The institute for Measurement, Methodology, Analysis, and Policy (IMMAP) is hosting Chuck Huber of Stata Corporation to speak on Latent Class Analysis (LCA) with Stata software. LCA is a subset of structural equation modeling (SEM), and uses model-based clustering and classification methods. Clustering is a multivariate method that finds groups of “similar” subjects in contrast to principal component or factor analytic methods that find groups of variables. Clusters are formed based on individual’s response patterns on the dataset variables, where individuals within each cluster are as similar as possible and individuals in other clusters differ substantially. Traditional clustering methods use a heuristic (rule-based) approach, while LCA invokes a latent class variable as part of a model-based approach. LCA is receiving increased attention in applied research settings, and has been used in many disciplines such as social sciences (psychology, education), medicine & health science, marketing & business, and areas of hard-science.
This presentation will demonstrate step-by-step how to conduct LCA (and related latent profile analysis, LPA) using Stata software and a real empirical data set from the published literature. Three important aspects will be presented:
1. Adding covariates to LCA models
2. LCA models with known groups
3. Latent profile analysis (LPA) versus latent class analysis (LCA)
The speaker, Chuck Huber from Stata, will cover basic concepts of LCA and LPA, as well as walk through doing the analysis in Stata, and how to interpret the results. This will be an introductory talk, but some prior familiarity with SEM and latent variable analysis will be helpful.
For further inquiry or directions, contact Roham Agha, (roham.agha@ttu.edu)
