📚 Exam Techniques and Marking Criteria for GCSE Cambridge Statistics | GCSE Cambridge 统计：答题技巧与评分标准
Mastering GCSE Cambridge Statistics requires more than just knowing the formulas – you need a clear strategy to present your answers exactly as examiners expect. This guide unpacks the marking criteria, highlights common pitfalls, and shares proven techniques to help you secure every mark available. From interpreting stem-and-leaf diagrams to justifying sampling methods, you will learn how to structure responses that demonstrate both statistical reasoning and clarity of communication.
要掌握 GCSE Cambridge 统计学，光知道公式是不够的——你需要清晰的策略，用考官期望的方式呈现答案。本指南拆解评分标准，点明常见陷阱，并分享行之有效的技巧，帮助你拿到每一个可得的分数。从解读茎叶图到证明抽样方法，你将学会如何组织答题，既展现统计推理能力，又保证表达清晰。
1. Understanding the Marking Criteria | 理解评分标准
Cambridge examiners allocate marks using a method–accuracy–communication (MAC) approach. Method marks are awarded for setting up a correct process, such as choosing the right probability formula or calculating cumulative frequencies. Accuracy marks follow correct working, even if a small arithmetic slip occurs earlier, as long as the subsequent steps are valid. Communication marks reward clarity – these require labelled axes, properly stated hypotheses, and conclusions expressed in the context of the question.
剑桥考官遵循“方法–准确性–表达”的给分方式。方法分给予正确的解题思路，比如选用正确的概率公式或计算累积频率。只要后续步骤合理，即使前面出现算术小错误，准确性分数依然可以获得。表达分奖励清晰度——这要求坐标轴有标签、假设明确陈述，以及结合题目情境给出结论。
It is crucial to read the mark allocation beside each question. A 3-mark question usually expects one mark for method, one for accuracy, and one for interpretation. Never leave a partially answered question blank when you can still show a logical approach, because the method mark may be independent of the final answer.
仔细看每道题旁的分值分配至关重要。一道3分的题目通常预期1分给方法、1分给准确性、1分给解释。如果你还能展示逻辑思路，就绝不要空着未答完的题目，因为方法分可能不依赖于最终答案。
For 'explain' or 'comment' questions, marks are explicitly for using statistical vocabulary. Phrases like 'positive skew', 'median is a better measure of central tendency than mean for skewed data', or 'the sample is not representative because it excludes...' directly align with the communication category.
对于“解释”或“评论”类题目，分数明确奖励使用统计术语。“正偏态”、“对于偏态数据，中位数比均值更适合作为集中趋势的度量”，或“样本不具备代表性，因为它排除了……”等表述，直接对应表达类得分。
2. Using Statistical Terminology Accurately | 准确使用统计术语
Examiners are strict about the difference between 'discrete' and 'continuous' data. Stating that shoe sizes are discrete when they are actually continuous (if measured precisely) can cost you a definition mark. Similarly, 'primary data' means you collected it yourself; 'secondary data' comes from existing sources. Describing a scatter graph, use 'positive correlation' rather than just 'correlation'.
考官对“离散数据”和“连续数据”的区别要求严格。如果鞋码实际是连续的（若精确测量），却说它是离散的，就可能丢掉定义分。同样，“一手数据”指你自己收集的数据；“二手数据”来自已有来源。描述散点图时，要用“正相关”而非仅仅“相关”。
When referring to averages, emphasize the appropriate choice: 'The median is unaffected by extreme values, so it is better to describe typical wages than the mean.' For spread, state 'interquartile range' instead of 'range' when the question includes outliers. Confidence in terms like 'bivariate data', 'explanatory variable', and 'response variable' distinguishes a high-tier answer.
提及平均值时，要强调恰当的选择：“中位数不受极端值影响，所以比均值更适合描述典型工资。”对于离散程度，若题目包含异常值，就用“四分位距”而非“极差”。自信地使用“二元数据”、“解释变量”、“响应变量”等术语，是高分答卷的标志。
In probability, distinguish 'experimental probability' (from trials) and 'theoretical probability' (from symmetry). If you carry out a simulation, you are estimating probability, not calculating an exact value. Use 'relative frequency' correctly rather than incorrectly calling it 'probability' in an experimental context.
在概率中，要区分“实验概率”（来自试验）和“理论概率”（来自对称性）。如果进行模拟，你是在估计概率，而非计算精确值。在实验情境下，正确使用“相对频率”，而非错误地称之为“概率”。
3. Showing Working Steps Clearly | 清晰展示计算步骤
Cambridge statistics papers require you to write down intermediate values. For standard deviation, do not just key numbers into a calculator and write the final answer; you risk losing method marks if a transcription error occurs. Show the sum of squares, the variance, and then the square root. Use the formula sheet wisely – annotate which symbols correspond to your data.
剑桥统计试卷要求写出中间过程。对于标准差，不要只将数字输入计算器然后写下最终答案；如果发生抄录错误，你可能会丢掉方法分。展示平方和、方差，再开方。善用公式表——注明哪些符号对应你的数据。
For grouped frequency calculations, explicitly list midpoints, multiply by frequency, and show the cumulative frequency column where needed. When computing Spearman's rank correlation coefficient, lay out the ranks in a table, calculate the differences d, and square them step by step. This structured approach not only earns more method marks but also helps you spot errors.
对于分组频率计算，明确列出组中值，乘以频数，并按要求展示累积频率列。计算斯皮尔曼等级相关系数时，在表格中排好等级，计算差值 d，再逐步平方。这种结构化的方式不仅能获得更多方法分，还能帮助你发现错误。
When constructing tree diagrams for probability, label branches with probabilities as fractions or decimals, and write the combined probability at the end of each path. Some marks are specifically for completing the diagram and adding all end-point probabilities that sum to 1.
为概率绘制树状图时，以分数或小数标注各分支概率，并在每条路径末端写上组合概率。有些分数专门用于完成图表并确保所有终结点概率之和为1。
4. Drawing and Annotating Graphs and Charts | 绘制与标注图表
Diagrams attract multiple marks for accuracy, scale, and labelling. A histogram for unequal class widths must use frequency density on the vertical axis – simply plotting frequency will lose all marks. Include a title, label axes with the variable name and unit, and use a ruler for straight lines. Bar charts for discrete data should have gaps between bars, while histograms for continuous data must have bars touching.
图表因准确性、刻度和标注而获得多项分数。不等组距的直方图必须在纵轴使用频率密度——若只是绘制频率，则会丢失全部分数。加上标题，用变量名称和单位标注坐标轴，并使用直尺画直线。离散数据的条形图，条形间应有空隙；连续数据的直方图，条形必须紧挨。
Cumulative frequency graphs require plotting points at the upper class boundary against the cumulative total. Many candidates mistakenly use midpoints. Connect the points with a smooth curve, not straight line segments. When reading off medians or quartiles, draw horizontal lines back to the y-axis and show the reading clearly on the axis.
累积频率图需要根据各组上限值描绘点，对应累积总数。许多考生误用组中值。用平滑曲线连接各点，而非直线段。读取中位数或四分位数时，画水平线回到y轴，并在轴上清晰显示读数。
Scatter diagrams must have a line of best fit drawn through the centroid (mean of x, mean of y) with equal points above and below. For correlation questions, you may need to describe the relationship, note any outliers, and comment on the strength. Always place the independent variable on the horizontal axis and the dependent on the vertical.
散点图必须画一条通过中心点（x均值, y均值）的最佳拟合线，两侧点数大致相等。对于相关性问题，你可能需要描述关系，注明任何异常值，并评论强度。始终将自变量放在水平轴，因变量放在垂直轴。
5. Interpreting and Explaining Results | 解释与诠释结果
When a question asks 'compare' two sets of data, you must use comparatives: 'On average, the heights of boys are greater than girls, because the median is 162 cm compared to 155 cm.' Then address spread: 'The interquartile range for boys is larger, indicating greater variation.' Do not just list statistics; you need a comparative statement for central tendency and one for dispersion.
当题目要求“比较”两组数据时，必须使用比较级：“平均而言，男孩的身高大于女孩，因为中位数为162厘米，而女孩为155厘米。”然后讨论离散程度：“男孩的四分位距更大，表明变异更大。”不要只列出统计量；你需要分别对集中趋势和离散度做出比较性陈述。
In probability contexts, after calculating an experimental probability of 0.37, you might interpret: 'This suggests about 37% of trials resulted in success, but with more trials the relative frequency may approach the theoretical probability.' Link back to the law of large numbers if appropriate.
在概率情境中，计算出实验概率0.37后，可以这样解释：“这表明约37%的试验结果为成功，但随着试验次数增加，相对频率可能趋近于理论概率。”如果合适，可联系大数定律。
For correlation, avoid causal language unless the context justifies it. Say 'There is a strong positive correlation between hours of revision and test score', not 'More revision causes higher scores.' Examiners deduct marks for assuming causation without evidence from a designed experiment.
对于相关性，除非情境合理，否则避免因果用语。应说“复习时数和测试分数之间存在强正相关”，而不是“更多复习导致更高分数”。考官会因在没有实验设计证据的情况下假定因果关系而扣分。
6. Addressing Reliability and Bias | 处理可靠性与偏差
Questions about sampling methods are common. A random sample is unbiased if every member of the population has an equal chance of selection. If you describe a systematic sample (every 10th person), mention the starting point was chosen randomly. A stratified sample requires proportional representation from groups – you must explain how you calculate the number from each stratum and why this improves representativeness.
关于抽样方法的问题很常见。如果总体中每个成员被选中的机会均等，随机样本就是无偏的。如果你描述系统抽样（每隔10人抽一人），要提及起始点是随机选择的。分层抽样需要各组按比例代表——必须说明如何计算每层的数量，以及为什么这样能提高代表性。
When evaluating reliability, link to sample size: 'A larger sample size would reduce sampling error and make the estimate more precise.' Identify possible sources of bias, such as voluntary response bias, convenience sampling, or measurement error. Always suggest a concrete improvement, e.g., 'Use a numbered list and a random number generator to select participants.'
评估可靠性时，要联系样本量：“更大的样本量会减小抽样误差，使估计更精确。”找出可能的偏差来源，如自愿回应偏差、便利抽样或测量误差。始终提出具体的改进措施，例如“使用编号列表和随机数生成器选择参与者”。
In surveys, acknowledge leading questions or poorly designed response categories. If a question asks how to improve a questionnaire, mention open/closed question format, pilot testing, and anonymity to encourage honest answers.
在调查中，承认诱导性问题或设计不当的响应类别。如果题目问如何改进问卷，提及开放式/封闭式问题格式、试点测试，以及为鼓励诚实回答而采用的匿名方式。
7. Time Management and Paper Strategy | 时间管理与试卷策略
Cambridge IGCSE Statistics papers often contain a mixture of short structured questions and longer problem-solving tasks. Begin by scanning the entire paper to gauge question weight. Allocate time proportionally: if a question carries 10% of the total marks, spend roughly 10% of the exam time on it. Leave the final 10 minutes for checking calculations.
剑桥 IGCSE 统计学试卷通常混合了简短的建构式问题和较长的解题任务。开始时浏览全卷，判断各题分值权重。按比例分配时间：如果一道题占总分的10%，就花大约10%的考试时间做它。留出最后10分钟用于检查计算。
Do not get stuck on one diagram. If a cumulative frequency graph is taking too long because you are unsure of a scale, move on and return later. An incomplete graph that you can partially plot may still earn method marks, whereas spending 20 minutes on it may cause you to miss easier marks elsewhere.
不要在一张图上卡住。如果累积频率图因为不确定刻度而耗时太久，就跳过，稍后再回来看。你不完整但已部分绘制的图仍可能得到方法分，而在上面耗费20分钟则可能导致你错过其他地方的简单分数。
In multiple-part questions, the later parts often depend on earlier calculations. Write down all your intermediate values clearly so that even if an early answer is wrong, the follow-through marks can be awarded for correct use of your own figures.
在分部分设问的题目中，后续部分常依赖于前面的计算。清晰写下所有中间值，这样即使前面的答案错了，后续能正确使用你自己的数据，仍可获得连带分数。
8. Common Mistakes and How to Avoid Them | 常见错误与避免方法
One frequent error is confusing 'interquartile range' with 'range'. IQR = Q3 – Q1, not max – ing off values from both ends to find the median first, then find the median of each half.
一个常见错误是混淆“四分位距”和“极差”。IQR = Q3 – Q1，而非最大减最小。从茎叶图中准确识别四分位数，要从两端划去数值先找出中位数，然后找出每半部的中位数。
Another mistake is plotting cumulative frequency at the midpoint of an interval. The correct point is the upper class boundary. Check the horizontal axis scale; if it starts at 20 but your first upper boundary is 25, you must plot at 25, not 22.5.
另一个错误是将累积频率绘制在区间的中点。正确的点是组上限。检查横轴刻度；如果从20开始，但第一个上限是25，则必须在25处描点，而非22.5。
he events are mutually exclusive. Use a sample space diagram or tree diagram to visualize the problem and avoid this confusion.
在概率中，考生有时本应相乘却用了加法。对于独立事件，P(A 且 B) = P(A) × P(B)。仅当事件互斥时才相加。使用样本空间图或树状图来将问题可视化，避免混淆。
9. Tackling Long-Answer Questions | 应对长篇回答题
Extended response questions in Statistics often present a scenario – such as comparing two distributions or designing a data collection plan. Structure your answer logically: state what you are comparing, choose appropriate measures, present calculations, then interpret in plain English within the context. Use subheadings mentally to organize your paragraphs.
统计学中的拓展回答题常常给出一个情境——例如比较两个分布或设计数据收集计划。按逻辑组织结构你的回答：说明要比较什么，选择适当的度量，展示计算，然后用通俗的英语结合情境进行解释。在脑海中用小标题来组织段落。
If asked to 'plan an investigation', specify your sampling method, how you will ensure randomness or stratification, the sample size, the data you will collect (types and units), and how you will present results. Mention ethical considerations if relevant, such as confidentiality. Examiner reports often note that answers lacking justification for sample size lose marks.
如果被要求“规划一项调查”，明确说明抽样方法、如何确保随机性或分层、样本量、将收集的数据（类型和单位），以及如何呈现结果。如有必要，提到伦理考量，如保密性。考官报告常指出，未证明样本量合理性的答案会丢分。
For a 'critique' question, examine the entire process: sampling bias, question wording in a survey, measurement accuracy, and whether the conclusion matches the data. Use terms like 'validity' and 'reliability' precisely. A strong critique also proposes specific improvements, not vague suggestions.
对于“批判性评价”类题目，审视整个过程：抽样偏差、调查中的问题措辞、测量准确性，以及结论是否与数据相符。准确使用“效度”和“信度”等术语。强有力的批判还会提出具体的改进措施，而非模糊的建议。
10. Checking and Reviewing Answers | 检查与回顾答案
Reserve at least 5–10 minutes at the end to review your answers. Check directions: Did you list the data in order when finding the median? Did you express probability as a fraction in its simplest form if required? Recalculate key figures like totals to ensure they add up correctly. Confirm that graphs have titles and axis labels.
最后至少留出5-10分钟检查答案。检查要求：求中位数时，你是否按顺序列出了数据？若要求，你是否将概率化为最简分数？重新计算总数等关键数字，确保加总正确。确认图表有标题和坐标轴标签。
Proofread interpretive statements for references to the context. Avoid generic phrases like 'the data is spread out' – specify 'the scores are more varied, with an IQR of 15 marks'. If a comparison question, ensure you have used comparatives and mentioned both data sets.
校对解释性陈述，确保提到了情境。避免通用说法，如“数据很分散”——具体化：“分数差异更大，四分位距为15分”。若是比较类问题，确保使用了比较级，并提及了两组数据。
Finally, ensure you have answered all sub-parts. Sometimes a question has an instruction hidden at the end, such as 'Give your answer correct to 3 significant figures' or 'State an assumption you made'. Missing these can cost a mark even if your working is perfect.
最后，确保已回答所有小问。有时题目末尾隐藏着要求，如“答案精确到3位有效数字”或“陈述你做出的一个假设”。即使你的计算完美，遗漏这些也可能扣分。
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