Nonlinear Growth Curves in Developmental Research
Linear model is an instrument that is quiet friendly for us to analyze data. However, it does not always to fulfill the theoretical concept of data and fit the data. This paper proposed existing model including easy linear and quadratic models, and complicate nonlinear latent growth models. Among these models, the Preece-Baines model is only one model that belongs to theory-driven. The authors adopted four models to fit longitudinal data of height. Results illustrated that the Preece-Bainesb model had the best fit of AIC and BIC indices, followed by the Quadratic model, and the linear model showed the worst fit on these two indices. I was wondering how to apply nonlinear growth curve model in IRT field, which means if dependent change from continuous height to categorical responses. Second, theory driven and data driven are totally different. Do we have any guidance to apply these two approaches? Or it just connects to viewpoint of philosophy while we are doing researches. If a researcher with mind of theory driven, he/she would develop theoretical model to fit corresponding data, and so does researchers who hold on data-basis.