📚 Exam Technique and Marking Criteria for SQA Statistics | SQA 统计：答题技巧与评分标准
Statistics papers in the SQA National 5 course test your ability to interpret data, calculate probabilities, and draw conclusions from statistical evidence. However, many marks are lost not because of a misunderstanding of the content, but due to poor exam technique or a failure to follow the marking criteria precisely. This guide will show you how to structure your answers, avoid common pitfalls, and maximise every available mark by understanding exactly what the marker is looking for.
SQA 国家5级统计试卷考察你解读数据、计算概率以及从统计证据中得出结论的能力。然而，许多失分并非因为对内容的不理解，而是因为答题技巧欠佳或未能严格遵循评分标准。本指南将为你展示如何组织答案、避开常见陷阱，并通过精准理解阅卷官的期望来最大化每一分。
1. Reading the Question with Statistical Precision | 精确阅读统计问题
Before you write a single number, highlight the command words – 'calculate', 'compare', 'justify', 'determine whether' – because each triggers a different expectation in the marking scheme. A 'compare' question, for example, requires at least two linked statements, not a solitary observation. Look for the specific statistical measure requested: mean, median, interquartile range, probability, or standard deviation, and note the required number of decimal places or significant figures.
在你写下任何一个数字之前，先圈出指令词——'calculate' (计算)、'compare' (比较)、'justify' (论证)、'determine whether' (判断是否)——因为每一个词在评分标准中都会引发不同的期望。例如，一道 'compare' 题要求至少两个相关联的陈述，而非孤立的一个观察。看清题目要求的特定统计量：平均值、中位数、四分位距、概率还是标准差，并注意要求的小数位数或有效数字位数。
2. Making Orderly Working Your Priority | 让条理清晰的解题步骤成为优先
The SQA marking scheme awards marks for 'method' and 'process' just as generously as for the final answer. If you calculate a mean, write down the sum of the values, show the division by n, and then give the result. A missing bracket or a skipped substitution step can cost a process mark. For a standard deviation, show the calculation of the mean first, then the squared deviations, then the sum, then the division by (n – 1), and finally the square root. Even if your final figure is wrong, you can still collect most of the marks.
SQA 评分标准对"方法"和"过程"的给分与最终答案的给分同样慷慨。如果你计算平均值，要写下数值的总和，展示除以 n 的过程，再给出结果。遗漏一个括号或跳过一个代入步骤都会导致过程分丢失。对于标准差，先写出平均值的计算，再写出离差平方、求和、除以 (n – 1)、最后开平方。即使最终数字有误，你仍能拿到大部分分数。
3. Probability Answers That Satisfy the Marking Criteria | 满足评分标准的概率答案
at least two decimal places unless the paper specifies otherwise. When a question asks for 'the probability that the event does not occur', compute 1 – early. For combined events, always state the multiplication or addition rule you are using – markers look for 'P(A and B) = P(A) × P(B)' or 'P(A or B) = P(A) + P(B) – P(A and B)' ltiply along branches; if you use a Venn diagram, define the sets and show the arithmetic.
除非试卷另有说明，概率必须表示为最简分数或至少保留两位小数的数字。当题目要求"事件不发生的概率"时，要计算 1 – P(事件) 并清楚地陈述。对于组合事件，始终在解题步骤中明确写出所使用的乘法或加法法则——阅卷官期望看到 'P(A and B) = P(A) × P(B)' 或 'P(A or B) = P(A) + P(B) – P(A and B)' 被清晰地写出。如果你使用树形图，标出所有分支的概率并沿分支相乘；如果你使用韦恩图，要定义集合并展示运算过程。
4. Comparing Data Sets Using Proper Statistical Language | 用恰当的统计语言比较数据集
A comparison answer that reads 'the median is higher' gains only one mark if no data values are referenced. A full-mark answer always mentions the measure of central tendency (mean or median) together with its numerical value for both data sets, then compares them in context – 'The median waiting time for Hospital A was 18 minutes, while for Hospital B it was 24 minutes, meaning that on average patients at B waited 6 minutes longer.' Then you must comment on the spread using range or interquartile range (IQR), again quoting figures and interpreting them. The phrase 'on average' must appear when you use the mean or median, and 'consistent' or 'more varied' when describing spread.
一个写成"中位数更高"的比较答案，如果没有引用具体数值，只能得到一分。一个满分答案总是先提及集中趋势的量数（平均数或中位数）连同两个数据集的数值，然后在上下文中进行比较——"甲医院候诊时间的中位数为 18 分钟，而乙医院为 24 分钟，这意味着乙医院的患者平均多等 6 分钟。"然后你必须用极差或四分位距 (IQR) 对离散程度进行评论，同样要引用数字并加以解释。在使用平均数或中位数时必须出现 'on average'，在描述离散程度时使用 'consistent' 或 'more varied'。
5. Tackling Box Plots and Cumulative Frequency Diagrams | 应对箱形图和累积频数图
When you are asked to draw a box plot, the marking scheme allocates marks for the median, quartiles, and extreme values all being correctly plotted and aligned on a scaled axis. Draw whiskers with a ruler, label the five-key values clearly, and never extend whiskers beyond the minimum or maximum data values. In a cumulative frequency diagram, marks are awarded for accurate plotting of points at the upper class boundary, for joining points with a smooth curve or straight lines as instructed, and for reading the median and quartiles correctly. When finding the median from a cumulative frequency graph, show a clear line from the half-total frequency across to the curve and down to the axis.
当你被要求绘制箱形图时，评分标准给分的点在于：中位数、四分位数和极值全部正确绘制并且对齐在有刻度的数轴上。用直尺绘制须线，清楚地标注五个关键值，而且决不能让须线超出数据的最小值或最大值。在累积频数图中，分数会分配给：在上组界精确描点，按照题目要求用平滑曲线或直线连接各点，以及正确读取中位数和四分位数。当从累积频数图中查找中位数时，要展示一条清晰的直线，从半数总频数水平引向曲线，再向下引到数轴。
6. Interpreting Correlation and Lines of Best Fit | 解读相关性与最佳拟合线
For a scatter graph question, start by describing the correlation as 'positive', 'negative', or 'no correlation', and then qualify its strength as 'strong', 'moderate', or 'weak' with reference to how close the points lie to a straight line. If you are asked to draw a line of best fit, it must have roughly equal numbers of points on either side and pass through the mean point (x̄, ȳ) unless told otherwise. When using the line to estimate a value, always show the construction lines on the graph – a vertical line from the given x to the best-fit line and then a horizontal line to the y-axis, or vice versa. Explicitly state whether your answer is an interpolation or an extrapolation, as extrapolation is less reliable.
对于散点图问题，首先将相关性描述为 'positive' (正)、'negative' (负) 或 'no correlation' (无相关)，然后根据数据点接近一条直线的程度限定其强度为 'strong' (强)、'moderate' (中等) 或 'weak' (弱)。如果要求你绘制最佳拟合线，它必须使线两侧的点数大致相等，并且除非另有说明，应通过平均值点 (x̄, ȳ)。当利用该直线估计数值时，始终在图上展示作图线——从给定的 x 值作垂线与最佳拟合线相交，再作水平线与 y 轴相交，或反之。明确说明你的答案是内插值还是外推值，因为外推法的可靠性较低。
7. Stem-and-Leaf and Dot Plot Perfection | 茎叶图与点图的完美呈现
In a stem-and-leaf diagram, order the leaves from smallest to largest and add a key that explains the place value, e.g. '3 | 7 means 37'. Both stems and leaves must be aligned, and a missing key loses a communication mark immediately. For back-to-back stem-and-leaf diagrams, the leaves on the left-hand side also increase from the stem outwards, so write them in reverse order. For a dot plot, use evenly spaced scales and only one dot per observation; avoid turning it into a bar chart. When asked to find the mode or median from these diagrams, state the values clearly.
在茎叶图中，将叶从最小到最大排序，并添加一个解释位值的图例，例如 '3 | 7 表示 37'。茎和叶都必须对齐，任何漏掉的图例会立刻扣掉表达分。对于背靠背茎叶图，左侧的叶也是从茎部向外增加，因此要按逆序书写。对于点图，使用等间距的尺度，每个观测值只用一个点表示；不要将其画成条形图。当要求你从这些图中找出众数或中位数时，要清楚地陈述数值。
8. The Critical Importance of Units and Context | 单位与上下文的至关重要性
Every mark scheme for SQA Statistics contains a specific note: 'units must be stated where appropriate'. If the question involves pounds, seconds, kilograms, or points, your final answer is not complete without the unit. Even an intermediate calculation that is a count of people should say '12 people' rather than just '12'. In interpretation questions, link your statistical conclusion back to the context using the exact wording from the problem – 'the probability that a randomly selected pupil walks to school is 0.35' rather than 'the probability is 0.35'.
每份 SQA 统计评分标准都有一条专门的注释：'适当处必须注明单位'。如果题目涉及英镑、秒、千克或分数，你的最终答案若不包含单位就是不完整的。即便是一个表示人数的中间计算步骤，也应写成 '12 people' 而不是区区 '12'。在解读类题目中，运用题目中的确切措辞将你的统计结论与上下文联系起来——'随机抽取一名学生步行上学的概率为 0.35' 而非 '概率是 0.35'。
9. Mastering Sampling Methods and Bias Explanations | 掌握抽样方法和对偏差的解释
When asked to name a sampling method, choose from 'simple random', 'stratified', 'systematic', or 'opportunity' (or 'convenience') based on the description provided. Then, you must justify why that method is suitable in the given context – for instance, 'stratified sampling guarantees that each year group is represented in proportion to its size, making the results more reflective of the whole school.' When identifying bias, avoid vague statements; specify which group is over-represented or under-represented and how that skews the results: 'An opportunity sample taken outside the library in the morning is likely to over-represent studious students and under-represent those who arrive later, biasing the estimate of average study time upwards.'
当被要求说出抽样方法的名称时，根据提供的描述从 'simple random' (简单随机)、'stratified' (分层)、'systematic' (系统) 或 'opportunity' (即 'convenience'，便利) 中选择一种。然后，你必须论证该方法在给定情境下为何合适——例如，'分层抽样确保每个年级按其人数比例被代表，使结果更能反映全校情况。' 在识别偏差时，避免含糊其辞的陈述；指明哪个群体被过度代表或代表不足，以及这如何扭曲了结果：'上午在图书馆外抽取的便利样本很可能过度代表勤奋的学生，而对到校较晚的学生代表不足，从而将平均学习时间的估计值向上偏倚。'
10. Understanding the 'Evaluate' and 'Justify' Command Words | 理解 'Evaluate' 与 'Justify' 指令词
An 'evaluate' or 'justify' question expects you to weigh up evidence and give a supported decision. Use a two-step approach: first, state your conclusion ('I agree because...' or 'New Flame's claim is not supported because...'), then provide at least two distinct pieces of statistical evidence with figures. For example, 'The claim that the new packaging increased sales is justified because median daily sales rose from 45 to 62, and the interquartile range decreased from 13 to 8, indicating more consistent sales.' Markers are trained to look for the use of comparative language and precise values.
一道 'evaluate' 或 'justify' 题期望你去权衡证据，并给出一个有依据的决定。采用两步法：首先，陈述你的结论（'我同意是因为...' 或 '新火焰的说法未被支持是因为...'），然后提供至少两个不同的带有数字的统计证据。例如，'新包装增加了销量这一说法是合理的，因为日均销量的中位数从 45 上升到了 62，同时四分位距从 13 缩小到了 8，表明销量更加稳定。' 阅卷官受过训练，会寻找比较性的语言和精确的数值。
11. Checking Rounding and Accuracy Rules | 检查舍入和精度规则
In SQA Statistics, money answers must be rounded to two decimal places unless pence is not applicable. Probabilities should be given as exact fractions or decimals to at least two decimal places; rounding to one decimal place is accepted only when the subsequent digit is 0. For intermediate steps in a standard deviation calculation, keep at least four decimal places or the full calculator value to avoid premature rounding errors. If the question states 'give your answer to three significant figures', apply this only to the final answer, not to each step, and ensure you know how to count significant figures starting from the first non-zero digit.
在 SQA 统计中，涉及货币的答案必须四舍五入到两位小数，除非便士并不适用。概率应以精确分数或至少两位小数的数字给出；仅当后续数字为 0 时，才接受舍入到一位小数。在标准差计算的中间步骤中，至少保留四位小数或使用完整的计算器数值，以避免提前舍入带来的误差。如果题目要求 '将你的答案保留三位有效数字'，仅将此应用于最终答案而非每一步骤，并且要确保你能从首个非零数字开始数有效数字。
12. Tackling Hypothesis Tests and Drawing Conclusions | 应对假设检验与得出结论
ve hypothesis H₁: p <, >, or ≠ value. Then calculate the test statistic, find the p-value from tables or a given distribution, and compare it with the significance level (commonly 5%). Your conclusion must be in plain English and directly address the context: 'Since the p‑ upport the claim that the new fertiliser increases the proportion of seeds that germinate.' Never write 'accept H₀'; only 'do not reject'.
当一道题目要求你用二项分布模型或一个比率来检验某个论断时，你必须先陈述原假设 H₀: p = 数值和备择假设 H₁: p <、> 或 ≠ 数值。然后计算检验统计量，从表格或给定的分布中找出 p 值，并将其与显著性水平（通常为 5%）进行比较。你的结论必须用通俗易懂的英文陈述并直接回扣上下文：'由于 p 值 0.032 小于 0.05，我们拒绝 H₀，并得出结论：有充分证据支持新化肥能提高种子发芽率这一论断。' 永远不要写 '接受 H₀'，只能写 '不拒绝'。
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