📚 Year 13 SQA Statistics: Exam Techniques and Marking Criteria | Year 13 SQA 统计：答题技巧与评分标准
This article provides comprehensive guidance on mastering the SQA Statistics examination for Year 13 students. Understanding the marking criteria and applying effective exam techniques can significantly boost your performance. We will break down the key areas assessed, common pitfalls, and how to present your answers to maximise marks.
本文为 Year 13 学生提供全面的 SQA 统计考试通关指南。理解评分标准并运用有效的答题技巧可以显著提高你的成绩。我们将详细分析考查的重点领域、常见失分点，以及如何呈现答案以获得最高分。
1. Understanding the SQA Marking Scheme | 理解 SQA 评分方案
In SQA Statistics, marks are awarded for method, accuracy, and interpretation. Even if your final answer is incorrect, you may still earn most of the marks by showing correct steps and reasoning. Examiners follow a positive marking approach, looking for evidence of understanding.
在 SQA 统计考试中，评分依据方法、准确性和解读。即使最终答案错误，只要展示了正确的步骤和推理，仍可能获得大部分分数。考官遵循正向评分原则，寻找理解的证据。
Always refer to past paper marking instructions to identify 'key idea' marks, such as stating hypotheses in hypothesis testing, selecting the correct test, and giving a conclusion in context. Often, marks are split: one for calculation, one for explanation.
始终参考历年真题评分说明，识别“关键点”分数，例如在假设检验中陈述假设、选择正确的检验，并结合背景给出结论。通常，分数会分开：计算一分，解释一分。
Key mark types in SQA Statistics:
SQA 统计中的关键分数类型：
- Method marks for selecting and applying the correct statistical technique
- Accuracy marks for correct numerical results and units
- Interpretation marks for linking the statistical finding to the real‑world scenario
- 方法分：选择并应用正确的统计方法
- 准确性分：正确的数值结果和单位
- 解读分：将统计结果与现实情境联系起来
2. Show Full Working – The Golden Rule | 展示完整过程——黄金法则
SQA statistics papers award marks for intermediate steps. Writing down the formula, substitution, and final value with units can make the difference between partial and full marks. Even if you use a calculator, show the expression you entered.
SQA 统计试卷会对中间步骤给分。写下公式、代入数据、带单位的最终结果，可能是部分分与满分之间的区别。即使使用计算器，也要展示输入的计算式。
xample, when calculating a standard deviation, do not simply write s = 4.67. Instead show:
例如，计算标准差时，不要只写 s = 4.67。要展示：
s = √(Σ(x - x̄)² / (n - 1)) = √87.25 / 4 ≈ 4.67
The examiner can then award a method mark even if the final number is mistyped. This approach is essential in all calculation‑based questions, including hypothesis tests and confidence intervals.
这样即使最终数字录入有误，考官也能给方法分。这种方法在所有计算题中都至关重要，包括假设检验和置信区间。
3. Effective Use of the Formulae Sheet | 有效使用公式表
SQA provides a statistical formulae sheet in the exam. Familiarise yourself with its layout before the test. Know exactly where to find the t‑test statistic, pooled variance, degrees of freedom, and critical values. Wasting time searching through pages can cost you marks on later questions.
SQA 在考试中提供统计公式表。考前要熟悉其排版。准确知道在哪里查找 t 检验统计量、合并方差、自由度和临界值。在公式表中浪费时间查找，可能会让你在后半部分题目丢分。
Do not rely on the formula sheet to teach you the meaning of symbols. While revising, annotate your own copy with notes on when each formula is used, and practise retrieving them under timed conditions. The sheet is a tool, not a textbook.
不要指望公式表来教你符号的含义。复习时，在自己的副本上标注每个公式的使用条件，并在限时条件下练习查阅。公式表是工具，不是教科书。
4. Interpreting Statistical Software Output | 解读统计软件输出
The SQA Statistics exam often includes computer output from packages like Minitab or R. You will be expected to extract test statistics, p‑values, confidence interval bounds, and regression coefficients directly from the tables, rather than calculating them from scratch.
SQA 统计考试常包含来自 Minitab 或 R 等软件的计算机输出。你需要直接从表格中提取检验统计量、p 值、置信区间界限和回归系数，而不是从零开始计算。
A typical question will give you a session window with headings such as 'Two‑Sample T‑Test and CI'. You must identify the correct row for your test (e.g. assuming equal or unequal variances), then read the t‑value, p‑value, and decision‑relevant figures. Always check the degrees of freedom and whether the test is one‑sided or two‑sided.
典型题目会给出带有“Two‑Sample T‑Test and CI”标题的会话窗口。你必须识别出正确的检验行（如假设方差相等或不等），然后读取 t 值、p 值以及决策相关数字。务必检查自由度以及检验是单侧还是双侧。
5. Hypothesis Testing – Structured Approach | 假设检验——结构化答题法
Hypothesis tests carry a significant proportion of marks. SQA examiners award marks for each key component: hypotheses, test selection, calculation, and conclusion in context. Use a clear, step‑by‑step layout.
假设检验占分值比例很大。SQA 考官对每个关键部分给分：假设、检验选择、计算和情境化结论。采用清晰的分步式排版。
For instance, when comparing two population means:
例如，比较两个总体均值时：
Step 1: State H₀: μ₁ = μ₂, H₁: μ₁ ≠ μ₂ (or one‑sided).
步骤 1: 陈述 H₀: μ₁ = μ₂, H₁: μ₁ ≠ μ₂（或单侧）。
Step 2: Choose the appropriate test, e.g. two‑sample t‑test with pooled variance. Justify briefly.
步骤 2: 选择合适的检验，如合并方差的双样本 t 检验。简要说明理由。
the test statistic using the formula, then compare it with the critical value or use the p‑value approach. If using p‑value, state p = 0.023 (or as given).
步骤 3: 用公式计算检验统计量，然后与临界值比较或使用 p 值法。如果使用 p 值，写出 p = 0.023（或给定值）。
Step 4: Make a decision: 'Since p < 0.05, we reject H₀.'
步骤 4: 做出决定：'因为 p < 0.05，我们拒绝 H₀。'
Step 5: Give a non‑statistical conclusion: 'There is sufficient evidence to suggest a difference in mean response between the two treatments.'
步骤 5: 给出非统计性结论：'有足够证据表明两种处理方式的平均响应存在差异。'
The conclusion in context is often worth 2 marks and is frequently missed. Never end with only 'reject H₀'. Always relate back to the problem statement.
语境结论通常值 2 分，且经常被遗漏。不要只以“拒绝 H₀”结束。务必联系题目表述。
6. Confidence Intervals – Precision and Interpretation | 置信区间——精度与解读
When constructing a confidence interval, always show the formula, the critical value, and the interval bounds. For a 95% CI for a mean, write:
在构建置信区间时，始终展示公式、临界值和区间界限。对于均值的 95% 置信区间，写出：
x̄ ± t* × (s / √n)
Then substitute the values. Marks are often deducted if the t* value is not reported or the degrees of freedom are missing.
然后代入数值。如果未报告 t* 值或缺少自由度，经常会被扣分。
Interpretation matters. A correct statement is: 'We are 95% confident that the true population mean weight loss lies between 2.3 kg and 4.1 kg.' Avoid incorrect phrasing such as 'There is a 95% probability that the population mean is in this interval.' The probability statement applies to the method, not the specific interval.
解读很重要。正确的表述是：“我们有 95% 的把握认为总体平均减重在 2.3 kg 到 4.1 kg 之间。”避免错误说法，如“总体均值落在这个区间的概率是 95%”。概率陈述针对的是方法，而不是特定的区间。
7. Probability and Distribution Questions | 概率与分布问题
SQA questions require you to recognise the appropriate distribution: binomial for fixed trials with success/failure, Poisson for rare events in a fixed interval, and normal for continuous symmetric data. Clearly write the distribution and its parameters before calculating probabilities.
SQA 考题要求你识别恰当的分布：固定试验次数且有成功/失败的二项分布，固定区间内稀有事件的泊松分布，以及连续对称数据的正态分布。在计算概率前先明确写出分布及其参数。
For example, if X ~ When using normal approximation to binomial, apply the continuity correction and show it explicitly. Many students lose marks by forgetting to add or subtract 0.5.
例如，若 X ~ B(20, 0.3)，写出符号并说明你需要的是 P(X = k) 还是 P(X ≤ k)。当使用正态近似二项分布时，运用连续性校正并明确展示。许多学生因忘记加减 0.5 而丢分。
When using standard normal tables, always sketch the distribution and shade the required area. This helps avoid errors with cumulative probabilities and ensures you are reading the correct table entry.
使用标准正态表时，始终画出分布并涂色标记所求区域。这有助于避免累积概率错误，并确保你读取正确的表值。
8. Linear Regression and Correlation | 线性回归与相关
In regression problems, do not simply write down the equation from software output. Interpret the coefficients: 'For every additional unit increase in x, the predicted y increases by 3.2 units, on average.' The intercept may have no practical meaning, and stating that can gain an interpretation mark.
在回归问题中，不要只从软件输出抄下方程。要解读系数：“x 每增加一个单位，预测的 y 平均增加 3.2 个单位。”截距可能没有实际意义，说明这一点也能获得解读分。
When asked to assess the model, comment on r² and residual plots. A high r² alone does not guarantee a good model; you must check for patterns in residuals, constant variance, and outliers. SQA mark schemes frequently reward comments on the limitations of prediction beyond the range of the data.
当要求评估模型时，要评论 r² 和残差图。高 r² 本身并不能保证好模型；你必须检查残差模式、方差恒定性和异常值。SQA 评分方案经常奖励关于超出数据范围预测局限性的评论。
9. Data Collection and Experimental Design | 数据收集与实验设计
Questions on sampling and design test your ability to apply statistical principles to real contexts. Be precise in your language. Instead of 'take a random sample', write 'assign each individual a number and use a random number generator to select a simple random sample of size 50'.
抽样和设计类问题考查你将统计原理应用到真实情境的能力。语言要精确。不要写“随机抽样”，而要写“为每个个体分配一个号码，并使用随机数生成器选择一个大小为 50 的简单随机样本”。
When evaluating a study, point out possible sources of bias such as volunteer samples, non‑response, or measurement error. The SQA mark scheme expects you to link the type of bias to the specific context and suggest practical improvements, e.g. using a matched pairs design to reduce variability.
在评估一项研究时，指出可能的偏倚来源，如志愿者样本、无响应或测量误差。SQA 评分方案期望你将偏倚类型与具体情境联系起来，并提出实际改进建议，例如使用配对设计以减少变异性。
10. Checking Answers and Spotting Mistakes | 检查答案与发现错误
Allocate the last 10 minutes of the exam to review your answers. First, check that all parts of each question have been attempted. Then re‑read your conclusion statements to ensure they are in context and grammatically correct. A meaningful context link can be the difference between a 1‑mark and a 2‑mark conclusion.
在考试最后 10 分钟用来检查答案。首先，检查每道题的每个部分是否都已作答。然后重读你的结论陈述，确保上下文正确、语法无误。有意义的语境联系可能是 1 分和 2 分结论之间的差别。
Verify your test statistic and p‑value are consistent. If you reject H₀, the p‑value should be less than α; if you fail to reject, it should be greater. Detectable arithmetic errors often arise from squaring negative numbers or misplacing the decimal point. Use rough estimates to sense‑check final answers.
验证检验统计量和 p 值是否一致。如果拒绝 H₀，p 值应小于 α；如果未拒绝，应大于 α。常见的算术错误来自负数平方或小数点错位。使用粗略估算来检验最终答案是否合理。
11. Time Management and Paper Strategy | 时间管理与考试策略
The SQA Statistics paper is designed to be completed within the allocated time if you work strategically. A sensible rule is 1 minute per mark, plus some reading time. If you are stuck on a 3‑mark subsection for more than 4 minutes, move on and return later if time allows.
如果策略得当，SQA 统计试卷可以在规定时间内完成。一条明智的规则是每分钟 1 分，外加一些阅读时间。如果你在一道 3 分的小题上卡住超过 4 分钟，就继续前进，如果时间允许再回来。
Begin by scanning the whole paper to identify easier questions or familiar contexts. Tackle high‑confidence sections first to secure marks and build momentum. The final questions often integrate several topics; leave sufficient time for these because they typically reward deeper understanding with more marks.
先浏览整份试卷，找出较简单的题目或熟悉的背景。先做有把握的部分，确保得分并建立答题势头。最后的题目往往综合多个主题；要给这些题留出足够时间，因为它们通常奖励更深度的理解，分值也更高。
12. Analysing Real Marking Instructions – A Case Study | 评分标准实例解析
Consider a typical SQA question: 'A random sample of 12 bottles from machine A and 15 bottles from machine B gave mean fill volumes of 502.3 ml and 498.7 ml respectively, with standard deviations of 3.1 ml and 2.8 ml. Test whether the machines produce different mean fills, assuming equal variances.'
考虑一道典型的 SQA 题目：“从机器 A 随机抽取 12 瓶，从机器 B 抽取 15 瓶，平均填充量分别为 502.3 ml 和 498.7 ml，标准差为 3.1 ml 和 2.8 ml。假设方差相等，检验两台机器的平均填充量是否不同。”
The mark breakdown might look like this (total 12 marks):
评分分配可能如下（共 12 分）：
- H₀: μ₁ = μ₂, H₁: μ₁ ≠ μ₂ (2 marks) – one for each hypothesis stated in symbols and words.
- Test identified as two‑sample t‑test with pooled variance (1 mark).
- ariance sₚ² = [(n₁-1)s₁² + (n₂-1)s₂²] / (n₁ + n₂ - 2) written with substitution (2 marks).
- Correct calculation of sₚ² ≈ 8.68 (1 mark).
- Test statistic t = (x̄₁ - x̄₂) / [sₚ√1/n₁ + 1/n₂] = 3.6 / (2.946 × 0.399) ≈ 3.07 (2 marks).
- Degrees of freedom = 25, critical value t₀.₀₂₅,₂₅ = 2.060 (or p‑value ≈ 0.005) (1 mark).
- Decision: 3.07 > 2.060, so reject H₀ (1 mark).
- Conclusion in context: average fills are significantly different; likely machine A overfills compared to machine B (2 marks).
- H₀: μ₁ = μ₂, H₁: μ₁ ≠ μ₂ (2分) – 每个假设用符号和文字陈述得 1 分。
- 检验确定为合并方差的双样本 t 检验 (1分)。
- 写出合并方差公式 sₚ² = [(n₁-1)s₁² + (n₂-1)s₂²] / (n₁ + n₂ - 2) 并代入 (2分)。
- 正确计算 sₚ² ≈ 8.68 (1分)。
- 检验统计量 t = (x̄₁ - x̄₂) / [sₚ√1/n₁ + 1/n₂] = 3.6 / (2.946 × 0.399) ≈ 3.07 (2分)。
- 自由度 = 25，临界值 t₀.₀₂₅,₂₅ = 2.060 (或 p‑值 ≈ 0.005) (1分)。
- 决策: 3.07 > 2.060，因此拒绝 H₀ (1分)。
- 语境结论: 平均填充量存在显著差异，机器 A 可能比机器 B 多装 (2分)。
A student who omitted the substitution step would lose those marks even with a correct final t‑value. Likewise, a conclusion that just says 'reject H₀' would drop 2 marks. This case study shows why thorough, structured answers are vital.
一位考生若遗漏了代入步骤，即使最终 t 值正确也会丢分。同样，只写“拒绝 H₀”的结论会丢掉 2 分。这一案例说明为何完整、结构化的答案至关重要。
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