Bayesian analysis offers several fit indices such as PPP, DIC, BIC, Bayes factor, etc.
Among these, what can be comparable to fit statistics by ML.
PPP may be comparable to a p-value for a chi-square test of model fit, although there are some arguments. DIC may not be comparable to BIC and AIC by ML especially under small samples and random effect models, right?
1. How about BIC? BIC obtained from Mplus version 7 can be comparable to BIC by ML?
2. Since Bayes factor is a indicator when comparing two models, is it similar to chi-square to df ratio?