Message-ID: <41ACECDB.1000201@stat.wisc.edu>
Date: 2004-11-30T21:57:47Z
From: Douglas Bates
Subject: Random effects and non-linear models
In-Reply-To: <33A19A6E6AC77141A5148EABFC80830BA525F6@email.mbs.tamu.edu>
Certo, Trevis wrote:
> I am trying to model growth over time for a number of companies (my unit
> of analysis). My dependent variable is a count (the number of
> acquisitions each firm made each year), and I'm having some trouble
> figuring out exactly how to model this. My assumption is that modeling
> the effect of time on this dependent variable with a linear model is
> inappropriate. Does anyone have any guidance regarding how to model such
> a dependent variable such that I could include both random and fixed
> effects? Thanks in advance.
You may want to consider a generalized linear mixed model (GLMM) for the
Poisson family. These can be fit with the GLMM function in the lme4
package.