📚 Edexcel Mathematics: Common Mistakes in Statistical Analysis for Psychology | Edexcel 数学：心理学统计分析易错点总结
Many Edexcel Mathematics students encounter statistical methods that are frequently applied in psychology and social sciences. While the underlying mathematics is clear, common misinterpretations can lead to serious errors in data analysis and conclusions. This article highlights those pitfalls, helping you refine your understanding for both your Edexcel exams and future research.
许多学习 Edexcel 数学的同学都会接触到心理学和社会科学中常用的统计方法。虽然基本原理明确，但对结果的常见误读却会造成数据分析与结论推断上的严重错误。本文重点总结这些易错点，帮助你在 Edexcel 考试和未来的研究中更精准地运用统计知识。
1. Understanding Null and Alternative Hypotheses | 零假设与备择假设的理解误区
One of the most frequent errors is formulating hypotheses incorrectly. The null hypothesis (H₀) should always state that there is no effect or no difference, while the alternative hypothesis (H₁) represents the research prediction. Students often reverse them or write a directional alternative when a non-directional one is appropriate.
最常见的错误之一是假设表述不当。零假设 (H₀) 应始终陈述没有效应或没有差异，而备择假设 (H₁) 才代表研究预测。学生经常把它们写反，或在本该使用非定向备择假设时错误地使用了定向表述。
For example, in a study comparing memory recall for words and images, H₀: μ_words = μ_images; H₁: μ_words ≠ μ_images. Stating H₁ as ‘words are better’ without justification introduces bias.
例如，在比较文字与图像记忆回忆的研究中，H₀：μ_words = μ_images；H₁：μ_words ≠ μ_images。若无依据将 H₁ 写成“文字记忆更好”，就引入了偏差。
2. Type I and Type II Errors | I 型与 II 型错误混淆
Students frequently confuse Type I and Type II errors. A Type I error occurs when the null hypothesis is incorrectly rejected (a false positive), while a Type II error is failing to reject a false null hypothesis (a false negative). Remembering the phrase ‘Type I, you saw a difference that was not there; Type II, you missed a difference that was there’ may help.
学生经常混淆 I 型错误与 II 型错误。I 型错误发生在错误地拒绝了真的零假设（假阳性），而 II 型错误则是未能拒绝一个错误的零假设（假阴性）。记住“I 型：无中生有；II 型：视而不见”或许有帮助。
In psychological testing, setting the significance level at 0.01 instead of 0.05 reduces the risk of a Type I error but increases the risk of a Type II error. Understanding this trade-off is crucial.
在心理学测验中，将显著性水平设为 0.01 而非 0.05 可以降低 I 型错误的风险，但会增加 II 型错误的风险。理解这种权衡至关重要。
3. Misinterpretation of p-values | p 值的常见误解
A p-value of 0.03 does not mean there is a 3% chance that the null hypothesis is true. It indicates that if the null hypothesis were true, the probability of obtaining a test statistic at least as extreme as the observed one is 3%. Many students incorrectly interpret it as the probability that the alternative hypothesis is false.
p 值为 0.03 并不表示零假设为真的概率是 3%。它表示如果零假设为真，得到当前乃至更极端检验统计量的概率是 3%。很多学生错误地将其解释为备择假设不成立的概率。
Another common mistake is treating a p-value just above 0.05 (e.g., p = 0.051) as ‘almost significant’ and drawing conclusions as if it were significant. Avoid this in Edexcel examinations.
另一个常见错误是把略高于 0.05 的 p 值（如 p=0.051）当作“几乎显著”，并做出好像结果显著的结论。在 Edexcel 考试中要避免这种处理。
4. Choosing the Correct Statistical Test | 正确选择统计检验方法
Psychology-related problems often require choosing an appropriate test: parametric tests like the t-test assume normal distribution and interval data, while non-parametric alternatives like Mann-Whitney U or Wilcoxon signed-rank test are used when assumptions are violated. Students may apply a t-test on ordinal data, leading to invalid results.
心理学相关问题常常需要选择合适的检验：参数检验如 t 检验假设数据呈正态分布且为等距数据，而当假设不满足时应使用非参数替代方法，如曼-惠特尼 U 检验或威尔科克森符号秩检验。学生可能对顺序数据使用 t 检验，导致结果无效。
Additionally, for correlation, Pearson’s r requires both variables to be normally distributed; otherwise, Spearman’s rho should be used. Always check the level of measurement and distribution before deciding.
此外，对于相关性，皮尔逊相关系数 r 要求两变量均呈正态分布；否则应使用斯皮尔曼秩相关系数。决定检验方法前务必先检查测量尺度和分布形态。
5. Normal Distribution Assumptions | 正态分布假设的误用
Assuming data is normally distributed without checking can invalidate an entire analysis. In psychology, reaction times, for instance, are often positively skewed. Using a z-test directly on such data is a classic mistake. Students should first examine histograms, Q-Q plots, or perform a Shapiro-Wilk test.
不经验证就假设数据服从正态分布会使整个分析失效。在心理学中，比如反应时间通常是正偏态分布。对此类数据直接使用 z 检验是经典错误。学生应首先观察直方图、Q-Q 图或进行夏皮罗-威尔克检验。
Remember the Central Limit Theorem: for large samples (n ≥ 30), the sampling distribution of the mean is approximately normal even if the original data is not, but this applies to the distribution of sample means, not individual data points.
记住中心极限定理：对于大样本（n ≥ 30），即使原始数据不服从正态分布，样本均值的抽样分布也近似正态，但这适用于样本均值的分布，而非个体数据点。
6. Correlation and Causation | 相关关系和因果关系的混淆
A strong correlation between two variables does not imply that one causes the other. In Edexcel exam questions on scatter graphs and regression, students often erroneously conclude causality. A correlation coefficient r = 0.9 between hours of sleep and test scores does not mean that sleeping longer improves performance directly; confounders like study time may play a role.
两个变量间存在强相关并不意味着一个导致另一个。在 Edexcel 有关散点图和回归的题目中，学生常错误地得出因果关系结论。睡眠时间与考试成绩的相关系数 r = 0.9 不代表睡得更久直接提高成绩；学习时间等混杂因素也可能参与其中。
Always state that correlation indicates a linear association and that further experimental manipulation is required to establish causation.
始终应说明相关只表明线性关联，确立因果关系还需进一步的实验操控。
7. Linear Regression Pitfalls | 线性回归的陷阱
Extrapolating beyond the range of the observed data is a frequent mistake. The regression line y = a + bx is only valid within the range of x values used to calculate it. Predicting behavior for extreme values can yield absurd outcomes.
超出观测数据范围进行外推是常见错误。回归直线 y = a + bx 仅在用于计算它的 x 取值范围之内有效。对极端值进行预测可能得出荒谬的结果。
Also, students sometimes forget that a regression equation is the line of best fit that minimizes squared residuals. They might misinterpret the slope b or intercept a without context.
同时，学生有时忘记了回归方程是通过最小化残差平方和得到的最佳拟合线。他们可能会脱离背景错误解读斜率 b 或截距 a。
y = a + bx, where b = Σ((xᵢ – x̄)(yᵢ – ȳ)) / Σ((xᵢ – x̄)²)
8. Interpreting Correlation Coefficients | 解释相关系数的常见错误
A correlation coefficient r = 0.2 may be statistically significant with a large sample, but its practical significance might be negligible. Conversely, r = 0.8 with a tiny sample might not be significant. Do not conflate the size of the correlation with its significance level.
相关系数 r = 0.2 在大样本下可能具有统计显著性，但它的实际重要性可能微乎其微。反之，小样本下 r = 0.8 可能并不显著。切勿混淆相关系数的大小与其显著性水平。
Also, the coefficient of determination R² = r² indicates the proportion of variance in one variable explained by the other. For r = 0.7, R² = 0.49, meaning 49% of the variance is shared, which is often misinterpreted as a ‘stronger’ relationship than it actually is.
再者，测定系数 R² = r² 表示一个变量的变异中可被另一个变量解释的比例。r = 0.7 时 R² = 0.49，意味着 49% 的变异被共享，人们常常高估这种关系的强度。
9. Effect Size vs. Significance | 效应量与显著性的区分
Statistical significance (p < 0.05) does not mean the effect is large. In psychological studies, a treatment might produce a statistically significant improvement with a tiny effect size (e.g., Cohen's d = 0.1). Edexcel problems may require you to comment on both p-values and effect size. Neglecting effect size is a typical error.
统计显著性（p < 0.05）并不意味着效应量大。在心理学研究中，一种干预可能产生统计上显著但效应量极小的改善（例如 Cohen's d = 0.1）。Edexcel 试题可能要求你同时对 p 值和效应量进行评论。忽略效应量是典型错误。
Cohen’s d = (mean₁ – mean₂) / pooled standard deviation. A small effect is around 0.2, medium 0.5, large 0.8. Report both significance and magnitude.
Cohen’s d = (均值₁ – 均值₂) / 合并标准差。小效应约为 0.2，中效应 0.5，大效应 0.8。既报告显著性也报告大小。
10. Chi-Squared Test Errors | 卡方检验的易错点
Chi-squared tests for independence are common in psychology contingency tables. A frequent error is using percentages instead of raw frequencies to compute expected values. The chi-squared statistic must be based on counts, not proportions.
独立性卡方检验在心理学列联表中十分常见。一个常见错误是使用百分比而非原始频数计算期望值。卡方统计量必须基于计数，而非比例。
Another pitfall is ignoring Yates’ continuity correction for 2×2 tables, or violating the condition that all expected frequencies should be ≥ 5. If this condition fails, Fisher’s exact test should be used.
另一个陷阱是忘了对 2×2 表格使用耶茨连续性校正，或者违反了所有期望频数应 ≥ 5 的条件。若该条件不满足，应使用费雪精确检验。
χ² = Σ (O – E)² / E
11. Degrees of Freedom Confusion | 自由度的混淆
Degrees of freedom (df) are often miscalculated. For a chi-squared test, df = (rows – 1) × (columns – 1). For an independent samples t-test, df = n₁ + n₂ – 2. Using the wrong df leads to incorrect critical values and mistaken rejection decisions.
自由度 (df) 经常被算错。对于卡方检验，df = (行数 – 1) × (列数 – 1)；对于独立样本 t 检验，df = n₁ + n₂ – 2。使用错误的自由度会得到错误的临界值，从而做出不当的拒绝判断。
In Edexcel probability distributions, students also mix up df for t-distribution and F-distribution. Always label your working clearly to avoid confusion.
在 Edexcel 概率分布中，学生也会混淆 t 分布和 F 分布的自由度。清晰标注计算过程可以避免混乱。
12. Reporting and Conclusion Errors | 报告与结论的误区
When writing conclusions, students sometimes state ‘accept H₀’. Instead, you should say ‘fail to reject H₀’ because a non-significant result does not prove the null is true; it merely indicates insufficient evidence against it.
撰写结论时，学生有时会写“接受 H₀”。正确的表述应是“未能拒绝 H₀”，因为不显著的结果并不证明零假设为真，只是说明没有足够证据反驳它。
Also, always mention the significance level used and the obtained test statistic along with the p-value. For example, ‘t(28) = 2.45, p = 0.021, two-tailed’. This complete reporting prevents ambiguity.
同时，始终应注明所使用的显著性水平、计算出的检验统计量以及 p 值。例如“t(28) = 2.45, p = 0.021, 双尾”。这样的完整报告可避免歧义。
Published by TutorHao | Edexcel Mathematics Revision Series | aleveler.com
更多咨询请联系16621398022（同微信）
屏轩国际教育cambridge primary/secondary checkpoint, cat4, ukiset,ukcat,igcse,alevel,PAT,STEP,MAT, ibdp,ap,ssat,sat,sat2课程辅导，国外大学本科硕士研究生博士课程论文辅导Cancel reply