Is this a common formula to find pseudo R-squared for this GLM?

Asked Dec 18 '16 at 16:56

Active Dec 18 '16 at 17:23

Viewed 346 times

I'm trying to figure out if this common pseudo R-squared formula is acceptable for the GLM I am running. Here is the code and formula:

m1<- glm((AdjustedActivity3+1)~HDtype, family = Gamma(link = log), data = q2)

McFaden's pseudo r2 formula:

1-(m1$deviance/m1$null.deviance)

I wonder: does the Gamma family have a specific formula for an R-squared value in this case, or is McFaden's usage generally accepted?

edited Dec 18 '16 at 17:23

smndpln

asked Dec 18 '16 at 16:56

Greg Hill

This Q&A http://stats.stackexchange.com/questions/11676/pseudo-r-squared-formula-for-glms?rq=1 has many useful links to resources for and against pseudo $R^2$. – mdewey Dec 18 '16 at 18:07

0 Answers0