s-news
[Top] [All Lists]

Re: overdispersion in logistic regression

To: mtexeira@agro.uba.ar
Subject: Re: overdispersion in logistic regression
From: Tim Hesterberg <timh@insightful.com>
Date: 13 Jul 2005 17:24:39 -0700
Cc: s-news@lists.biostat.wustl.edu
In-reply-to: <1121105272.42d2b5783d0a6@webmail3.agro.uba.ar> (mtexeira@agro.uba.ar)
References: <1121105272.42d2b5783d0a6@webmail3.agro.uba.ar>
>Everybody knows how can I correct for overdispersion in a logistic regression 
>problem ?
>
>thanks
>Marcos Texeira

I would typically estimate parameters ignoring the over-dispersion,
then specified the dispersion when calling summary().
I don't recall if I estimated the dispersion myself, or passed
dispersion = 0.
See help(summary.glm)

Tim Hesterberg


<Prev in Thread] Current Thread [Next in Thread>