协方差分析（以及部分方差分析）中因不可靠性导致的有偏参数估计和Ⅰ类错误率膨胀：替代方法和补救策略。
Biased parameter estimates and inflated Type I error rates in analysis of covariance (and analysis of partial variance) arising from unreliability: alternatives and remedial strategies.
机构信息
Department of Psychology, Northwestern University, Family Institute at Northwestern University, Evanston, IL 60208-2710, USA.
出版信息
J Abnorm Psychol. 2010 May;119(2):307-19. doi: 10.1037/a0017552.
Abstract
Miller and Chapman (2001) argued that 1 major class of misuse of analysis of covariance (ANCOVA) or its multiple regression counterpart, analysis of partial variance (APV), arises from attempts to use an ANCOVA/APV to answer a research question that is not meaningful in the 1st place. Unfortunately, there is another misuse of ANCOVAs/APVs that arises frequently in psychopathology studies even when addressing consensually meaningful research questions. This misuse arises from inflated Type I error rates in ANCOVA/APV inferential tests of the unique association of the independent variable with the dependent variable when the covariate and independent variables are correlated and measured with error. Alternatives to conventional ANCOVAs/APVs are discussed, as are steps that can be taken to minimize the impact of this bias on drawing valid inferences when conventional ANCOVAs/APVs are used.
摘要
米勒和查普曼（2001）认为，协方差分析（ANCOVA）或其多元回归对应物——部分方差分析（APV）的一大类误用源于试图使用 ANCOVA/APV 来回答一个本身没有意义的研究问题。不幸的是，即使在解决一致认为有意义的研究问题时，协方差分析/部分方差分析在心理病理学研究中也经常出现另一种误用。这种误用源于当协变量和自变量相关且存在测量误差时，对自变量与因变量的独特关联进行 ANCOVA/APV 推断检验时，I 型错误率膨胀。本文讨论了传统的协方差分析/部分方差分析的替代方法，以及在使用传统的协方差分析/部分方差分析时，为了尽量减少这种偏差对得出有效推论的影响，可以采取的步骤。
相似文献
1
Biased parameter estimates and inflated Type I error rates in analysis of covariance (and analysis of partial variance) arising from unreliability: alternatives and remedial strategies.协方差分析（以及部分方差分析）中因不可靠性导致的有偏参数估计和Ⅰ类错误率膨胀：替代方法和补救策略。
J Abnorm Psychol. 2010 May;119(2):307-19. doi: 10.1037/a0017552.
2
Analysis of covariance in parallel-group clinical trials with pretreatment baselines.具有治疗前基线的平行组临床试验中的协方差分析。
Biometrics. 1987 Dec;43(4):895-901.
3
The power of a paired t-test with a covariate.带有协变量的配对t检验的功效。
Soc Sci Res. 2015 Mar;50:277-91. doi: 10.1016/j.ssresearch.2014.12.004. Epub 2014 Dec 17.
4
Impact Evaluation Using Analysis of Covariance With Error-Prone Covariates That Violate Surrogacy.
Eval Rev. 2019 Dec;43(6):335-369. doi: 10.1177/0193841X19877969. Epub 2019 Oct 2.
5
Analysis of Covariance in Randomized Experiments with Heterogeneity of Regression and a Random Covariate: The Variance of the Estimated Treatment Effect at Selected Covariate Values.在存在回归异质性和随机协变量的随机实验中进行协方差分析：在选定协变量值下估计处理效应的方差。
Multivariate Behav Res. 2020 Nov-Dec;55(6):926-940. doi: 10.1080/00273171.2019.1693953. Epub 2019 Dec 3.
6
Using analysis of covariance (ANCOVA) with fallible covariates.使用有错误协变量的协方差分析（ANCOVA）。
Psychol Methods. 2011 Jun;16(2):166-78. doi: 10.1037/a0023355.
7
Type I Error Rates and Parameter Bias in Multivariate Behavioral Genetic Models.多变量行为遗传模型中的I型错误率和参数偏差
Behav Genet. 2019 Jan;49(1):99-111. doi: 10.1007/s10519-018-9942-y. Epub 2018 Dec 20.
8
The Comparison of Latent Variable Propensity Score Models to Traditional Propensity Score Models under Conditions of Covariate Unreliability.在协变量不可靠的情况下，潜在变量倾向评分模型与传统倾向评分模型的比较。
Multivariate Behav Res. 2020 Jul-Aug;55(4):625-646. doi: 10.1080/00273171.2019.1663136. Epub 2019 Sep 18.
9
Analysis of covariance (ANCOVA) with difference scores.采用差异分数的协方差分析（ANCOVA）。
Int J Psychophysiol. 2004 May;52(3):277-83. doi: 10.1016/j.ijpsycho.2003.12.009.
10
Misunderstanding analysis of covariance.协方差的误解分析
J Abnorm Psychol. 2001 Feb;110(1):40-8. doi: 10.1037//0021-843x.110.1.40.
引用本文的文献
1
Differential Treatment Response of Proactive and Reactive Partner Abusive Men: Results from a Laboratory Proximal Change Experiment.主动型和反应型施虐伴侣男性的差异化治疗反应：来自实验室近端变化实验的结果。
Psychosoc Interv. 2024 Jan 2;33(1):43-54. doi: 10.5093/pi2024a2. eCollection 2024 Jan.
2
The Questionable Practice of Partialing to Refine Scores on and Inferences About Measures of Psychological Constructs.对心理建构测量的分数和推论进行部分化以提高分数的可疑做法。
Annu Rev Clin Psychol. 2023 May 9;19:155-176. doi: 10.1146/annurev-clinpsy-071720-015436. Epub 2023 Feb 7.
3
How failure to falsify in high-volume science contributes to the replication crisis.大量科学研究中未能证伪导致了复制危机。
Elife. 2022 Aug 8;11:e78830. doi: 10.7554/eLife.78830.
4
Emotional and behavioural pathways to adolescent substance use and antisocial behaviour: results from the UK Millennium Cohort Study.青少年物质使用和反社会行为的情绪和行为途径：来自英国千禧年队列研究的结果。
Eur Child Adolesc Psychiatry. 2021 Nov;30(11):1813-1823. doi: 10.1007/s00787-020-01661-x. Epub 2020 Nov 2.
5
The Effects of Posttraumatic Stress and Trauma-Focused Disclosure on Experimental Pain Sensitivity Among Trauma-Exposed Women.创伤后应激和创伤聚焦披露对创伤后女性实验性疼痛敏感性的影响。
J Trauma Stress. 2020 Dec;33(6):1071-1081. doi: 10.1002/jts.22571. Epub 2020 Aug 13.
6
Compound Bias due to Measurement Error When Comparing Regression Coefficients.比较回归系数时因测量误差导致的复合偏差。
Educ Psychol Meas. 2020 Jun;80(3):548-577. doi: 10.1177/0013164419874494. Epub 2019 Sep 16.
7
It's the thought that counts: belief in suicide as an escape moderates the relationship between emotion dysregulation and suicidal ideation cross-sectionally and longitudinally.重要的是想法：将自杀视为一种逃避的信念会在横断面和纵向上缓和情绪失调与自杀意念之间的关系。
Borderline Personal Disord Emot Dysregul. 2019 Nov 5;6:16. doi: 10.1186/s40479-019-0112-5. eCollection 2019.
8
Musical training, bilingualism, and executive function: working memory and inhibitory control.音乐训练、双语能力与执行功能：工作记忆和抑制控制
Cogn Res Princ Implic. 2018;3(1):11. doi: 10.1186/s41235-018-0095-6. Epub 2018 Apr 11.
9
Statistically Controlling for Confounding Constructs Is Harder than You Think.对混杂结构进行统计控制比你想象的要难。
PLoS One. 2016 Mar 31;11(3):e0152719. doi: 10.1371/journal.pone.0152719. eCollection 2016.
10
The Roles of Dopamine and Hypocretin in Reward: A Electroencephalographic Study.多巴胺和下丘脑泌素在奖赏中的作用：一项脑电图研究。
PLoS One. 2015 Nov 23;10(11):e0142432. doi: 10.1371/journal.pone.0142432. eCollection 2015.
本文引用的文献
1
Indications for propensity scores and review of their use in pharmacoepidemiology.倾向评分的适应症及其在药物流行病学中的应用综述。
Basic Clin Pharmacol Toxicol. 2006 Mar;98(3):253-9. doi: 10.1111/j.1742-7843.2006.pto_293.x.
2
SPSS and SAS procedures for estimating indirect effects in simple mediation models.用于估计简单中介模型中间接效应的SPSS和SAS程序。
Behav Res Methods Instrum Comput. 2004 Nov;36(4):717-31. doi: 10.3758/bf03206553.
3
The integration of continuous and discrete latent variable models: potential problems and promising opportunities.连续和离散潜变量模型的整合：潜在问题与前景机遇。
Psychol Methods. 2004 Mar;9(1):3-29. doi: 10.1037/1082-989X.9.1.3.
4
Assessing psychological change in adulthood: an overview of methodological issues.评估成年期的心理变化：方法学问题概述
Psychol Aging. 2003 Dec;18(4):639-57. doi: 10.1037/0882-7974.18.4.639.
5
Potential problems with "well fitting" models.“拟合良好”模型的潜在问题。
J Abnorm Psychol. 2003 Nov;112(4):578-98. doi: 10.1037/0021-843X.112.4.578.
6
The statistical power of abnormal-social psychological research: a review.异常社会心理学研究的统计功效：一项综述。
J Abnorm Soc Psychol. 1962 Sep;65:145-53. doi: 10.1037/h0045186.
7
Mediation in experimental and nonexperimental studies: new procedures and recommendations.实验性研究与非实验性研究中的中介作用：新程序与建议
Psychol Methods. 2002 Dec;7(4):422-45.
8
A comparison of methods to test mediation and other intervening variable effects.检验中介效应和其他干预变量效应的方法比较。
Psychol Methods. 2002 Mar;7(1):83-104. doi: 10.1037/1082-989x.7.1.83.
9
Misunderstanding analysis of covariance.协方差的误解分析
J Abnorm Psychol. 2001 Feb;110(1):40-8. doi: 10.1037//0021-843x.110.1.40.
10
Estimating causal effects from large data sets using propensity scores.使用倾向得分从大型数据集中估计因果效应。
Ann Intern Med. 1997 Oct 15;127(8 Pt 2):757-63. doi: 10.7326/0003-4819-127-8_part_2-199710151-00064.
Zotero 插件