Asymptotic Results for Some Longitudinal Regression Models

Abstract

The scope of this thesis is to examine asymptotic properties for some longitudinal regression models. These models are often used in biostatistics, econometrics and psychology and involve the collection of several measurements, at different occasions, on the same individuals or clusters. First, we examine the classical regression model for longitudinal data and define the asymptotically optimal estimating equation, with strongly consistent solutions. Then, we consider a simple model with error in covariates, adapted to longitudinal data, obtain estimators of the regression parameters and examine their asymptotic behavior.