Specifying random effects in lme()

Wed, May 9, 2007 10:40 AM

On Wed, 2007-05-09 at 16:58 +0200, Nilsson Fredrik X wrote:

Hi Nilsson,

Many thanks for this. I had actually tried this but when I tried coef()
on the resulting model, I got what looked like only random effects for
the log.cl parameter. Now that I have looked at the resulting object
more closely after your email, I have noticed that the magnitudes of the
random effects for log.so4 are very small indeed, and so when coef()
prints the values for log.so4, they do all look constant, but only
because they are printed to finite decimal places.

I should look more closely next time!

Many thanks, I now have some indication that the effect of log.cl does
vary between sites, but not log.so4.

All the best,

G

I made a script which is able to do so, but perhaps the addition of
this extra variability in your case over-specifies the model? 

Hope this helps. Otherwise I suppose that you have to supply more details.

Cheers,

Fredrik
--------------------------------------------------------------------------
Example:

nsite<-10
nwithin<-10
prosw<-nsite*nwithin

conc<-vector(length=prosw)
temp<-rnorm(prosw)
log.so4<-seq(-5,5, length=nwithin)
log.cl<-rnorm(prosw)
int<-17
acl<-sqrt(2)
aso4<-21
atemp<- -14
rho<--0.9
for (i in 1:nsite)
{
sf<-0.1*rnorm(1)
sc<-2*rnorm(1)
ss<-rnorm(1)
err<-0.1*rnorm(1)
for (j in 1:nwithin)
{
conc[nwithin*(i-1)+j]<-sf +  int + (acl+sc)*log.cl[nwithin*(i-1)+j] + (aso4+ss)*log.so4[j] +
                  atemp*temp[nwithin*(i-1)+j] + err
err<-err*rho + 0.1*rnorm(1)
}
}

temp<-rep(temp,nsite)
log.so4<-rep(log.so4, nsite)
log.cl<-rep(log.cl,nsite)
site<-rep(seq(1,nsite), each=nwithin)

DF<-data.frame(conc,temp, log.so4, log.cl, site)
DF$site<-as.factor(site)

DFgd<-groupedData(conc~temp|site, data=DF)
plot(DFgd)

DF.lme<-lme(conc~log.so4+temp+log.cl, random=~log.so4+log.cl|site, data=DF)
plot(ACF(DF.lme, resType="n", form=~1|site, maxLag=floor(nwithin/2-.5)), alpha=0.05)
DF2.lme<-update(DF.lme, correlation=corAR1(form=~1|site))



-----Ursprungligt meddelande-----
Fr?n: r-sig-mixed-models-bounces at r-project.org [mailto:r-sig-mixed-models-bounces at r-project.org] F?r Gavin Simpson
Skickat: den 9 maj 2007 15:24
Till: R-SIG-Mixed-Model
?mne: [R-sig-ME] Specifying random effects in lme()

Dear List,

I am modelling repeated measures data from 11 sites in the UK using
lme() from the nlme package. I have the following call for a random
intercept model with an AR1 error structure:

mod <- lme(doc ~ temp + log.so4 + log.cl, random = ~ 1 | site,
           correlation = corAR1(), data = hydro)

I would like to fit a model that has random effects for log.so4 and
log.cl within site, so that I am fitting a random slop and intercept
model --- we want the effect of log.so4 and log.cl to vary between sites
and to compare the fits of the two models.

If I am following lmer() in lme4 correctly, I believe I could fit this
model as:

hydro.lmer <- lmer(doc ~ temp + log.so4 + log.cl +
                   (log.so4 | site) + (log.cl | site),
                   data = hydro)

but without the AR1 correlation.

I am struggling to get the syntax correct for random in lme() to add
random effects to log.so4 and log.cl. I have only managed to get a model
where one of the coefficients (log.cl) also has a random effect.

I would be most grateful if someone could explain how to fit the model
in lme() by showing me the correct form for the random argument.

Many thanks,

Gav

%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%
 Gavin Simpson                 [t] +44 (0)20 7679 0522
 ECRC, UCL Geography,          [f] +44 (0)20 7679 0565
 Pearson Building,             [e] gavin.simpsonATNOSPAMucl.ac.uk
 Gower Street, London          [w] http://www.ucl.ac.uk/~ucfagls/
 UK. WC1E 6BT.                 [w] http://www.freshwaters.org.uk
%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%

Specifying random effects in lme()

Thread (3 messages)