s-news
[Top] [All Lists]

Re: Generating overdispersed poisson data

To: s-news@lists.biostat.wustl.edu
Subject: Re: Generating overdispersed poisson data
From: Gregor Gorjanc <gregor.gorjanc@bfro.uni-lj.si>
Date: Tue, 31 Jul 2007 21:10:01 +0000 (UTC)
References: <0552C65FE0AE49458A43BB6538148C920640AD@SAGE.forestry.oregonstate.edu>
User-agent: Loom/3.14 (http://gmane.org/)
Huso, Manuela <manuela.huso <at> oregonstate.edu> writes:
> I am carrying out some simulations in which I need to generate overdispersed
poisson data.  The rpois function only takes sample size and mean as its
arguments.  What do I need to do to create overdispersed poisson data?  I
apologize if this is a naïve question.  A search of the archives revealed
nothing on this subject.

Hi!

First sample different lambdas and then use rpois i.e. your model is
something like

y ~ Poisson(lambda)
log(lambda) = \mu + e
e ~ Normal(0, sigma^2)

Gregor


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