📚 Common Misconceptions in SQA Year 13 Statistics and How to Correct Them | SQA 13年级统计常见误区与纠正方法
In Advanced Higher Statistics, students often carry forward subtle misunderstandings that can undermine their analytical reasoning and exam performance. This article highlights the most prevalent misconceptions encountered in SQA Year 13 Statistics and provides clear, targeted corrections. By addressing these pitfalls, you can strengthen your statistical literacy and approach both written papers and project work with greater confidence.
在高级统计课程中，学生常常会延续一些不易察觉的误解，这些误解会削弱他们的分析推理能力和考试表现。本文指出了 SQA 13年级统计中最常见的误区，并提供了清晰、有针对性的纠正方法。通过解决这些陷阱，你可以加强统计素养，更自信地应对书面考卷和项目作业。
1. Confusing Population Parameters with Sample Statistics | 混淆总体参数与样本统计量
Many learners treat the sample mean x̄ as if it were the true population mean μ. They write x̄ = μ without acknowledging sampling variability.
许多学习者将样本均值 x̄ 当作真实的总体均值 μ 来对待，在不承认抽样变异性的情况下写出 x̄ = μ。
The sample mean is an estimate computed from a random sample. Different samples yield different x̄ values. We use the sample statistic to infer about the parameter, but never equate them with certainty. The uncertainty is expressed through standard error s / √n and confidence intervals.
样本均值是根据随机样本计算得出的估计值。不同的样本会产生不同的 x̄ 值。我们使用样本统计量来推断参数，但绝不能将它们确定地等同起来。这种不确定性通过标准误差 s / √n 和置信区间来表达。
A common side-effect is writing a hypothesis as H₀: x̄ emember that hypotheses are statements about population parameters, not sample statistics.
一个常见的连带错误是将假设写成 H₀: x̄ = μ₀ 而不是 H₀: μ = μ₀。请记住，假设是关于总体参数的陈述，而不是关于样本统计量的。
2. Misinterpreting the p-value | 误解 p 值
Students frequently believe the p-value is the probability that the null hypothesis is true. This is incorrect.
学生经常认为 p 值是原假设为真的概率。这是不正确的。
The p-value is the probability of obtaining a test statistic at least as extreme as the one observed, assuming the null hypothesis is true. It does not give the probability that H₀ is correct, nor does it measure the size of an effect.
p 值是在原假设为真的条件下，获得一个与观测值同样极端或更极端的检验统计量的概率。它既不给出 H₀ 为真的概率，也不衡量效应的大小。
Correct interpretation: A small p-value (e.g., p < 0.05) indicates that such an extreme result would be unlikely if H₀ were true, leading us to question H₀. A large p-value simply means the data are consistent with H₀, not that H₀ has been proven.
正确的解释：一个小的 p 值（例如 p < 0.05）表明，如果 H₀ 为真，这样极端的结果是不太可能出现的，这促使我们质疑 H₀。一个大的 p 值仅仅意味着数据与 H₀ 一致，而不是证明了 H₀。
In SQA exams, you must describe the p-value in context and compare it with the significance level α, stating whether there is evidence against H₀.
在 SQA 考试中，你必须结合具体情境描述 p 值，并将其与显著性水平 α 进行比较，说明是否有证据否定 H₀。
3. Confidence Interval Misunderstandings | 置信区间的常见误解
A typical error is saying 'There is a 95% probability that the population mean lies in the computed 95% confidence interval.'
一个典型错误是说“总体均值有 95% 的概率落在计算出的 95% 置信区间内”。
Once the interval is calculated, the parameter either is or is not inside it. The 95% confidence refers to the long-run proportion of intervals, generated by the same method, that would capture the true parameter. For a single realised interval, we say we are 95% confident that the interval contains μ, which means the method works 95% of the time.
一旦区间被计算出来，参数要么在里面，要么不在里面。95% 置信度是指在长期重复抽样下，用相同方法构建的区间中有 95% 会包含真实参数。对于单个已实现的区间，我们说我们有 95% 的信心认为该区间包含 μ，这意味着该方法在 95% 的情况下是有效的。
Also, students often confuse the confidence level with the interval width. A higher confidence level (99% vs 95%) produces a wider interval, reflecting greater certainty at the expense of precision.
此外，学生常常混淆置信水平与区间宽度。更高的置信水平（99% 对比 95%）会产生更宽的区间，以牺牲精确度为代价来获得更大的确定性。
For a mean with known σ, the interval is:
x̄ ± z* × σ / √n
Ensure you interpret both the point estimate and the margin of error correctly.
确保你正确解释点估计值和误差范围。
4. Correlation Versus Causation | 相关关系与因果关系的混淆
Seeing a high Pearson correlation coefficient r does not mean one variable causes the other to change. Many students leap to causal conclusions.
看到一个高的皮尔逊相关系数 r 并不意味着一个变量导致另一个变量变化。许多学生会贸然得出因果结论。
Correlation indicates an association and the strength of a linear relationship. Causation requires a controlled experiment where confounding variables are eliminated. A strong correlation could arise from a lurking variable, reverse causation, or pure coincidence.
相关关系表示的是关联性和线性关系的强度。因果关系需要对照实验来消除混杂变量。强相关可能源自潜在变量、反向因果或纯粹巧合。
In SQA project work, you must comment on whether an observed association might be causal, but always use cautious language: 'The data suggest a positive linear association. This does not necessarily imply that A causes B.'
在 SQA 项目工作中，你必须评论观察到的关联是否可能是因果关系，但要始终使用谨慎的语言：“数据表明存在正向线性关联。这并不一定意味着 A 导致 B。”
A classic example is the strong positive correlation between ice cream sales and drowning incidents. The lurking variable is warm weather, which increases both.
一个经典例子是冰淇淋销量与溺水事件之间的强正相关。潜在的变量是温暖的天气，它同时导致了两者的增加。
5. Assumptions of Normality | 正态性假设的误用
Students often apply z-tests or t-tests blindly without checking whether the underlying population is approximately normal or whether the sample size is large enough for the Central Limit Theorem (CLT).
学生常常盲目地应用 z 检验或 t 检验，而不检查基础总体是否近似正态，或样本量是否足够大以应用中心极限定理 (CLT)。
if n is large (usually n ≥ 30 for the CLT to take effect on the sampling distribution of x̄). For small samples from a non-normal population, consider non‑parametric alternatives such as the Wilcoxon signed‑rank test.
当总体不是正态分布时，如果 n 足够大（通常 n ≥ 30，此时 CLT 对 x̄ 的抽样分布起作用），t 检验可能仍然有效。对于来自非正态总体的小样本，可以考虑非参数替代方法，例如威尔科克森符号秩检验。
For the chi‑squared test and ANOVA, normality of the errors or subgroups must be checked, but students frequently skip this. Always discuss assumptions in any hypothesis‑testing question and back them up with graphical checks like a histogram or normal probability plot.
对于卡方检验和方差分析，必须检查误差或子组的正态性，但学生常常忽略这一点。在任何假设检验问题中都要讨论假设，并通过直方图或正态概率图等图形检查来支持这些假设。
The assumption of equal variances in two‑sample t‑tests is another area of confusion. If an F‑test or Levene’s test shows unequal variances, use Welch’s t‑test, which does not assume equal variances.
双样本 t 检验中的方差齐性假设是另一个容易混淆的领域。如果 F 检验或 Levene 检验显示方差不相等，则应使用 Welch t 检验，它不假定方差相等。
6. Type I and Type II Errors | 第一类错误与第二类错误
Many students cannot distinguish between the two error types in context. They might say 'The probability of making a Type I error is β' or confuse the consequences.
许多学生无法在具体情境中区分两类错误。他们可能会说“犯第一类错误的概率是 β”，或者混淆它们的后果。
A Type I error occurs when H₀ is true but we reject it. Its probability is α (the significance level). A Type II error happens when H₀ is false but we fail to reject it; its probability is β. The power of a test is 1 − β, the probability of correctly rejecting a false H₀.
第一类错误发生在 H₀ 为真却被拒绝时，其概率为 α（显著性水平）。第二类错误发生在 H₀ 为假却未能拒绝时，其概率为 β。检验的功效为 1 − β，即正确拒绝一个错误 H₀ 的概率。
In practice, reducing α also reduces the chance of a Type I error but increases the chance of a Type II error (holding sample size constant). Only larger sample sizes can simultaneously reduce both error probabilities.
在实践中，降低 α 会减少第一类错误的概率，但会增加第二类错误的概率（在样本量不变的情况下）。只有更大量的样本才能同时减少两种错误的概率。
In exam questions, you might be asked to identify which error is more serious in a given real‑world scenario, such as a medical trial. Be prepared to justify your choice.
在考试题目中，你可能需要根据给定的现实情境，例如医学试验，确定哪种错误的后果更严重。准备好为你的选择提供理由。
7. Extrapolation in Regression | 回归分析中的外推问题
A very common flaw in project work is using a fitted regression line to predict response values for x‑values far outside the range of the original data.
项目作业中一个非常常见的缺陷是使用拟合的回归线来预测远超出原始数据范围的 x 值对应的响应值。
The least‑squares line ŷ bserved range of the explanatory variable. Outside that range, the linear relationship may not hold, leading to grossly inaccurate predictions. Always state that extrapolation is unreliable and should be avoided.
最小二乘线 ŷ = b₀ + b₁ x 仅在解释变量的观测范围内是可靠的。超出该范围，线性关系可能不成立，从而导致极其不准确的预测。务必说明外推是不可靠的，应当避免。
Furthermore, a high R² does not make extrapolation safe. Even if the line fits the observed points well, the underlying pattern could change. The SQA expects you to restrict predictions to the domain of the data or to highlight the uncertainty when extrapolating is necessary.
此外，高 R² 值并不能令外推变得安全。即使回归线很好地拟合了观测点，潜在的模式也可能发生变化。SQA 期望你将预测限制在数据所在的范围之内，或者在必须外推时强调其中的不确定性。
8. The Gambler's Fallacy and Independence | 赌徒谬误与独立性的误解
When studying probability, many students believe that after a run of heads in coin tosses, tails becomes 'due'. This is the gambler's fallacy, a misunderstanding of independence.
在学习概率时，许多学生认为在连续多次抛出硬币正面之后，反面就“应该出现了”。这就是赌徒谬误，是对独立性的一种误解。
If events are independent, past outcomes do not influence future ones. The probability of a head remains ½ every toss, regardless of previous results. This fallacy often appears in probability tree questions or in simulations.
如果事件是独立的，过去的结果不会影响未来的结果。每次抛掷出现正面的概率始终是 ½，与之前的结果无关。这种谬误经常出现在概率树题目或模拟中。
In statistics, independence is a crucial assumption for many tests (e.g., t‑test, chi‑squared test). Data collected over time or within clusters can violate independence. Always assess whether observations influence each other. If in doubt, consider paired tests or repeated‑measures approaches.
在统计学中，独立性是许多检验（例如 t 检验、卡方检验）的关键假设。按时间或按群组收集的数据可能违反独立性。务必评估观测值之间是否相互影响。如果有疑问，请考虑配对检验或重复测量方法。
9. Ignoring the Effect of Sample Size | 忽视样本量的影响
Some learners think that a statistically significant result implies a large or practically important effect. They ignore the fact that with a very large sample, even trivially small differences can become 'significant'.
一些学习者认为，统计上显著的结果意味着效应量很大或者具有实际重要性。他们忽视了这样一个事实：在样本量非常大的情况下，即使是微不足道的微小差异也可能变得“显著”。
Statistical significance (p < 0.05) simply suggests the observed effect is unlikely to be due to random chance under H₀. It does not indicate the magnitude. Always report effect size (e.g., Cohen’s d for means, or the correlation coefficient r) together with the p-value and confidence interval.
统计显著性（p < 0.05）仅仅表明，在 H₀ 下，观察到的效应不太可能是由随机偶然性引起的。它并不表示效应的大小。务必同时报告效应量（例如用于均值的 Cohen’s d, 或相关系数 r）以及 p 值和置信区间。
Conversely, with a very small sample, an important effect may fail to be detected due to low power. A non‑significant result does not prove absence of an effect; it may mean the study lacked the power to detect it.
相反，在样本量非常小的情况下，一个重要的效应可能因为统计功效低而未被检测到。不显著的结果并不能证明效应不存在；这或许意味着研究缺乏检测该效应的功效。
When planning a study, power analysis is essential to determine the appropriate sample size. In SQA questions, you may need to comment on why a study might have failed to find significance despite an apparent trend.
在规划一项研究时，功效分析对于确定适当的样本量至关重要。在 SQA 题目中，你可能需要评论为什么一项研究尽管有显著趋势，却未能发现统计显著性。
10. Misapplication of Chi-Squared Tests | 卡方检验的误用
The chi‑squared test for independence is often applied to tables where expected frequencies are too low, or the data are not counts.
独立性卡方检验经常被应用于期望频数过低的表格，或者应用于非计数数据。
The chi‑squared test requires that expected frequencies in each cell are at least 5 for the approximation to the χ² distribution to be valid. If more than 20% of expected frequencies fall below 5, you should combine categories or use Fisher’s exact test. The test also assumes observations are independent and that data are frequency counts.
卡方检验要求每个单元格的期望频数至少为 5，这样近似 χ² 分布才有效。如果超过 20% 的期望频数低于 5，你应当合并类别或使用 Fisher 精确检验。该检验还假设观测值相互独立，且数据为频数计数。
Students often confuse the chi‑squared test of independence with the chi‑squared goodness‑of‑fit test. The former tests whether two categorical variables are associated, while the latter tests whether observed frequency counts match a theoretical distribution.
学生常常混淆独立性卡方检验和卡方拟合优度检验。前者检验两个分类变量是否有关联，而后者检验观测频数计数是否符合某个理论分布。
In both cases, the test statistic is calculated as:
χ² = Σ (O − E)² / E
he χ² distribution appropriately. For independence in an r × c table, df = (r − 1)(c − 1).
务必列出自由度，并适当地查阅 χ² 分布。对于 r × c 表格中的独立性检验，自由度 df = (r − 1)(c − 1)。
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