Unrestricted Normal Distribution Type Symmetry Model for Square Contingency Tables with Ordered Categories

Journal of Biometrics & Biostatistics

article

Yusuke Saigusa¹  Shimpei Goda²  Kouji Yamamoto³  Sadao Tomizawa² 

[1] Department of Biostatistics, Yokohama City University School of Medicine;Department of Information Sciences, Faculty of Science and Technology, Tokyo University of Science;Department of Medical Statistics, Osaka City University, Graduate School of Medicine

关键词: Bivariate normal distribution;    Contingency table;    Normal distribution type symmetry;    Ordered category;   

DOI  :  10.4172/2155-6180.1000395

来源: Hilaris Publisher

【 摘 要 】

For square contingency tables with ordered categories, Yamamoto, Nakane and Tomizawa [1] proposed the restrictednormal distribution type symmetry (RNDS) model that the cell probability has a similar form as a bivariate normal densitywith equal marginal means and variances. This paper proposes a new model that has more relaxed constraints based ona bivariate normal density. It also gives a decomposition of the RNDS model into the proposed model and marginal meansand variances equality. Moreover it is shown that the test statistic for goodness-of-fit of the RNDS model is asymptoticallyequivalent to the sum of those for the decomposed models.

【 授权许可】

Unknown