Dear Friends,
I have been following the discussions of mixed-effects models with
some confusion, and I realize that much of this is work in progress,
and that much of the discussion is beyond my knowledge of statistics.
My question is simple, though: Is there a set of commands that will
produce an output equivalent to the exceedingly useful
predict(bl.lm, newdata = bl.new, type = 'response', interval =
'confidence')
for (at least) gaussian mixed-effects model (but preferably for glmm
models)? I have looked at
predict.glmmPQL {MASS}
predict.lme {nlme}
and neither of them offer exactly what I need, and I just don't know
enough to go from what they do offer to what I need. I have Pinheiro
& Bates as well as MASS, and both come tantalizingly close to what I
need, but I can't figure out the next step.
AFAIK, there's nothing of the sort for aov(. ~ ., Error(...) ...)
either.
In order to publish results of designed psychological experiments
(most of which are of the classic anova variety, with all the
predictors being factors), we need to plot error bars on our
interaction plots. I suspect (from the discussion on this list) that
my colleagues are using SPSS or SAS, and reporting incorrect CIs.
Not being a statistician I have hit a wall here. I'm not sure if the
transition to the results I need is staring me in the face, and I
don't know enough to take the next step, or if the tools aren't yet
available. In any event, I would be very grateful to you for guidance
on how to proceed.
_____________________________
Professor Michael Kubovy
University of Virginia
Department of Psychology
USPS: P.O.Box 400400 Charlottesville, VA 22904-4400
Parcels: Room 102 Gilmer Hall
McCormick Road Charlottesville, VA 22903
Office: B011 +1-434-982-4729
Lab: B019 +1-434-982-4751
Fax: +1-434-982-4766
WWW: http://www.people.virginia.edu/~mk9y/
