ABSTRACT
A review of the structural equation modeling literature reveals a dearth of studies involving multigroup comparisons of latent mean structures. This situation is likely a function of the many complexities associated with the implementation of these analyses using earlier versions of the LISREL program (up to and including version 6). Indeed, it was this very issue that led to the writing of the first paper to explicitly model and demonstrate the necessary specification “tricks” required in testing for differences in mean structures (see Byrne et al., 1989). Fortunately, beginning with LISREL 7, the actual specification procedure has been made relatively straightforward. As a consequence of these changes, I expect that more studies of latent mean structures will be forthcoming, albeit they are still relatively far and few between.
