ABSTRACT
When the comparability of social science scales in international surveys is to be assessed, current survey researchers almost habitually think of multigroup conrmatory factor analysis (MGCFA) as the method of choice (He, Merz, & Alden, 2008; Schmitt & Kuljanin, 2008). Alternative methodological approaches do however exist, with conceptual and practical advantages of their own. Recently, international educational attainment testing surveys such as the Program for International Student Assessment (PISA; OECD, 2005) have given some visibility to models collectively known under the name of item response theory (IRT), which have their origin in educational testing research.* Among these, the Rasch model (Rasch, 1960) is the most basic, but due to its parsimony perhaps also the most elegant approach. Originally designed for dichotomous questions, it has long been extended to also handle the polytomous ordinal response formats more common in attitude research (Andrich, 1978; Masters, 1982). e present chapter will, from an applied perspective, explore the use of a particular version of the polytomous Rasch model to investigate measurement equivalence in comparative research settings. is particular version is the combination of the polytomous Rasch model with latent class analysis in the so-called mixed Rasch model (von Davier & Rost, 1995).
