COMPARISON OF GROWTH CURVES FITTING IN LONGITUDINAL AND CROSS-SECTIONAL DESIGNS.

Luiz de SOUZA[1]

§    ABSTRACT: In this paper the results of longitudinal and cross-sectional studies are compared. The model is a composition of Potthoff & Roy (1964), with restricted variances matrix, for longitudinal data, and linear regression for cross-sectional data. The Fisher Scoring method is applied to solve the likelihood equations. The Wald statistic for testing linear hypotheses in model is used. This paper stems from solving practical problems in Medicine.

§    KEYWORDS: Longitudinal and cross-section studies; growth curves; linear regression

 

 



[1] Departamento de Genética e Matemática Aplicada à biologia da Faculdade de Medicina de Ribeirão Preto - USP - 14049-000 - Ribeirão Preto - SP - Brasil.