ABSTRACT
The focus of this chapter is on latent variable structural equation modeling (SEM) of cross-cultural data. Within the context of SEM, two general approaches for modeling between-culture differences and similarities exist. The first approach is multiple group mean and covariance structures (MACS) modeling (e.g., Browne & Arminger, 1995; Card & Little, 2006; Sörbom, 1982), and the second is multilevel SEM (ML SEM; e.g., Cheung & Au, 2005; Hox, 2002; Mehta & Neale, 2005). As we will elaborate in more detail below, the first approach is generally utilized when the number of cultures under consideration is small or relatively finite (e.g., <20 or so) and one is generally interested in comparing structural relations across specific groups (e.g., tests of measurement invariance; see below).
