Linda, Bengt, and/or Tihomir - Do you have any recommendations for papers that examine ILD (or data collected with ecological momentary assessment) using growth mixture modeling? I have data with up to 25 assessments for each patient that I'd like to examine for latent growth classes. I'm guessing that simple models with only linear and nonlinear slopes will not be sufficient to capture the change, and that piecewise effects or other components will be necessary. Thanks! Bruce
Are these assessments made at the same time points for each subject or are the time points random? There was a 2008 Psychometrika article by Shedden & Zucker who considered mixtures of flexible growth shapes, but I think that requires approximately the same time points.
Thank you, Bengt! This will give me some help as I start! The assessments are regular for my data -- baseline then monthly for the first year, then quarterly for each of the four following years. Haven't read it closely yet, but I found a "related record" in Web of Knowledge when I located the Shedden & Zucker article. It also looked interesting re examining how many assessments are important for LCA with ILD, although the authors (from The Methodology Center) used PROC TRAJ and the usual linear & quadratic slopes. FYI, that article is Tan, Dierker, et al. (2011), How spacing of data collection may impact estimates of substance use trajectories. Substance Use & Misuse, 46(6), 758-768. Thanks! Bruce