ann feng posted on Tuesday, November 23, 2010 - 3:20 pm
I have a quick question relating to the information criteria in the output of a random effect logistic regression model (or a multilevel model estimated by PML method). Is the AIC/BIC (Schwarz's Bayesian Information Criteria?) error- based (i.e. SSE) or likelihood-based? Thanks in advance for clearing this up for me!