Likelihood Ratio Criterion for Mean Structure in the Growth Curve Model with Random Effects

✍ Scribed by Hironori Fujisawa

• Publisher: Elsevier Science
• Year: 1997
• Tongue: English
• Weight: 734 KB
• Volume: 60
• Category: Article
• ISSN: 0047-259X
• DOI: 10.1006/jmva.1996.1644

✦ Synopsis

In the present paper, we consider the likelihood ratio criterion (LRC) for mean structure in the growth curve model with random effects. It is difficult to express the LRC as a closed form because of a restriction on parameters. The lower bound and upper bound of the LRC are suggested as a closed form. By using them, it is shown that the least favorable distribution of the LRC is 4-type distribution. 1997 Academic Press

where ; i is the random coefficient vector, e i is the p_1 error vector distributed as N(0, _ 2 I p ), ' i is the q_1 random effect vector distributed as N(0, 1), e i 's and ' i 's are mutually independent. The variations of the article no. MV961644 90 0047-259XÂ97 25.00