FSS – Typologies in Data

We start with an overview of logistic and multinomial logistic regression. Then we work on categorical latent variables that identify typologies of subjects, these are “Finite mixture models”. In direct applications, one assumes that the overall population heterogeneity with respect to a set of manifest variables is due to the existence of two or more distinct homogeneous subgroups (latent classes) of individuals. This seminar will introduce participants to the prevailing “best practices” for direct applications of finite mixture modelling to cross-sectional data (Latent Class Analysis, and Latent Profile Analysis), in terms of model assumptions, specification, estimation, evaluation, selection, and interpretation. Models that allow for the inclusion of correlates and predictors of latent class membership as well as distal outcomes of latent class membership will be presented. The course will end by presenting longitudinal extensions of mixture models. These models estimate the probability to stay or switch classes over time, allowing us to study the process of change of meaningful typologies.  

Course objectives:

• Learn how to conduct exploratory data analysis and find typologies in the data using mixture modelling (e.g., latent class modelling).  
• Learn how to run and interpret logistic regression models
• Learn to run, and interpret Latent Class Analysis (LCA, cross-sectional typologies)
• Learn to run, and interpret Hidden Markov Models (HMM, longitudinal typologies)
• Learn how to use obtained typologies in further analysis (e.g., to establish associations between individual- characteristics and cluster membership).  
• Apply LCA and HMM in R
• Learn how to report the results