Dear all,
I have a rather basic question and apologize in advance because of it.
I am using a panel data Poisson regression, my dependent variable is the number of disease incidents per country per year and I have several covariates. Since different countries with different population sizes can have a different number of diseases, I control for the population of the countries. But I am aware that I can introduce the population size as an exposure variable. My question is that if I enter population simply as a covariate should I get similar results as when I enter population as an exposure variable?
And also, if I simply divide the number of diseases in each country (to have a per capita dependent variable) and then use OLS regression, should I get similar results as when I use a Poisson regression but with a count dependent variable (and not a per capita and hence a noninteger variable)?
Thanks,
Marzieh
0 Response to Question about comparing the results of OLS regression and Poisson regression
Post a Comment