Anonymous posted on Sunday, July 25, 2004 - 10:49 am
I have some questions on multiple group multilevel modeling:
1) Is it possible to undertake structured means modeling (if so, are there any limitations to such modeling of latent means at the within-group or between-group level)?
2) I am somewhat confused about how Mplus treats multiple groups in the multilevel analysis. If the grouping variable is a "within variable" (i.e., gender of students within schools), it appears that each gender and school is now treated as a cluster and no overall between group modeling (school-based, not school-gender-based) is possible. Is this correct?
bmuthen posted on Sunday, July 25, 2004 - 12:06 pm
1) Yes. In line with regular multilevel modeling such as regression analysis, mean-related parameters are in the between part of the model.
2)The grouping variable is expected to be a between-level variable with no within variation. Special approaches are needed for grouping variables that vary across within units.