集合适的数据，使用恰当的方法处理数据，并得出有理有据的结论。调查的每一个阶段都有相应分值，因此一个逻辑清晰且记录详尽的过程至关重要。
Examiners will look for evidence of independent thinking and the ability to link the statistical techniques you choose to the context of your investigation. A purely mechanical repetition of calculations without meaningful commentary will limit your score. Always show your working and explain why you selected a particular diagram or measure.
考官会寻找独立思考的证据，以及你将所选统计方法与调查情境联系起来的能力。仅仅机械地重复计算而缺乏有意义的评论，会限制你的得分。始终展示你的运算过程，并解释为何选择某一图表或度量指标。
2. Formulating a Research Question and Hypothesis | 形成研究问题与假设
Start by identifying a topic that allows for measurable variables. Your research question should be specific and feasible within the time available. For example, instead of 'Are students fit?', use 'Is there a difference in the median time Year 11 boys and girls can hold a plank?' A well-framed question leads naturally to a testable hypothesis.
首先要确定一个允许有可测量变量的主题。你的研究问题应当具体且在可用时间内可行。例如，不要用“学生健康吗？”，而要用“Year 11 男生和女生平板支撑时间的中位数是否存在差异？”一个设计良好的问题会自然地导向可检验的假设。
Express your hypothesis using both a null hypothesis (H₀) and an alternative hypothesis (H₁). In the SQA context, these may be stated in words rather than symbols. For instance, H₀: There is no difference in the median plank times between Year 11 males and females; H₁: There is a difference. Ensure your hypothesis relates directly to the statistical test you plan to use.
使用零假设 (H₀) 和备择假设 (H₁) 来表达你的假设。在 SQA 的情境中，通常用文字而非符号来陈述。例如，H₀：Year 11 男女生平板支撑时间的中位数没有差异；H₁：存在差异。确保你的假设与你计划使用的统计检验直接相关。
3. Sampling Strategies: Random, Stratified and Systematic | 抽样策略：随机、分层与系统抽样
A reliable investigation depends on how well your sample represents the population. You must describe your sampling frame and the specific method you used. Simple random sampling gives every member an equal chance but may not capture subgroups equally; stratified sampling ensures that key subgroups (e.g. gender, year group) are proportionally represented.
一项可靠的调查取决于样本对总体的代表程度。你必须描述你的抽样框和所使用的具体方法。简单随机抽样让每个成员都有均等的机会，但可能无法平等地捕捉到各个子群；分层抽样则确保关键子群（如性别、年级）按比例被代表。
If you use systematic sampling (e.g. selecting every 10th pupil from a register), be clear about the starting point and interval. Always justify your choice of sampling method and comment on any potential bias introduced by it. For instance, convenience sampling using your own friend group is easy but rarely representative and will be criticised in the evaluation.
如果你使用系统抽样（例如从名单中每隔10人选取一名），请明确说明起始点和间隔。始终对你选择的抽样方法进行辩护，并评论该方法可能引入的任何潜在偏差。例如，使用你自己的朋友圈进行便利抽样虽然简单，但很少具有代表性，会在评估环节受到批评。
4. Designing Effective Data Collection Instruments | 设计有效的数据收集工具
Your data collection sheet or questionnaire must be clear, efficient and free from ambiguous questions. Pre-test your questions on a small pilot group to check for understanding. Include fields for key identifiers and the variables you need to measure, and consider whether data will be discrete, continuous, ordinal or categorical.
你的数据收集表或问卷必须清楚、高效且不含含糊的问题。在一个小型的试点小组中预测试你的问题，以检查其是否易于理解。包含关键标识符和你需要测量的变量的字段，并考虑这些数据将是离散型、连续型、顺序型还是分类型。
When collecting quantitative data, specify the units and the precision of measurement. For example, record times to the nearest second or lengths to the nearest centimetre. Use tally charts for grouped data where appropriate. The design of your instrument directly affects the reliability of your raw data, so invest time in getting this right.
收集定量数据时，请指明单位和测量精度。例如，将时间记录到最接近的秒，或将长度记录到最接近的厘米。在适当的情况下使用分组数据的划记表。你的工具设计直接影响原始数据的可靠性，因此值得花时间把它做好。
5. Ethical Considerations and Bias Control | 伦理考量与偏差控制
You must consider ethical practice: obtain informed consent from participants, assure them of anonymity and confidentiality, and allow them to withdraw at any time. State clearly that data will be stored securely and only used for the purpose of the investigation. These points must be documented in your report.
你必须考虑伦理实践：获得参与者的知情同意，向他们保证匿名和保密，并允许他们随时退出。明确说明数据将被安全存储，且仅用于本次调查的目的。这些要点必须在报告中加以记录。
Bias can creep in at many stages: leading questions, non-response, measurement errors, or an unrepresentative sample. Identify where bias might occur in your own investigation and describe the steps you took to minimise it. Acknowledging the limitations of your method shows a mature statistical understanding.
偏差可能出现在许多阶段：诱导性问题、无回答、测量误差或样本不具代表性。要识别你自己的调查中可能发生偏差的地方，并描述你为最小化偏差所采取的步骤。承认你方法的局限性，这体现了成熟的统计理解。
6. Organising, Cleaning and Tabulating Data | 整理、清理与列表呈现数据
Raw data is rarely ready for analysis. Remove any obvious outliers that result from recording errors, but do not discard valid extreme values without justification. Document any cleaning steps, such as removing a response where a participant clearly misunderstood the question. Sort data into ordered lists or frequency tables as appropriate.
原始数据很少已准备好进行分析。移除任何由记录错误引起的明显异常值，但没有正当理由不要丢弃有效的极端值。记录所有的清理步骤，例如移除参与者明显误解了问题的回答。根据情况将数据整理成有序列表或频数表。
For grouped data, choose sensible class intervals—usually 5 to 10 groups—and ensure each interval is of equal width unless there is a strong contextual reason not to. Include columns for frequency, cumulative frequency and relative frequency if needed. A well-structured table makes subsequent calculations and graph plotting much easier.
对于分组数据，选择合理的组距——通常为5到10组——并确保每个区间的宽度相等，除非有强烈的背景理由不允许这样。根据需要包含频数、累积频数和相对频数列。一个结构良好的表格会使随后的计算和图表绘制容易得多。
7. Selecting and Drawing Suitable Statistical Diagrams | 选择与绘制合适的统计图表
The diagram you choose must match your data type and your investigation aim. Use bar charts for categorical data, histograms for continuous grouped data, and box plots for comparing distributions of continuous data. Scatter graphs are essential for exploring relationships between two quantitative variables.
你选择的图表必须与你的数据类型和调查目的相匹配。分类数据使用条形图，连续的分组数据使用直方图，比较连续数据的分布使用箱线图。散点图对于探索两个定量变量之间的关系至关重要。
Ensure all graphs are fully labelled with a title, axes labels and units. Scales must be linear and consistent; break the scale only if absolutely necessary and make the break obvious. When drawing a scatter graph, add a line of best fit only if the correlation is reasonably strong, and justify whether you are using a straight line or a curve.
确保所有图表都有完整的标签：标题、坐标轴标签和单位。刻度必须是线性的且一致的；只有当绝对必要时才断开刻度，并使断开处显而易见。绘制散点图时，只有当相关性相当强时才添加最适拟合线，并论证你使用的是直线还是曲线。
8. Calculating Summary Statistics Accurately | 准确计算汇总统计量
You are expected to compute measures of central tendency (mean, median, mode) and measures of dispersion (range, interquartile range, standard deviation). Show your working clearly, especially for standard deviation, which is often tested. The SQA formula for sample standard deviation is usually given, but you should be able to apply it step by step.
你需要计算集中趋势的度量（均值、中位数、众数）和离散度的度量（极差、四分位距、标准差）。请清晰地展示运算过程，尤其是经常被考核的标准差。SQA 通常提供样本标准差公式，但你应当能够逐步应用它。
Use the formula: s = √( Σ(x - x̄ n using calculator functions; you may need to write out the intermediate columns (x, x - x̄, (x - x̄)²) to gain full marks. Interpret what these statistics tell you in the context—for example, a smaller standard deviation indicates more consistent results.
使用公式：s = √( Σ(x - x̄)² / (n-1) )，其中 n 为样本大小。使用计算器功能时要小心；你可能需要列出中间栏（x, x - x̄, (x - x̄)²）以获得满分。要在情境中解释这些统计量告诉你什么——例如，较小的标准差表明结果更一致。
For comparative investigations, calculate summary statistics for each group separately. Use box plots to compare medians and IQRs visually. Then state whether your findings support or refute your original hypothesis, but do not claim absolute proof—statistical evidence is about likelihood, not certainty.
对于比较性的调查，分别计算每个组的汇总统计量。使用箱线图来直观地比较中位数和四分位距。然后说明你的发现是支持还是反驳了你最初的假设，但不要宣称绝对证明——统计证据涉及的是可能性，而非确定性。
9. Interpreting Results and Drawing Conclusions | 解释结果与得出结论
Interpretation is where you demonstrate deeper understanding. Link your calculated statistics and graphs back to your original question. If you found a difference in medians, ask whether this difference is meaningful in the real world, not just statistically detectable. Quote actual values from your data to support every claim.
在解释环节，你将展示更深入的理解。将你计算出的统计量和图表与最初的问题联系起来。如果你发现中位数存在差异，要问这个差异在现实世界中是否有意义，而不仅仅是在统计上可检测。引用数据中的实际数值来支持每一个论断。
Avoid overgeneralising—your conclusions apply only to your sample and the wider population it represents. Use phrases like 'The sample suggests that...' or 'Within this investigation, there is some evidence that...'. Acknowledge that a different sample might produce different results.
避免过度概括——你的结论仅适用于你的样本及其所代表的更广泛总体。使用诸如“样本表明……”或“在本次调查中，有一些证据表明……”这样的措辞。承认不同的样本可能会有不同的结果。
10. Evaluating the Investigation Thoroughly | 彻底评估调查过程
A robust evaluation discusses both strengths and limitations of your approach. Strengths might include a carefully designed sampling plan, large sample size, or effective piloting. Limitations could be time constraints, a sampling frame that missed certain groups, or measurement inaccuracies.
一份扎实的评估会讨论你方法的优点和局限性。优点可能包括精心设计的抽样计划、较大的样本量或有效的试点。局限性可能是时间限制、忽略了某些群体的抽样框，或测量不准确。
Be specific and avoid generic comments like 'I could have gathered more data.' Instead, say 'Increasing the sample size to at least 50 per group would improve the reliability of the standard deviation estimates.' Suggest concrete improvements for each limitation you identify, and reflect on how these changes might affect the conclusions.
要具体，避免像“我可以收集更多数据”这样的笼统评论。相反，应说“将每组样本量增加到至少50个，将会提高标准差估计值的可靠性。”为你发现的每个局限性提出具体的改进建议，并思考这些改变会如何影响结论。
11. Structuring Your Report for Maximum Clarity | 以最清晰的结构撰写报告
Your final report should follow a logical sequence: Title, Introduction (including hypothesis), Method (sample and data collection), Results (tables, calculations, graphs), Analysis and Interpretation, Conclusion, and Evaluation. Use clear headings and subheadings to guide the reader. Number all tables and figures.
你的最终报告应遵循逻辑顺序：标题、引言（包括假设）、方法（样本与数据收集）、结果（表格、计算、图表）、分析与解释、结论，以及评估。使用清晰的标题和子标题引导读者。为所有表格和图表编号。
Within the report, embed your reasoning at every stage. Do not simply place a graph; explain what features to notice. For calculations, show the formula, the substituted values, and the result. Keep the language formal but accessible. Proofread to eliminate errors that could undermine the credibility of your work.
在报告中，要在每个阶段嵌入你的推理。不要只是放置一张图表；要说明需要注意哪些特征。对于计算，要展示公式、代入的数值和结果。保持语言正式但易于理解。仔细校对，以消除可能损害你工作可信度的错误。
12. Common Pitfalls and How to Avoid Them | 常见陷阱及其避免方法
A common mistake is confusing discrete and continuous data, which leads to inappropriate graphs or calculations. Another is ignoring units or using mixed units within a dataset. Always standardise measurements before analysis. Also, avoid performing calculations on data that has not been cleaned—an outlier due to recording '180' instead of '18' can distort your mean dramatically.
一个常见的错误是混淆离散数据和连续数据，这会导致不恰当的图表或计算。另一个错误是忽视单位，或在数据集中混用不同单位。在分析之前，始终将度量标准化。此外，避免对未经清理的数据进行计算——一个因将“18”误录为“180”而产生的异常值会极大地扭曲你的均值。
Drawing a line of best fit through a scatter graph with no visible correlation is a frequent error. Only add a line if the relationship is reasonably linear and you can justify using it for estimation. Finally, do not confuse correlation with causation; if two variables are correlated, state that they are associated, not that one necessarily causes the other.
在没有明显相关性的散点图上画一条最适拟合线，是一个频繁出现的错误。只有当关系相当线性且你能证明其可用于估计时，才添加该线。最后，不要混淆相关与因果；如果两个变量相关，说明它们有关联，而不是一个必然导致另一个。
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