[R-sig-ME] checking for overdispersion in lmer?
john.maindonald at anu.edu.au
Fri Aug 6 00:31:09 CEST 2010
The totals (input as weights if you are fitting proportions) need
to be supplied, with family=binomial. You need to model the
extra variance, with observation level random effects. If you
need an example, some of can supply one. I gave an example
for overdispersed Poisson in a Dec 12 message last year.
John Maindonald email: john.maindonald at anu.edu.au
phone : +61 2 (6125)3473 fax : +61 2(6125)5549
Centre for Mathematics & Its Applications, Room 1194,
John Dedman Mathematical Sciences Building (Building 27)
Australian National University, Canberra ACT 0200.
On 05/08/2010, at 11:37 PM, touchtonj wrote:
> How can I check for overdispersion in a lmer with proportion data? There is no residual deviance output like in a glm
> thanks ~ janeene
> R-sig-mixed-models at r-project.org mailing list
More information about the R-sig-mixed-models