[R-sig-ME] comparing results across software packages

[Ben Bolker]

Ben Bolker <bbolker at ...> writes:

> 
> Price, Emily <ep311508 <at> ...> writes:
> 
> > 
> > Dear R-sig-mixed-models,
> 

  [ MAJOR SNIPPAGE]

  Doug Bates points out that I meant to use glmer() and 
not glmmadmb() below:

> 
> (3) lme4 should also be able to handle this model, either with
> 
glmer(cbind(dv,trial-dv)~trt.f + classb.f + 
   (trt.f|id.f), 
  data=lam_hlm, family=binomial(link="logit"))
> 
> (unstructured variance-covariance matrix is the default) or

glmer(dv/trial~trt.f + classb.f + 
   (trt.f|id.f), weights=trial,
  data=lam_hlm, family=binomial(link="logit"))
 
> (I think glmmML can also do this one).
> 
>   Ben Bolker
> 
>