0

I am using a Tweedie GLM to analyse the effect of sex and location on site fidelity for a mark-recapture study. The data is heavily zero inflated as a lot of the animals were not recaptured, therefore I believe that Tweedie is the best fit. Tweedie acts as an add on to the glm() function from the statmod package.

My question concerns the output of my GLM using the summary() function in R. I understand that R chooses one of the levels from each category as baseline for the intercept though the output does not list all possible differences between locations.

At the moment the output I have is: P (Intercept) 0.35 SexM 0.00 LocationOH 0.00 LocationRF 0.04

I have excluded all other values for ease of reading. From my interpretation this would suggest that there is a significant difference in fidelity between males and females, there is a significant difference between OH and NH (listed under intercept as the baseline along with females) and there is a significant difference between RF and NH. However, there is no output to tell me if there is a significant difference between OH and RF. Is there some code that I can use to find this value?

Phil
  • 7,287
  • 3
  • 36
  • 66
Tom Johnson
  • 1
  • Apologies, the structure of the output was changed after posting. There are 3 locations (NH, OH, RF) and 2 sexes (M, F) and the number listed next to each category is the p value – Tom Johnson Jun 07 '23 at 21:33
  • don't have time to write out an answer, but check the `pairs()` function in the `emmeans` package: https://cran.r-project.org/web/packages/emmeans/vignettes/comparisons.html#pairwise – Ben Bolker Jun 08 '23 at 01:46

0 Answers0