Modelling Multivariate Longitudinal Data

  • Inês Sousa
  • Departament of Mathematics and Applications of the University of Minho
  • Local: ZOOM – 13:00 – Link
  • Quarta-feira, 2 de dezembro de 2020
  • Seminário Conjunto CEAUL e CEMAT
  • Referência Projeto: UIDB/00006/2020 and UIDB/04621/2020
In observational longitudinal studies it is common to have repeated measures of more than one process that changes over time, being the main objective understanding the association between the two processes and if there is a more relevant one. For example, in a weekly psychotherapy process several evaluations are made to the client as symptomatology, empathy to the psychoterapist or proportion of innovative moments. In these studies it is important to understand in which way symptomatology and innovative moments are associated and to identify which process drives individual changes. In this talk we propose a multivariate model for two longitudinal variables in the context of linear mixed models, exploring different correlation structures in such a way to be possible to answer to the scientific questions. Different distributions will be discussed for longitudinal variables