ABSTRACT
Path models are structural equation models that consist of complex paths between latent and/or observed variables, possibly including both direct and indirect effects, and reciprocal effects. A multilevel path model uses the same approaches outlined in Chapter 14 for multilevel factor analysis. With multilevel path models, we may have the complication that there are pure group-level variables (global variables in the terminology of Chapter 1). An example would be the global variable group size, which simply does not exist on the individual level. Current software can deal with this in a straightforward manner.
