Key Practical Assessment Points for Year 10 OCR Statistics | 实验/实践考核要点

📚 Key Practical Assessment Points for Year 10 OCR Statistics | 实验/实践考核要点

In the OCR GCSE Statistics course, practical investigations and experiments form a vital part of developing your statistical thinking. Whether you are designing a survey, conducting a controlled experiment, or interpreting real data, you must demonstrate a clear understanding of how to plan, carry out, and evaluate a statistical enquiry. This article brings together the essential practical assessment skills you will need, from formulating a hypothesis to avoiding common pitfalls. Use these points to strengthen your coursework preparation and to perform confidently in any data‑handling task.

在 OCR GCSE 统计课程中，实践调查和实验是培养统计思维的重要环节。无论是设计问卷、进行对照实验，还是解读实际数据，你都需要展示出清晰规划、实施和评估一项统计调查的能力。本文汇集了你必须具备的关键实践考核技能，从提出假设到避开常见陷阱，帮助你在课程作业和数据处理任务中胸有成竹。

1. Understanding the Task Brief | 理解任务简介

Read the brief several times and identify exactly what is being asked. Highlight key words such as “compare”, “estimate”, “investigate”, or “relationship”. This will tell you whether your analysis should focus on differences, trends, or associations.

反复阅读任务简介，准确找出要求你做什么。圈出“比较”“估计”“调查”“关系”等关键词，它们会告诉你分析应该侧重于差异、趋势还是关联。

Determine the type of variables involved: categorical (nominal or ordinal) or numerical (discrete or continuous). Knowing the variable type dictates which graphs and summary statistics are appropriate later.

确定涉及的变量类型：分类变量（名义或有序）或数值变量（离散或连续）。明确变量类型才能正确选择后续的图表和汇总统计量。

Identify the population and the sample you will work with. A clear definition prevents later confusion about the scope of your conclusions.

明确总体和你要使用的样本。清晰的定义能避免日后对结论适用范围产生混淆。

2. Formulating a Hypothesis | 提出假设

Write a null hypothesis (H₀) that states there is no effect, no difference, or no association. For example, H₀: μ₁ = μ₂, where μ₁ and μ₂ are the population means of two groups.

写出原假设 (H₀)，表明没有效应、没有差异或没有关联。例如 H₀: μ₁ = μ₂，其中 μ₁ 和 μ₂ 是两个组的总体均值。

Write an alternative hypothesis (H₁) that represents what you expect to find. Use directional (one‑tailed) or non‑directional (two‑tailed) wording based on prior knowledge. For instance, H₁: μ₁ > μ₂ suggests a specific direction.

写出备择假设 (H₁)，代表你期望发现的结果。根据已有知识选择定向（单尾）或非定向（双尾）表述。例如 H₁: μ₁ > μ₂ 表明了特定的方向。

Avoid vague hypotheses like “there will be a difference”. Always link the hypothesis to the measurable variables and state the population parameters clearly.

避免像“会存在差异”这样含糊的假设。始终将假设与可测量的变量联系起来，并清楚说明总体参数。

3. Choosing a Sampling Method | 选择抽样方法

Selecting the right sampling technique greatly influences the reliability of your results. The table below summarises common methods.

选择合适的抽样方法会极大地影响结果的可靠性。下表总结了常见的方法。

Method	Description	Advantage	Disadvantage
Simple Random	Every member has an equal chance of selection. 每个成员被选中的机会均等。	Unbiased, easy to understand. 无偏，易于理解。	Requires a full list; may miss subgroups. 需要完整名单；可能忽略子群体。
Systematic	Choose every k‑th individual from a list. 从列表中每隔 k 个选取一个。	Simple to implement. 实施简便。	Periodic patterns can introduce bias. 周期性模式可能引入偏差。
Stratified	Population divided into strata; random sample taken from each. 总体分成层；每层内随机抽样。	Ensures representation of key groups. 确保关键群体的代表性。	Need information on strata sizes. 需要知道各层大小。
Quota	Interviewers select a fixed number from each category. 访问员从各类别中选取固定数量。	Quick, no sampling frame needed. 快捷，无需抽样框。	Non‑random selection may introduce bias. 非随机选择可能引入偏差。

Match the method to your research question and available resources. For experiments, random allocation of subjects to treatment groups is essential.

让抽样方法与你的研究问题和可用资源匹配。对于实验，将受试者随机分配到处理组至关重要。

4. Designing the Data Collection Sheet | 设计数据收集表

Create a table or spreadsheet that clearly labels all variables. Include columns for the independent variable, the dependent variable, and any control variables that must be recorded (e.g., temperature, time).

制作一个表格或电子表格，清晰标注所有变量。为自变量、因变量以及必须记录的控制变量（如温度、时间）分别留出栏位。

Use closed‑format questions wherever possible to simplify data entry and reduce ambiguity. For example, list pre‑coded categories for responses.

尽可能使用封闭式问题，以简化数据录入并减少歧义。例如，给出预先编码的回答类别。

Always include space for notes about unexpected events or anomalies. This helps when you need to explain outliers later.

务必留出位置记录意外事件或异常。这有助于后续解释离群值。

5. Minimising Bias | 减少偏差

Bias can creep in at every stage. Selection bias occurs when the sample is not representative. Counter it by using random sampling or random allocation.

偏差可能渗透到每个阶段。当样本不具有代表性时便产生选择偏差。可通过随机抽样或随机分配加以避免。

Measurement bias arises if instruments are not calibrated or if questions are leading. Blind or double‑blind trials, where participants and/or assessors do not know group assignments, reduce this in experiments.

测量偏差的产生原因是工具未经校准或提问具有诱导性。在实验中，使用单盲或双盲试验（参与者及/或评估者不知分组情况）可减少此类偏差。

Non‑response bias happens when people who decline to participate differ systematically from those who do. Always record the response rate and consider its impact on findings.

无回应偏差是指拒绝参与的人与参与者存在系统性差异。始终记录回应率，并思考其对结论的影响。

6. Conducting the Experiment / Pilot Study | 进行实验 / 试点研究

Before the main data collection, run a small‑scale pilot study. This lets you test your procedure, check timing, and spot unanticipated problems.

在正式数据收集之前，先进行小规模试点研究。这能让你检验流程、核对时间安排并发现预料之外的问题。

During the actual experiment, keep conditions as constant as possible for all groups. If you are testing fertiliser, water every plant on the same schedule and use identical pots and soil.

在真正的实验中，尽可能让所有组的条件保持恒定。若测试肥料，应按相同时间表给每株植物浇水，并使用相同的花盆和土壤。

Repeat measurements (replication) improves reliability. Taking the mean of several readings helps smooth out random errors.

重复测量（重复）能提高信度。取几次读数的平均值有助于消除随机误差。

7. Recording Data Accurately | 准确记录数据

Record data to a consistent level of precision. If your ruler measures to the nearest 0.1 cm, write every length as 12.3 cm, not 12.30 or 12 cm.

以一致的精密度记录数据。如果尺子最小刻度是 0.1 cm，那么每个长度都应记录为 12.3 cm，而非 12.30 或 12 cm。

Use significant figures appropriately, and never adjust raw data without documenting the reason. Suspected outliers should be flagged, not erased.

恰当地使用有效数字，切勿在未记录原因的情况下改动原始数据。疑似离群值应做出标记，而不是删除。

Transfer data into a master file as soon as possible, double‑checking entries against the original record sheets to avoid transcription errors.

尽快将数据誊写到主文件，对照原始记录表核实输入，避免转录错误。

8. Organising and Presenting Data | 整理与呈现数据

Start with a grouped frequency table for continuous data. Choose class intervals of equal width, and check that boundaries do not overlap. For discrete data, a simple frequency distribution may be enough.

处理连续数据时，先编制分组频数表。选择等宽的组距，并确保组界不重叠。离散数据则用简单的频数分布表即可。

Graph selection depends on your purpose:

Bar chart – compare categorical frequencies.
条形图——比较分类频数。
Pie chart – show proportions of a whole.
饼图——展示整体各部分的比例。
Histogram – display the shape of a continuous distribution (area proportional to frequency).
直方图——显示连续分布的形状（面积与频数成正比）。
Box plot – summarise median, quartiles, and outliers.
箱线图——概括中位数、四分位数及离群值。
Scatter graph – explore relationship between two numerical variables.
散点图——考察两个数值变量之间的关系。

Label axes clearly, include units, and give every chart a title. A well‑presented graph communicates your findings instantly.

清晰标注坐标轴、注明单位，并为每张图表加上标题。良好的图表呈现能让你的发现一目了然。

9. Analysing Data: Measures and Graphs | 数据分析：度量与图表

Calculate appropriate averages: the mean (x̄ = Σx / n) for symmetric data without outliers, and the median for skewed data or when outliers are present. The mode can be useful for categorical data.

计算适当的平均值：对于无离群值的对称数据使用均值 (x̄ = Σx / n)，对于偏态数据或存在离群值时使用中位数。众数可用于分类数据。

Measure spread using the range (max − min) or, better, the interquartile range:

IQR = Q₃ − Q₁

This tells you the spread of the middle 50 % of the data and is resistant to outliers.

使用极差（最大值 − 最小值）或更好的四分位距度量离散程度：

IQR = Q₃ − Q₁

它反映中间 50 % 数据的分散情况，且不受离群值影响。

Draw a box plot to compare groups visually. Side‑by‑side box plots quickly reveal differences in median and spread. Look for overlaps and shifts.

绘制箱线图以直观比较各组。并列箱线图能迅速揭示中位数和离散度的差异，注意重叠和位移。

For bivariate data, calculate the line of best fit and describe the correlation (positive, negative, or none). Remember: correlation does not imply causation.

对于双变量数据，计算最佳拟合线并描述相关性（正、负或无）。切记：相关不意味着因果。

10. Drawing Conclusions and Evaluating | 得出结论并评估

Return to your original hypothesis and state whether the evidence supports it. Use comparative language such as “the median height of fertilised plants was 3.2 cm greater than the control group”.

回到最初假设，说明证据是否支持它。使用对比性语言，如“施肥植物的中位数高度比对照组高 3.2 cm”。

Discuss the reliability of your findings. Mention the sample size, the presence of any outliers, and whether the data was consistent across repeats. A confidence interval or reference to the spread can strengthen your argument.

讨论结论的可靠性。提及样本量、是否存在离群值，以及多次重复的数据是否一致。引用置信区间或散度信息能增强论证。

Critically evaluate your method. What worked well? What would you change? Suggest specific improvements, such as increasing the sample size, using more precise instruments, or controlling an additional variable.

严格评估你的方法：哪些做得好？哪些需要改变？提出具体的改进措施，如扩大样本量、使用更精密的仪器或控制另一个变量。

11. Common Pitfalls to Avoid | 需要避免的常见陷阱

Small sample size leads to high variability and makes it hard to detect real effects. Always aim for the largest feasible sample within your constraints.

样本量过小会导致变异度高，难以检测真实效应。在条件允许的情况下，始终争取最大的可行样本量。

Ignoring confounding variables can produce misleading conclusions. If you are testing a new revision technique, the time of day or prior ability could affect results – control or randomise these factors.

忽视混杂变量可能得出误导性结论。若测试一种新复习方法，学习时段或先前能力都可能影响结果——应控制或随机化这些因素。

Handling outliers blindly can distort the analysis. Instead, check whether the outlier is a recording error, a natural variation, or a sign of something interesting, then decide on a justified treatment.

盲目处理离群值会扭曲分析。正确的做法是检查离群值是记录错误、自然变异还是有趣现象的征兆，然后做出有依据的处理决定。

Confusing correlation with causation is a classic mistake. Just because two variables rise together does not mean one causes the other; there may be a lurking third factor.

混淆相关与因果是经典错误。两个变量同向变化并不意味着一个导致另一个；可能存在隐藏的第三因素。

Finally, poor time management during an assessment can mean rushed analysis. Plan your timetable backwards from the deadline, allocating clear slots for design, data collection, analysis, and write‑up.

最后，考核期间时间管理不当会导致仓促分析。从截止日期倒推制定时间表，为设计、数据收集、分析和报告撰写安排明确的时间段。

Published by TutorHao | Statistics Revision Series | aleveler.com

更多咨询请联系16621398022（同微信）

Key Practical Assessment Points for Year 10 OCR Statistics | 实验/实践考核要点