Which are the exact formulae for MUML and MUMLM in the case of a two-level factor model with unbalanced groups? And how are the standard errors calculated? Are these standard errors robust? Although there are some hints in the MPLUS user's guide, it's difficult to figure out how MUML and MUMLM are calculated. Thanx in advance.
PS If there is a publication on the exact formulae, maybe you could put it in the paper download area, since most of the methodological journals are difficult to come by in Switzerland.
bmuthen posted on Friday, August 22, 2003 - 9:41 am
The formulas for MUML are in the User's Guide Technical Appendix. The MUMLM formulas are not, but they were an attempt to make the SEs and chi-square more correct and more robust to nonnormality - however, due to recent investigations we are not fully satisfied with MUMLM at this point and are planning to replace it; in the meanwhile MUML might be safer.