[R] Mixed models negative binomial, Error in eval(expr, envir, enclos)

[Ben Bolker]

Karina Charest Castro <jlonkc <at> gmail.com> writes:

> 
> Hi!

  This question is more appropriate for r-sig-mixed-models at r-project.org.
Please repost there (I will add a few questions/comments below that
you should probably address when you repost)

> I am trying to do a glmer.nb but get this error:
> Error in eval(expr, envir, enclos) :
>   ..2 used in an incorrect context, no ... to look in
> 
> My data structure is data.frame
> Here my data set: d1_2
> 
> Year ID Age Reproductive_status Rank_Residuals Asso.Y1 TotalY1
> 1994 109 8 Lactating 0.23947902 9 48
> 1994 110 6 Barren -0.030977951 56 62
> 1995 1000 9 Summer_yeld 0.08624 31 55
> 1995 103 4 Barren -0.06861 29 52
> 1996 23 8 Barren -0.015726749 64 64
> 1997 1000 11 Lactating 0.148665859 5 35
> 1997 103 6 Barren -0.116539886 51 51
> 2001 105 9 Lactating -0.036222842 27 28
> 2001 106 9 Barren 0.073777158 40 40
> 2001 116 8 Barren -0.015710059 5 31

  is this your whole data set, or just the first few lines?
(If the former, then you won't be able to fit the model you want --
there are not nearly enough data.)

> library(lme4); library(MASS)
> nb.glmer <- glmer.nb(Asso.Y1 ~ Reproductive_status * Age +
> Reproductive_status * Rank_Residuals + Reproductive_status * I(Age^2) +
> (1|ID) + (1|Year), offset = log(d1_2$TotalY1), data=d1_2)

  The fixed-effect part should be identical to:

~ Reproductive_status*(Rank_Residuals+poly(Age,2,raw=TRUE))


> 
> I have also tried because I found that glmer.nb is not well understood:
> library(glmmADMB)
>     random <- MCMCglmm(cbind(Asso.Y1,TotalY1-Asso.Y1) ~ Reproductive_status
> * Age + Reproductive_status * Rank_Residuals + Reproductive_status *
> I(Age^2), random = ~ ID+Year, data = d1_2)
> BUT: Error in MCMCglmm(cbind(Asso.Y1, TotalY1 - Asso.Y1) ~
> Reproductive_status *  :
>   family must have the same length as the number of responses
> 
> I am not able to solve those problems and I have looked on forums too.
> 
> Tkx in advance!

  You're loading the glmmADMB package but calling MCMCglmm, which
is a completely different function from a completely different package,
with different syntax ...