I'm using the lavaan package for my sem models. In the next months, I'll need to fit a mixed, multilevel sem models. However, lavaan doesn't fit this kind of model yet. Does anyone know of an alternative package?
MPlus, made by Bengt and Linda Muthen can fit multilevel SEMs. They have a very usable manual available online and excellent help via their online forums. I believe the student version is free to download. It can also do things like multilevel mediation.
I too endorse Mplus. However, particularly promising is gsem (Generalized Structural Equation Modeling) by the most recent version of Stata (Stata 13). It can allow the estimation of models to handle multilevel analysis while incorporating Bernoulli procedures.