ABSTRACT
This chapter describes the group-as-covariate approach, focuses on unordered latent class models, and explains levels of invariance and procedures required to test them making strong links with factor analysis. It also describes a general latent class analysis (LCA) model and demonstrates the application of the method using empirical data. The assessment of measurement invariance with confirmatory LCA involves modeling the LCA together with an observed group. In the LCA context, the chapter discusses at least four invariance levels: no measurement invariance, configural invariance, full metric invariance, full scalar invariance, as well as partial scalar and metric invariance. It also discusses some tools available to test for measurement invariance of latent classes across groups. The chapter assesses the invariance of value classes in the three European regions as two groups: countries of Western and Northern Europe comprise the first group and the countries of Eastern Europe comprise the second group.
