Reviving social class with latent profile analysis for cross-sectional research

Abstract

This contribution picks up the debates on social class and links them to the recent focus on subjective evaluations - namely values. The paper elaborates on a potential measurement construct integrating a subjective evaluation with an objective stratification scale to grasp similarities in modernization and individualization trends in society. Running an explorative latent profile analysis (LPA) on the first 5 waves of the European Social Survey including all country samples available, an optimal number of potential subgroups is extracted, related to socio-demographics and meaningful interpreted. Results show an astonishingly stable number of classes extracted across different national contexts and time points.
In the context of the conference, the presentation aims to reflect critical upon the potential and limitations of the explorative application of LPA for cross-sectional research. With time and countries as covariates, the contribution compares the effort of grasping fine grained differences across time and countries in a complex modeling situation against more pragmatic strategy choices like estimating models for each time point separately or testing for shifts in time for each country. The discussion will focus on resource requirements as well as issues of measurement equivalence in order to assess the different strategies.