Analysing longitudinal data via nonlinear models in randomized block designs

Abstract

Two ways of introducing block effects into nonlinear models are considered. One enters the block effects in a linear fashion and the other in a nonlinear fashion. In both cases block effects can be considered as fixed or random. Additionally to the block effects, we discuss the incorporation of different covariance structures into the model to take into account the longitudinal nature of the data. Existing methods for fitting nonlinear models for data from completely randomized designs are shown to be appropriate for this setup. A numerical growth curve example for eucalyptus trees is presented. © 2001 Elsevier Science B.V. All rights reserved.