Transformed statistics for tests of conditional independence in J x K x L contingency tables

JOURNAL OF MULTIVARIATE ANALYSIS

卷:171

Article

Taneichi, Nobuhiro¹  Sekiya, Yuri²  Toyama, Jun³ 

[1] Hokkaido Univ, Kita Ku, Sapporo Campus,1-5,Ainosato 5-3, Sapporo, Hokkaido 0028502, Japan

[2] Hokkaido Univ, Kushiro Campus,15-55,Shiroyama 1, Kushiro, Hokkaido 0858580, Japan

[3] Inst Pract Applicat Math, Sapporo, Hokkaido 0630001, Japan

关键词: Bartlett adjustment;    Bartlett-type adjustment;    Chi-square limiting distribution;    Improved transformation;    Test of conditional independence;    Three-way contingency table;   

DOI  :  10.1016/j.jmva.2018.12.001

来源: Elsevier

【 摘 要 】

Tests of the hypothesis of conditional independence in J x K x L contingency tables are considered. An expression to approximate the null distribution of the test statistics is derived. Using this expression, transformed statistics are obtained which converge to a chi-square limiting distribution faster than the original statistics do. Simulations are used to compare the transformed statistics with the original ones, and transformed statistics are proposed based on a Bartlett-type adjustment. Through this work and earlier ones, we cover testing hierarchical loglinear models in three-way tables except for a model having three interaction terms. (C) 2018 Elsevier Inc. All rights reserved.

【 授权许可】

Free