Re: [R] False convergence of a glmer model

2010-02-16 Thread Douglas Bates
On Tue, Feb 16, 2010 at 9:38 AM, Shige Song wrote: > Hi Doug, > > Thanks. Next time I will post it to the R-SIG0-mixed-models mailing > list, as you suggested. I have added R-SIG-mixed-models to the cc: list. I suggest we drop the cc: to R-help after this message. > With respect to your questio

Re: [R] False convergence of a glmer model

2010-02-16 Thread Shige Song
Hi Doug, Thanks. Next time I will post it to the R-SIG0-mixed-models mailing list, as you suggested. With respect to your question, the answer is no, these parameters do not make sense. Here is the Stata output from "exactly" the same model: . xi:xtlogit inftmort i.cohort, i(code) i.cohort

Re: [R] False convergence of a glmer model

2010-02-16 Thread Douglas Bates
On Tue, Feb 16, 2010 at 9:05 AM, Shige Song wrote: > Dear All, > I am trying to fit a 2-level random intercept logistic regression on a > data set of 20,000 cases.  The model is specified as the following: >  m1 <- glmer(inftmort ~ as.factor(cohort) + (1|code), family=binomial, data=d) > I got

[R] False convergence of a glmer model

2010-02-16 Thread Shige Song
Dear All, I am trying to fit a 2-level random intercept logistic regression on a data set of 20,000 cases. The model is specified as the following: m1 <- glmer(inftmort ~ as.factor(cohort) + (1|code), family=binomial, data=d) I got "Warning message: In mer_finalize(ans) : false convergence (8)

[R] False convergence of a glmer model

2010-02-16 Thread Shige Song
Dear All, I am trying to fit a 2-level random intercept logistic regression on a data set of 20,000 cases. The model is specified as the following: __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the pos