Repeated-measure analyses: Which one? A survey of statistical models and recommendations for reporting

详细信息    查看全文

• 作者：Jacques P. Maurissen^a ; ^{JMaurissen@yahoo.com} ; Thomas J. Vidmar^b
• 关键词：Repeated-measure analysis of variance ; Variance/covariance matrix ; Univariate ; multivariate ; mixed effects ; Type I and II errors ; Reporting recommendations ; Accuracy ; completeness ; transparency
• 刊名：Neurotoxicology and Teratology
• 出版年：2017
• 出版时间：January-February 2017
• 年：2017
• 卷：59
• 期：Complete
• 页码：78-84
• 全文大小：293 K
• 卷排序：59

文摘

Several approaches to the analysis of repeated-measure study designs are available. The data structure is essential in selecting a model. Violating the sphericity assumption can result in undue increase in Type I or II error rates. The recommended procedure for analysis of most repeated-measure data in neurotoxicity studies is the mixed model. Accuracy, completeness and transparency should always be the guiding principles in reporting statistical analyses.