APS
2022 APS Annual Convention · 2022
Bayesian Model Fit and Selection Indices for Detecting Misspecification: The Case of Bayesian Piecewise Growth Modeling
- Ihnwhi Heo
University of California, Merced - Fan Jia
University of California, Merced - Sarah Depaoli
University of California, Merced
Abstract
We examined the performance of Bayesian model fit and selection indices (CFI, TLI, BIC, DIC, RMSEA, and posterior predictive p-value) to detect model misspecification in Bayesian piecewise growth models. The indices produced disparate performance from one another across various conditions of growth trajectory, sample size, knot location, and prior specification.
Other
