Kim De Roover, KU Leuven
Finding clusterwise measurement invariance with mixture multigroup factor analysis
Psychological research often builds on between-group comparisons of (measurements of) latent variables, for instance, to evaluate cross-cultural differences in mindfulness. A critical assumption in such comparative research is that the same latent variable(s) are measured in the same way across all groups (i.e., measurement invariance). Nowadays, measurement invariance is often tested across lots of groups. When (a certain level of) measurement invariance isuntenable across many groups, it is hard to unravel invariances from non-invariances and for which groups they apply. Mixture multigroup factor analysis (MMG-FA; De Roover, 2021; De Roover, Vermunt, & Ceulemans, 2020) was recently proposed to cluster groups based on the measurement parameters, whereas the structural parameters are allowed to differ between groups within a cluster. More specifically, MMG-FA clusters the groups according to a specific level of ‘clusterwise measurement invariance’ (e.g., based on factor loadings only to achieve metric invariance within clusters, or based on loadings and intercepts to achieve scalar invariance within clusters). In this presentation, the full framework of mixture multigroup factor analysis and the ‘mixmgfa’ R-package are presented, as well as how to take the steps from the initial overall level of invariance across all groups to the desired level of clusterwise invariance.
about the speaker
Kim De Roover obtained a PhD in methodology of educational sciences at KU Leuven (Belgium). After a few postdoc years, she was an assistant professor at Tilburg University (the Netherlands) for six years. Currently, she is an associate professor in the group of Quantitative Psychology and Individual Differences at KU Leuven. Her research interests include factor analysis, factor rotation, measurement invariance, structural equation modeling, mixture modeling and model selection. She has done a lot of work on multigroup factor analysis and multigroup structural equation modeling for many groups, combined with mixture modeling (e.g., mixture multigroup factor analysis). She received the Classification Society Distinguished Dissertation Award for her PhD and has obtained funding for several research projects, most recently a Vidi grant from the Netherlands Organization for Scientific Research (800 000 euros) and a Starting Grant from the European Research Council (1 500 000 euros).