[R-sig-ME] family() in probit-glmer

[Ben Bolker]

   This does indeed look like a bug, now fixed
<https://github.com/lme4/lme4/commit/7f7350751ef1dfbd1598c24d4075fe18ea005860>
.  Thanks.

  In the meantime, you could also extract the family and associated
information via model at resp$family (although we do encourage everyone to
use the accessor methods when they work!)

  Ben Bolker


On 14-09-05 06:22 AM, Alessandro Moscatelli wrote:
> Dear all,
> 
> I have a problem using the function family(). I am using lme4 version 1.1-7
> running on R 3.1.0
> I fit a glmer model with a probit link function:
> 
> formula.mod = cbind(Longer, Total - Longer) ~ X * condition + (1 + X|
> Subject)
> mod1 <- glmer(formula = formula.mod, family = binomial(link = "probit"),
> data = datafr)
>> mod1
> Generalized linear mixed model fit by maximum likelihood (Laplace
> Approximation) ['glmerMod']
>  Family: binomial  ( probit )
> Formula: cbind(Longer, Total - Longer) ~ X * condition + (1 + X | Subject)
>    Data: datafr
>       AIC       BIC    logLik  deviance  df.resid
>  330.1026  348.8775 -158.0513  316.1026       101
> Random effects:
>  Groups  Name        Std.Dev. Corr
>  Subject (Intercept) 1.452885
>          X           0.001128 -0.99
> Number of obs: 108, groups:  Subject, 6
> Fixed Effects:
>  (Intercept)             X    conditionB  X:conditionB
>   -7.1760859     0.0090841    -0.2190119     0.0001522
> 
> However, using the family() function:
>> family(mod1)
> 
> Family: binomial
> Link function: logit
> 
> Does anyone has the same problem? Is it a bug? For Now I solved the problem
> using:
>> summary(mod1)$link
> [1] "probit"
> 
> Best
> Alessandro
> 
> 	[[alternative HTML version deleted]]
> 
> _______________________________________________
> R-sig-mixed-models at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models
>