Skip to content

Reconciling Near Identical AIC Values and Highly Significant P Value (Ben Bolker)

1 message · Ed Merkle

#
Ben et al,
Slightly off topic, but speaking of Vuong's tests, I wonder about the 
ease with which we could obtain the necessary output from lme4. To carry 
out the tests, I believe we need individual observations' contributions 
to both the likelihood and gradient (evaluated at the ML estimates).  I 
believe these are not simple to obtain from models fit in lme4, but I 
wonder whether you have any hints here.

Thanks,
Ed