Latent Variable Modeling for Biomedical Applications.

Seminar Series

Friday, February 14, 2020 - 01:30
Hock Plaza 11025
BERD Core Seminars

Abstract: Latent variable models (LVMs) are ubiquitous in social and medical sciences. They provide key platforms for incorporating informative but unobservable variables into modeling frameworks. In this seminar, will explain the general concept of LVMs, when it is appropriate to use LVMs, how to use path diagrams to help construct LVMs and comparisons to other common models. We will also cover using MPlus software for implementing LVMs and how to connect results to R for summarizing results and making plots. Finally, we will present a working example of using LVMs for modeling joint exposure to SES and pollution in North Carolina.

Li Ren, MS 
Biostatistician II
Duke Cancer Institute

Alexandra Larson, PhD
Department of Biostatistics and Bioinformatics
Duke School of Medicine