Turning Point Detection in Longitudinal Data Using Piecewise Linear Models in SAS NLMIXED

• In Event: Multilevel Modeling: Nonnested Multilevel Data, Measurement Invariance, and Nonlinear Models

Abstract

A simple piecewise procedure can be used to approximate some nonlinear growth curves. Yet a main problem of using a piecewise approach lies in a judicious choice of the turning point where the change in the growth rate occurs. The purpose of this study is to explore the use of piecewise procedure in nonlinear multilevel modeling framework for estimating unknown turning points in longitudinal data. The study will propose guidelines in identifying good starting values for estimating piecewise linear models in SAS NLMIXED. A simulation study will be conducted to investigate the effectiveness of the proposed guidelines in detecting a medium change in the growth rate between two to three -piece linear models at the turning point(s).

Authors

• Ling Ning, University of California - Davis

• Wen Luo, Texas A&M University - College Station