辛普森悖论揭秘。
The Simpson's paradox unraveled.
机构信息
Department of Epidemiology, Harvard School of Public Health, Harvard-MIT Division of Health Sciences and Technology, Boston, MA 02115, USA.
出版信息
Int J Epidemiol. 2011 Jun;40(3):780-5. doi: 10.1093/ije/dyr041. Epub 2011 Mar 30.
Abstract
BACKGROUND
In a famous article, Simpson described a hypothetical data example that led to apparently paradoxical results.
METHODS
We make the causal structure of Simpson's example explicit.
RESULTS
We show how the paradox disappears when the statistical analysis is appropriately guided by subject-matter knowledge. We also review previous explanations of Simpson's paradox that attributed it to two distinct phenomena: confounding and non-collapsibility.
CONCLUSION
Analytical errors may occur when the problem is stripped of its causal context and analyzed merely in statistical terms.
摘要
背景
辛普森在一篇著名文章中描述了一个假设的数据例子，导致了明显的悖论结果。
方法
我们使辛普森例子的因果结构变得明确。
结果
我们展示了当统计分析受到主题知识的适当指导时，悖论是如何消失的。我们还回顾了以前对辛普森悖论的解释，这些解释将其归因于两种截然不同的现象：混杂和不可 collapsibility。
结论
当问题被剥夺其因果背景并仅从统计角度进行分析时，可能会出现分析错误。
相似文献
1
The Simpson's paradox unraveled.辛普森悖论揭秘。
Int J Epidemiol. 2011 Jun;40(3):780-5. doi: 10.1093/ije/dyr041. Epub 2011 Mar 30.
2
Simpson's paradox and experimental research.辛普森悖论与实验研究。
Nurs Res. 2009 Mar-Apr;58(2):123-7. doi: 10.1097/NNR.0b013e318199b517.
3
Misleading Epidemiological and Statistical Evidence in the Presence of Simpson's Paradox: An Illustrative Study Using Simulated Scenarios of Observational Study Designs.辛普森悖论存在时的误导性流行病学和统计证据：一项使用观察性研究设计模拟场景的说明性研究
J Med Life. 2020 Jan-Mar;13(1):37-44. doi: 10.25122/jml-2019-0120.
4
Response to: Simpson's Paradox is suppression, but Lord's Paradox is neither: clarification of and correction to Tu, Gunnell, and Gilthorpe (2008) by Nickerson CA & Brown NJL (https://doi.org/10.1186/1742-7622-5-2).回应：辛普森悖论是抑制现象，但洛德悖论并非如此：对尼克森CA和布朗NJL对涂、冈内尔和吉尔索普（2008年）的澄清与修正（https://doi.org/10.1186/1742-7622-5-2） 。
Emerg Themes Epidemiol. 2020 Mar 11;17:1. doi: 10.1186/s12982-020-00089-7. eCollection 2020.
5
Simpson's paradox in psychological science: a practical guide.心理学中的辛普森悖论：实用指南。
Front Psychol. 2013 Aug 12;4:513. doi: 10.3389/fpsyg.2013.00513. eCollection 2013.
6
The role of causal reasoning in understanding Simpson's paradox, Lord's paradox, and the suppression effect: covariate selection in the analysis of observational studies.因果推理在理解辛普森悖论、罗德悖论和抑制效应中的作用：观察性研究分析中的协变量选择
Emerg Themes Epidemiol. 2008 Feb 26;5:5. doi: 10.1186/1742-7622-5-5.
7
Bounds for pure direct effect.纯直接效应的界限。
Epidemiology. 2014 Sep;25(5):775-6. doi: 10.1097/EDE.0000000000000154.
8
The potential for Simpson's paradox in drug utilization studies.药物利用研究中出现辛普森悖论的可能性。
Ann Epidemiol. 1997 Oct;7(7):517-21. doi: 10.1016/s1047-2797(97)00084-7.
9
Confounding and Simpson's paradox.混杂因素与辛普森悖论
BMJ. 1994 Dec 3;309(6967):1480-1. doi: 10.1136/bmj.309.6967.1480.
10
Effect decomposition in the presence of an exposure-induced mediator-outcome confounder.存在暴露引起的中介结局混杂因素时的效应分解。
Epidemiology. 2014 Mar;25(2):300-6. doi: 10.1097/EDE.0000000000000034.
引用本文的文献
1
Assessing the spread risk of COVID-19 associated with multi-mode transportation networks in China.评估中国与多模式交通网络相关的新冠病毒传播风险。
Fundam Res. 2022 Apr 22;3(2):305-310. doi: 10.1016/j.fmre.2022.04.006. eCollection 2023 Mar.
2
CorrAdjust unveils biologically relevant transcriptomic correlations by efficiently eliminating hidden confounders.CorrAdjust通过有效消除隐藏的混杂因素，揭示了生物学上相关的转录组相关性。
Nucleic Acids Res. 2025 May 22;53(10). doi: 10.1093/nar/gkaf444.
3
The Same but Different?: A Systematic Review of the Impact of Selection and Collider Bias on Internal Validity.相同却又不同？：对选择偏倚和对撞机偏倚对内部效度影响的系统评价
Epidemiology. 2025 Jul 1;36(4):473-481. doi: 10.1097/EDE.0000000000001864. Epub 2025 Apr 1.
4
Protection from prior natural infection vs. vaccination against SARS-CoV-2-a statistical note to avoid biased interpretation.既往自然感染与接种 SARS-CoV-2 疫苗的防护作用——避免偏倚解读的统计学说明
Front Med (Lausanne). 2024 Jun 12;11:1376275. doi: 10.3389/fmed.2024.1376275. eCollection 2024.
5
Biologically meaningful genome interpretation models to address data underdetermination for the leaf and seed ionome prediction in Arabidopsis thaliana.用于解决拟南芥叶和种子离子组预测中数据不足问题的具有生物学意义的基因组解释模型。
Sci Rep. 2024 Jun 8;14(1):13188. doi: 10.1038/s41598-024-63855-6.
6
Estimation of the causal effect of sex on neonatal intensive care unit outcomes among very low birth weight infants.估计性别对极低出生体重儿新生儿重症监护病房结局的因果效应。
J Perinatol. 2024 Jun;44(6):844-850. doi: 10.1038/s41372-024-01989-1. Epub 2024 May 6.
7
Drug Shortages Prior to and During the COVID-19 Pandemic.药品短缺：新冠疫情前后
JAMA Netw Open. 2024 Apr 1;7(4):e244246. doi: 10.1001/jamanetworkopen.2024.4246.
8
Statistical modelling of determinants of child stunting using secondary data and Bayesian networks: a UKRI Global Challenges Research Fund (GCRF) Action Against Stunting Hub protocol paper.利用二手数据和贝叶斯网络对儿童发育迟缓决定因素进行统计建模：英国研究与创新署全球挑战研究基金（GCRF）遏制发育迟缓行动中心方案文件。
BMJ Paediatr Open. 2024 Mar 22;8(Suppl 1):e001983. doi: 10.1136/bmjpo-2023-001983.
9
Athletic Injury Research: Frameworks, Models and the Need for Causal Knowledge.运动损伤研究：框架、模型和因果知识的需求。
Sports Med. 2024 May;54(5):1121-1137. doi: 10.1007/s40279-024-02008-1. Epub 2024 Mar 20.
10
The Prevalence of Parkinson's Disease in Poland: Regional and Sex-Related Differences.波兰帕金森病的流行情况：地区和性别差异。
J Parkinsons Dis. 2024;14(3):521-532. doi: 10.3233/JPD-230291.
本文引用的文献
1
Causal directed acyclic graphs and the direction of unmeasured confounding bias.因果有向无环图与未测量混杂偏倚的方向
Epidemiology. 2008 Sep;19(5):720-8. doi: 10.1097/EDE.0b013e3181810e29.
2
A structural approach to selection bias.一种针对选择偏倚的结构化方法。
Epidemiology. 2004 Sep;15(5):615-25. doi: 10.1097/01.ede.0000135174.63482.43.
3
Causal knowledge as a prerequisite for confounding evaluation: an application to birth defects epidemiology.因果知识作为混杂因素评估的先决条件：在出生缺陷流行病学中的应用
Am J Epidemiol. 2002 Jan 15;155(2):176-84. doi: 10.1093/aje/155.2.176.
4
Data, design, and background knowledge in etiologic inference.病因推断中的数据、设计与背景知识。
Epidemiology. 2001 May;12(3):313-20. doi: 10.1097/00001648-200105000-00011.
5
Confounding: essence and detection.混杂因素：本质与检测
Am J Epidemiol. 1981 Oct;114(4):593-603. doi: 10.1093/oxfordjournals.aje.a113225.
6
Identifiability, exchangeability, and epidemiological confounding.可识别性、可交换性与流行病学混杂因素。
Int J Epidemiol. 1986 Sep;15(3):413-9. doi: 10.1093/ije/15.3.413.
7
Interpretation and choice of effect measures in epidemiologic analyses.流行病学分析中效应测量指标的解读与选择
Am J Epidemiol. 1987 May;125(5):761-8. doi: 10.1093/oxfordjournals.aje.a114593.
Zotero 插件