This project is about adapting chi-squared-based SEM fit indices for Bayesian SEM. The project is carried out by:

• Mauricio Garnier-Villarreal (faculty page)
• Terrence D. Jorgensen (faculty page)

The manuscript was accepted for publication in Psychological Methods in April 2019. The accepted version of the manuscript is available in the Files section of this project, and supplemental materials are organized as follows:

• Simulation: All materials related to our Monte Carlo simulation investigating the sampling behavior of SEM and BSEM fit indices. This contains 3 subfolders:
  • R syntax: source code for conducting our Monte Carlo study
  • data: the results of our Monte Carlo simulation
  • figures: figures summarizing results across conditions. Some are included in the manuscript, and others are supplementary online materials.
• Application: An illustrative example applied to a real data set, demonstrating how the proposed fit indices can be calculated using the R package blavaan.

Note that the simulation was run on a high-performance computing cluster, not a personal desktop or laptop (which could have taken years). Those interested in replicating the results in the "data" folder would need to adapt the R syntax file "run_simulation_on_HPC.R", but could still utilize the functions in the file "functions_for_simulation.R". The source code in "functions_for_blavaan.R" has since been incorporated into the R package blavaan (see the "blavFitIndices" function).