GROWTH
CURVES: ANALYSIS OF INCOMPLETE LONGITUDINAL DATA WITH AUTOCORRELATED ERRORS
Luiz de SOUZA[1]
Antonio
Francisco IEMMA[2]
§
ABSTRACT: The aim of this work is to study incomplete
longitudinal data using growth curve models. Under the assumption of a
first-order autoregressive process for the errors covariance matrix, maximum
likelihood estimators of the parameters of a proposed model, are obtained,. Test of hypothesis CxU = q are performed using
Wald’stastistics.
§ KEYWORDS: Longitudinal data; growth curves; autoregressive process.