I have a quick question regarding growth mixture modeling (GMM) with continuous outcomes. It seems that most of the literature cites examples where the measurements are taken over arbitrary, discrete points in time (i.e. y1...y4). We have similar measurements taken on both a monthly and weekly basis. With respect to the former monthly measures, each participant's measurement occasion coincides closely to the intended monthly value of time. However, for the weekly assessments, there is, of course, more variation in the timing and frequency of measurement in the sense that the data are unbalanced (i.e. the number and times of weekly assessments vary more across patients than do the monthly measures).
My question: while the monthly data seem applicable to the GMM model, per the literature examples, can GMM also be applied to the more severely unbalanced weekly data through the "individually-varying times of observations" utility? If so, how? I would suspect that weekly assessments provide better information, but again, it suffers from a larger imbalance of both timing and frequency of measurement occasion. Perhaps the weekly data would be better suited with traditional growth modeling, i.e. a random coefficients model.
Any thoughts would be greatly appreciated. Thanks for providing the mixture modeling literature on the website.