[6297] in s-news-athena
Using varcomp()
daemon@ATHENA.MIT.EDU (Ji-Xian Wang)
Sat Jan 28 11:17:32 1995
From: Ji-Xian Wang <eomc13@festival.ed.ac.uk>
To: S-news@stat.wisc.edu
Date: Sat, 28 Jan 95 16:03:34 GMT
I am using varcomp() to fit random effect models and I found that
varcomp() did not work for models with only one random effect. An
example(the one in Guide to Stat. and Math. Analysis) is as follows:
scores: outcomes
times: fixed effect
subject :random effect
>varcomp(scores~subject+times) works OK. But
>varcomp(scores~subject) gets:
Error in Ops.factor(values, m[,varnames[i]]): "*" not meaningful for
factors
Dumped
I also tried my own data with only one random effect and got the same error
message. However, it works well for models with two random effects. Is
it a bug or I did something wrong?
Thank you in advance for any help.
I am using Splus3.2 on SUN.
Jixian Wang
Medical Statistics Unit
University of Edinburgh
