Poster #13 - Sensitivity Analysis for Omitted Confounder in Structural Equation Modeling

• In Event: 47.099 - MTCC Poster Session 11
  In Poster Session: 47.099-1 - Division D Section 2 Statistical Theory and Quantitative Methodologies

Abstract

The SEM model misspecification is common among published studies. The related but unobserved variables were a major problem in model misspecification, and could be potential threatens to the validity of a study. This study discussed the sensitivity analyses beyond observed variables and provided a two parameters approach to assess to what extents the SEM models are sensitive to external misspecification. A simulation study indicates that the proposed approach could detect the unmeasured confounder within latent structures. The future study will apply this approach to applied dataset and demonstrate the capacity of the proposed approach.

Authors

• Huan Kuang, University of Florida

• Walter L. Leite, University of Florida

• Zeyuan Jing, University of Florida