[R-sig-ME] profile and/or bootstrapped CIs for GAMM random effects

Ben Bolker bbolker at gmail.com
Wed Sep 2 04:18:39 CEST 2015

Hash: SHA1

  This looks like it could get ugly, we might have to call on Simon
Wood.  I've posted it as https://github.com/lme4/lme4/issues/333 , but
I think it's really a gamm4 problem or at least a gamm4/lme4
communication problem.  The part of the problem I've been able to
identify so far is that the @call slot of the $mer part of the gamm4
object does *not* look like a normal lme4 call, which causes problems
(much farther down) where we use update() to try to reconstruct a
version of the model that is just a deviance function.  Getting a
deviance function back from a glmerMod object in a more
straightforward way is something that's been on our list ...

  Ben Bolker

On 15-09-01 06:51 PM, John Morrongiello wrote:
> Hi all I thought I'd re-post this message from last week as it may
> have got lost in the weekend traffic:
> I'm having trouble estimating CIs for random effects (i.e. CODE)
> from a gamm4 model fit with a gamma distribution, My model is:
> c1<-gamm4((longdist)~s(time,k=4),random=~(1|CODE),data=catfish1,family=Gamma(link=log))
> ##this returns Error in mkMerMod(rho = environment(devfun), opt =
> opt, reTrms = b$reTrms,  : unused argument (devFunOnly = TRUE)
> confint(c1$mer,method="boot") ##this returns Error in
> mkNewReTrms(object, newdata, compReForm, na.action = na.action,  : 
> random effects specified in re.form that were not present in
> original model Are these errors arising because I'm using a gamma
> distribution? Is there another way to get the random effect CIs?
> I can get random effect CIs for the following model: 
> c2<-gamm4((longdist)~s(time,k=4),random=~(1|CODE),data=catfish1,family=gaussian)
>  confint(c2$mer,method="profile")##works 
> confint(c2$mer,method="boot")##same error message as above
> Thanks for your time John
> -- Dr. John R. Morrongiello School of BioSciences University of
> Melbourne Victoria 3010, Australia T: +61 3 8344 8929 M: +61 403
> 338 554 E:
> john.morrongiello at unimelb.edu.au<mailto:%20jmorrongiell at unimelb.edu.au>
W: morrongiellolab.com<http://morrongiellolab.com/>
> [[alternative HTML version deleted]]
> _______________________________________________ 
> R-sig-mixed-models at r-project.org mailing list 
> https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models

Version: GnuPG v1.4.11 (GNU/Linux)


More information about the R-sig-mixed-models mailing list