Skip to content
Prev 11446 / 20628 Next

new version of lme4 - Log-Likelihood of model returned as "-Inf"

Actually, it got stripped by the mailing list software.  Can you post
it somewhere?  (You can also e-mail it to me.)  (Maybe it got censored
because your data frame is called "shits", which I assume is pronounced
"ess-hits" unless these are data on scat samples :-) ?)

 Perhaps you have the Poisson analogue of complete separation, i.e. a
combination of factor levels within which all the responses are zero?
If so, the bglmer function from the blme package might be helpful ...
On 14-02-12 06:14 PM, juwb08 at hampshire.edu wrote: