Another PGLS in R question. I really like the caper package, as I think it's the most intuitive for PGLS analyses. I'd like to include a factor variable in my models, namely here trophic level (herbivore, omnivore, carnivore). When I've run the pgls function with factors with only two levels (e.g., terrestrial versus arboreal), the anova function prints out a nice anova table telling me the general significance of that factor. However, when I try and use anova on a pgls object with this three-level factor, I get the following error message:
> full=pgls(zscore~diet+lifespan+trophic,data=cdat,lambda="ML")
> anova(full)
Error in terms.formula(formula, data = data) :
invalid model formula in ExtractVars
Has anyone else had this problem when including a 2+ level factor in PGLS? I did some google digging, but nothing has really turned up. If caper just doesn't like these kind of categorical variables, has anyone had better luck in another package?
Thank you
I think I found the problem.
Try using the attached function instead to perform the anova.
Let me know if it works for you!
I am having this exact same problem!
Have you found any solutions, yet?
I think I found the problem.
Try using the attached function instead to perform the anova.
Let me know if it works for you!
You're welcome!
The problem was that the pgls breaks multi-state categorical variables into a number of binary variables (like shown by the summary function).
This increases the number of variables that the anova then tries to iterate over, but there are fewer variables in the formula, so it hits an error where it doesn't know what other variables to include. I fixed this by having it iterate over the number of variables in the formula rather than the number of binary variables.
Dear William Gearty,
I had the same problem running anova in a pgls model with categorical variable with more then two levels. I used your function and it solved the problem!
Kind regards,
Gustavo
@Leandro Aristide we've coded lifespan as a continuous covariate (maximum lifespan measured in months) using data from PanTHERIA
- Badges
- Science topic
- Similar topics
- Biological Science
- Evolutionary Biology
- Phylogenetics