THE USE OF CONTRASTS IN MULTIVARIATE NONLINEAR MIXED MODELS TO COMPARE TREATMENTS IN LONGITUDINAL FACTORIAL EXPERIMENTS

Abstract

The purpose of this study was to establish contrasts in multivariate nonlinear mixed models to verify the effects of treatments in experiments with longitudinal data and multiple responses. The evaluated nonlinear functions were the three parameters curves logistic, Gompertz and von Bertalanffy. The random variables were added to the fixed parameters, asymptote α , abscissa of the inflection point  β, and parameter γ. The best fitted model was expanded with covariates, which establish orthogonal contrasts, in order to verify main effects and interactions in factorial experiments. The methodology was applied to analyse data of an experiment with citrus, in which case the logistic bivariate mixed effects model was the best fit. The chosen model allowed comparisons between treatments in a global context of more than one dependent variable and throughout the measurement period.