Past, Present, and Future of Model Fit: Modern Solutions to Historical Challenges

Session Type: Symposium

Abstract

Evaluating model fit is an essential aspect of latent variable modeling. In this session, we discuss (a) issues that lead to poor generalizability of traditional fit index cutoffs, (b) simulation-based techniques for deriving custom cutoffs for specific models, (c) modern computational environments that can remove barriers for implementing simulation-based techniques, (d) a tutorial demonstration of a software application that facilitates simulation-based techniques, (e) a comparison of simulation-based techniques to the recently proposed equivalence testing framework, (f) how extensions of these methods to growth models require a more refined definition of what fit indices mean in growth models, and (g) how this refined definition has implications for expanding SEM-type fit indices to evaluate fit of mixed effect models.

Sub Unit

• SIG-Structural Equation Modeling

Chair

• Daniel Katz, University of California - Santa Barbara

Papers

• Unresolved Issues With Hu and Bentler Cutoffs and a Method to Derive More Appropriate Cutoffs - Daniel McNeish, Arizona State University; Melissa Gordon Wolf, University of California - Santa Barbara

• Dynamic Model Fit Indices: A R Shiny Application - Melissa Gordon Wolf, University of California - Santa Barbara; Daniel McNeish, Arizona State University

• Extending the SRMR for SEMs With Mean Structures and Covariates - Tyler Matta, NWEA; Daniel McNeish, Arizona State University

• Moving Beyond Hu and Bentler: Comparing Adjusted Cutoffs From Dynamic Fit Indices and Equivalence Testing - Daniel McNeish, Arizona State University

• Global Model Fit for Mixed-Effect Models - Tyler Matta, NWEA

• A Practical Geometry of Data-Model Fit and Modification - Christian Meyer, University of Maryland; Gregory R. Hancock, University of Maryland

Discussant

• Denis Dumas, University of Georgia