Online-Vortrag „New Developments in Measurement Invariance Testing – How Data Science and Causal Inference Can Improve the Comparability of Measurements“

Über den Vortrag

Psychological and educational research often involves unobservable constructs such as cognitive abilities, personality traits, or motivation that have to be inferred from scale or test items aiming at measuring them. When comparing (the means of) the latent variables representing these constructs across groups, it is necessary that the measurements are invariant. However, standard procedures to test measurement invariance (MI) have several downsides and cannot be used to broadly explore MI.

In this talk, I will therefore introduce new statistical approaches inspired by developments in data science that can be used to investigate MI more thoroughly and briefly speak about a causal MI framework that helps researchers to find an adequate modeling strategy that allows for unbiased between-group comparisons. Its main focus lies on model-based recursive partitioning (MOB) and how it can be used to develop tools for MI exploration, e.g., EFA trees (Sterner & Goretzko, 2023).

Über die Veranstaltungsreihe

WRITE ist ein Projekt, das am Regensburger Universitätszentrum für Lehrerbildung (RUL) angesiedelt ist und das junge Nachwuchswissenschaftler:innen im Publikationsprozess unterstützt. Interdisziplinäre Teams arbeiten über ein Jahr an einem gemeinsamen englischsprachigen Artikel und werden von Expert:innen bis zur Publikation in einem Peer-Reviewed Journal begleitet. Parallel werden Workshops zu Forschungsmethodik, Datenanalyse und Wissenschaftlichem Schreiben angeboten. Ausführlichere Informationen zu den aktuellen WRITE-Projekten finden sich unter: https://www.uni-regensburg.de/rul/das-rul/forschungskolleg/write/index.html.