Title: Parameterizing nontraditional aspects of change as random coefficients

 

Abstract: One of the primary goals of longitudinal data analysis is to estimate and interpret model parameters that reflect meaningful aspects of change. Widely used models often can be reparameterized to yield statistically equivalent models with new, substantively useful parameters. I review general steps involved in the reparameterization of growth curve models within a structured latent curve framework. Subsequently, I describe two longitudinal empirical examples in which reparameterization allows a new parameter of substantive interest to be estimated as a random effect and to be predicted: an aperture point and a transition point, respectively.

 

Suggested Readings: 
           Preacher, K. J., & Hancock, G. R. (2012). On interpretable reparameterizations of linear and nonlinear latent growth curve models. In Harring, J. R., & Hancock, G. R. (Eds.), Advances in longitudinal methods in the social and behavioral sciences (pp. 25-58). Charlotte, NC: Information Age Publishing.