Hi, This might be a silly question, but how do I create a pooled model (binomial family) as a baseline for a likelihood-ratio test for a random-intercept lmer model? R won't compare a glm model with an lmer model (or glmmPQL ). Thanks in advance. -Bobby