Parameter Detection and Recovery Utilizing Bayesian Structural Equation Modeling: A Simulation Study (Poster 7)

• In Event: AERA Poster Session Five
  In Poster Session: Division D Section 2 Poster Session

Abstract

Mediation detection and parameter recovery has been explored in the frequentist framework but has been mostly untouched from a Bayesian framework. Utilizing the lavaan and BSEM packages in R, this paper investigates the characteristics of Bayesian Structural Equation Modeling via simulation to detect and recover parameters of mediators in a simple path diagram by way of highest posterior density intervals. Overall, the BSEM approach showed good results with non-informative priors, and priors with relatively large variances at a wide range of sample sizes. For utilizing smaller variances, much larger sample sizes were required.

Author

• Ethan Harris, University of Arkansas at Fayetteville
  Presenting Author