[ADMB Users] Incorporating more than one random effect into glmmADMB

Lena Lorenz lena.m.lorenz at gmail.com
Thu Jun 23 05:47:44 PDT 2011


I have a model with the following structure:
The response variable consists of mosquito counts with a strong
negative binomial distribution. Explanatory variables are two fixed
effects and their interaction, and two random effects (which have to
be specified as group="" to specify random intercepts in the glmmADMB
model rather than as random=~"").
I.e. I need a model such as:

model <- glmmADMB(y ~ treatment*month, random= ~1, group = "day",
group = "hut", data = XX, family = "nbinom").

Can this be run, or do you have any suggestions how to solve this
problem? Unfortunately I work on Windows, so I cannot take advantage
of the new alpha version that I saw on one website.

Any ideas of how to use the glmmADMB model with two random effects, or
an alternative for analysing negative binomial data with a mixed
effects model would be greatly appreciated!

Thank you very much,

Lena Lorenz

More information about the Users mailing list