Interpret scale parameter when estimating dispersion in Poisson model

[ashdederich]

Hi,

I'm working on checking dispersion for my poisson models. I saw in R documentation that a summary of the phyloglm poisson model gives me a scale estimate, which is the scale parameter to estimate overdispersion.

For one of my models, I have a scale of 0.2704646 and I am unsure how to interpret it. I know that an evenly dispersed Poisson model needs to have a dispersion close to 1. However, since scale is measuring overdispersion I'm unsure if it should be close to 0 or not. How do I interpret the scale parameter, and is there a way to test if the over/underdispersion is significant or not? I was not able to use the dispersiontest() function from the AER CRAN package because it is not a glm object that the dispersiontest() function recognizes.

Additionally, are there built-in ways in phylolm for correcting for over/underdispersion?

Thank you for such a great package! Ashley

[lamho86]

Hello Ashley,

The scale is the dispersion parameter of the Poisson model. Currently, there is no built-in function to correct for over/under-dispersion in phylolm.