Skip to content
Prev 1225 / 20628 Next

Different versions of lme4 and covariance of randomeffects

David:

I'm looking at this and am a little confused. You say the differences
are not too dramatic. But, they are. I can see that the syntax for your
model specification is the same, but the estimates of the fit statistics
differ (logLik), the estimates of the fixed effects differ, as well as
their standard errors, and the variance components differ.

Am I missing something? Was the same data set used in both cases?