[R] Help on glmmPQL

[Lorenz.Gygax at fat.admin.ch]

See help file of glmmPQL (the first thing you should be doing):

glmmPQL(fixed, random, family, data, correlation, weights,
        control, niter = 10, verbose = TRUE, ...)
                 ^^^^^^^^^^
Arguments
...
niter maximum number of iterations.  
...

Thus, you can increase the value of niter to say 25. In my experience it is likely that the model fit is problematic if 10 iterations are not enough. Perhaps if you are lucky you were close and only some few more iterations are needed. If that is not the case, you may have too few occurrences of your response or too many explanatory variables.

Regards, Lorenz


> -----Original Message-----
> From: r-help-bounces at stat.math.ethz.ch
> [mailto:r-help-bounces at stat.math.ethz.ch]On Behalf Of 
> Francisco Redelico
> Sent: Wednesday, May 31, 2006 4:54 PM
> To: r-help at stat.math.ethz.ch
> Subject: [R] Help on glmmPQL
> 
> 
> Dear R-Users,
> 
> I am trying to fit a generalized linear mixed models
> using glmmPQL, 
> 
> mymodel.glmm = glmmPQL(efect~ time+ cov1+ cov2,
>                     random = ~ fact1+ fact2+ fact3+
> fact1*time + fact2*time+ fact3*- cov1 - cov2- time|
> subject, family=poisson)
> 
> I get this error message: 
> 
> Error in lme.formula(fixed = zz ~ time + cov1 + cov2 ,
> random = ~fact1 +  : 
> iteration limit reached without convergence (9)
> 
> I used
> 
> meControl( maxIter = 10000, msMaxIter=10000,
> tolerance=1e-4, nlmStepMax=10000, msVerbose = TRUE)
> 
> But it does not work.
> 
> Could you tell me what is the error about? What does
> the number in brackets mean?
> 
> Thanks in advance
> 
> Francisco