An extended forward search algorithm to detect latent dimensionality, population heterogeneity and measurement invariance in latent variable models

Irini Moustaki, London School of Economics

  • Date: 25 FEBRUARY 2021  from 16:00 to 18:00

  • Event location: Modalità telematica, mediante sistema di videoconferenza su piattaforma Microsoft Teams

  • Type: Statistics Seminars

Abstract


In this paper, we extend the forward search algorithm developed initially to detect outliers in
regression models and later in factor models for continuous and binary data to address questions of
latent dimensionality, population heterogeneity and measurement invariance. We present simulation
studies and real examples that justify the successful use of the algorithm in all those cases. The FS
algorithm can be used in combination with goodness-of-fit test and fit measures and in some cases to
overcome their limitations.


Co-authors: Yan Lu and and Yunxiao Chen


L’Organizzatore                                  Il Direttore
Prof. Silvia Cagnone                          Prof. Angela Montanari

La S.V. è invitata