An Introduction to Latent Class Analysis in Mplus

Session Submission Type: Professional Development Course

Abstract

Traditional Latent Class Analysis (LCA) models the dependencies among observed categorical variables through an underlying latent categorical variable. There are many extensions of LCA that permit other indicator modalities, such as observed continuous and count outcomes, in addition to enabling the investigation of antecedents and consequences of latent class membership. These approaches are often termed “person-centered” analyses in contrast to the “variable-centered” analyses of conventional factor and SEM models. This course will provide participants with an introduction to cross-sectional analyses with categorical latent variables, focusing on LCA and latent profile analysis. Specification and estimation of all models in the Mplus V6 software will be demonstrated throughout. Participants will be provided syntax and annotated output for all models discussed in the workshop.

Sub Unit

• Professional Development and Training Committee

Directors

• Karen L. Nylund-Gibson, University of California - Santa Barbara

• Katherine E. Masyn, Harvard University