[R-sig-ME] lmer and overdispersion

[Ben Bolker]

  Please don't e-mail me personally: please e-mail the
r-sig-mixed-models help list ...

  It doesn't make sense to add overdispersion to a linear mixed model
(i.e. one with normally distributed
responses).  See
http://article.gmane.org/gmane.comp.lang.r.lme4.devel/6426 for example
...

On Sat, Aug 6, 2011 at 11:53 AM, Ahmad Rabiee <ahmadr at sbscibus.com.au> wrote:
> Dear Ben
>
>
>
> I’ve got a dataset (see attached) and I would like to run a mixed model
> logistic regression. I would like to account for overdispersion in this
> dataset.
>
>
>
> I checked the examples in “lme4 package manual”, but could understand how
> you did this? Below is my R codes, I would appreciate if you can help, what
> I need to do here to account for overdispersion in the following models?
>
>
>
> Your help and comments would be greatly appreciated.
>
> Ahmad
>
>
>
> #--------------------------------------------------
>
> library(lme4)
>
> setwd("G:/Data")
>
> ket <- read.table("z-score BHB data.csv", header=T, sep=",", na.string="NA")
>
>
>
> # Model 1: Country as random effects term
>
> ket.1a <- lmer(ketosis ~ z_bhb + bcs_pre + lact + twins + (1|country), ket)
>
>
>
> # Model 2: herdno & Country as random effects terms
>
> ket.1b <- lmer(ketosis ~ z_bhb + bcs_pre + lact + twins + (1|herdno) +
> (1|country), ket)
>
> ket.1b
>
>
>
> # Model 3: herdno nested within country
>
> ket.1c <- lmer(ketosis ~ z_bhb + bcs_pre + lact + twins + (herdno|country),
> ket)
>
> ket.1c
>
>
>
>
>
>
>
> "Try not to become a man of success, but rather try to become a man of
> value"
> Albert Einstein
>
>
>
> Please note my new email address is mailto:ahmadr at sbscibus.com.au. Please
> update your records.
>
>