The title says it all. I'm wondering if someone can help me understand the difference (if there even is one) between a Quasi-Poisson model and fitting a Poisson Regression Model using GEE? It is my understanding that both of these methods can be effective ways for handling over/under-dispersion and they seem similar, but I can't tell if they are the same thing or not.

I'm really surprised no one has addressed this? As often as I see the terms quasi-poisson and GEE used in the literature, maybe I'm not the only one confused by the seemingly dual usage of these terms?

  • $\begingroup$ "As often as I see the terms quasi-poisson and GEE used in the literature" -- can you provide a reference/example? $\endgroup$ – Jon Oct 21 '16 at 17:44

Quasi-poisson GLM is a special case of Poisson GEE.

The specification of GEE (copied from wikipedia) is that

$$U(\beta)=\sum_i \frac{\partial\mu_{ij}}{\partial \beta_k} V_i(Y_i-\mu_i(\beta))$$

where $V_i$ is the variance of observation $i$. In the case that these are a constant multiple of the (standard dispersion) poisson mean, this reduces to a poisson GLM with a dispersion parameter. However, the GEE is more general than that; it can cope when there is additional variance structure on top of this.

| cite | improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.