On 8/04/2009, at 1:27 PM, Ben Bolker wrote:
<snip>
I agree that that the individual-level random effect is probably
the issue.
I played with this some today but didn't manage to resolve it --
tried JAGS/R2jags and glmer from lme4 but didn't manage to
get an estimate of epsilon that matched the input value. I'm
a little worried about binary data with an underlying random
effect, I think there's an identifiability problem there ...
Oh dear. This is getting hairier than I expected ... and a bit
over my head. The toy model proposed *looks* simple enough at
first blush. I would've thought that if a model which is that
simple doesn't ``work'', then what hope is there for the more
complicated models that one needs for analyzing real data?
cheers,
Rolf
######################################################################
Attention:\ This e-mail message is privileged and confid...{{dropped:9}}
______________________________________________
R-help@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.