[6523] in s-news-athena
No subject found in mail header
daemon@ATHENA.MIT.EDU (Jose C. Pinheiro)
Mon Feb 27 10:54:50 1995
Date: Mon, 27 Feb 95 09:35:28 -0600
From: pinheiro@stat.wisc.edu (Jose C. Pinheiro)
To: bates@stat.wisc.edu, lindstro@stat.wisc.edu
Cc: about@stat.wisc.edu, from@stat.wisc.edu, is@stat.wisc.edu,
person@stat.wisc.edu, S-news@stat.wisc.edu, the@stat.wisc.edu,
This@stat.wisc.edu, to@stat.wisc.edu, varcomp@stat.wisc.edu,
who@stat.wisc.edu, wrote@stat.wisc.edu
Yes, we didn't think about this case when we modified the code. Before, it was
messing up even when the variables were defined in the same data frame. It seems
that people tend to use the software in ways we would never expected them to.
I guess what we can do is to create an indicator vector mapping the missing
values to the original observations and remove them at the end of predict.lme().
I will take a look at it, but I will only be able to do it after Wednesday when
I have a meeting on a clinical trial I am working on.
--Jose'
