Statistical Inference for Random-Coefficient Growth Curve Models with Hierarchical Within-Individuals Design Matrices
by
Akari Sato
Department of Mathematics, Hiroshima University, JAPAN
Coauthors: Yasunori Fujikoshi (Department of Mathematics, Hiroshima University, JAPAN), Megu Ohtaki (Department of Environmentrics and Biometrics, Research Institute forRadiation Biology and Medicine, Hiroshima University, JAPAN)

This paper deals with some inferential problems under random-coefficient growth curve models with several hierarchical within-individuals design matrices. The models include the one whose mean structure consists of polynomial growth curves with different degrees. We consider two types of random-coefficient covariance structures. For each of the covariance structures some inferential procedures are proposed, modifying the unbiased estimators of the covariance parameters. We also examine selection criteria for the models. The dental measurement data (see, e.g., Potthoff and Roy (1964, Biometrika, 51, 313-326)) is reexamined, based on random-coefficient growth curve models.

Date received: October 12, 2000