Estimating Regression Coefficients for Balanced Growth Curve Model When Time Trend of Baseline is Not Specified

摘要

This article presents a method for estimating the regression coefficients for a growth curve model when the time trend of the baseline has not been specified. The concept of this method is similar to that of the Cox proportional hazard model. No particular shape is assumed for the baseline time trends, or, alternatively, it can be assumed that they are estimated nonparametrically. Because of these nuisance parameters for the baseline trends, we do not have to pay attention to model those shapes. In addition to the simplicity of modeling baseline curves, we can also nonparametrically describe the baseline trends by using the residuals after the regression coefficients have been estimated.