[R-sig-ME] lme4, failure to converge with a range of optimisers, trust the fitted model anyway?

Hans Ekbrand hans.ekbrand at gmail.com
Sun Apr 5 12:14:20 CEST 2015


On Sun, Apr 05, 2015 at 11:33:26AM +0200, Ben Pelzer wrote:
> Dear Hans,
> 
> You could try Ben's suggestion for nAGQ = 5 in Stata. I believe that
> the routine"s "xtlogit" and "xtmelogit"  (in Stata 13 they are named
> differently, I don't remember the names right now)  are able do work
> with nested random effects and more than 1 quadr. point. These
> routines are however pretty time-consuming, may be even more so than
> glmer. Best regards,

OK. Even if will handle this case with aggregation, it is good to know
what stata can do for future challenges.



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