How can I compute Marginal Effects in a Negative Binomial Model??
Posted: Thu Aug 05, 2021 5:36 am
In testing my Poisson regression (using the Eviews guide: http://www.eviews.com/help/helpintro.ht ... %23ww38558) I can see overdispersion, so am switching to a negative binomial distribution.
I am comfortable with estimating marginal effects for my Poisson model using the basic formula exp(c(1)+c(2)x)*c(2). However, I understand the negative binomial distribution is a poisson/gamma mix and I am not clear and can't find any guidance on how to approach estimating expected values or marginal effects with this distribution.
Grateful for any advice!!!
Thanks in advance.
Tymon
I am comfortable with estimating marginal effects for my Poisson model using the basic formula exp(c(1)+c(2)x)*c(2). However, I understand the negative binomial distribution is a poisson/gamma mix and I am not clear and can't find any guidance on how to approach estimating expected values or marginal effects with this distribution.
Grateful for any advice!!!
Thanks in advance.
Tymon