I have been looking for a package to do commonality analyses in a mixed model.
In your PeerJ article, you said “The package does not report partial R2 and commonalities, although they could be calculated from the partR2 output.”. I am now trying to use the formulae you gave in Figure 1 of the article.
I would greatly appreciate it if you can double-check my calculation below.
From the biomass example, here are the Part R2 results from the package.
Thank you so much for creating a neat package!
I have been looking for a package to do commonality analyses in a mixed model.
In your PeerJ article, you said “The package does not report partial R2 and commonalities, although they could be calculated from the partR2 output.”. I am now trying to use the formulae you gave in Figure 1 of the article.
I would greatly appreciate it if you can double-check my calculation below.
From the biomass example, here are the Part R2 results from the package.
> R2 (marginal) and 95% CI for the full model:
> R2 CI_lower CI_upper nboot ndf
> 0.5133 0.4136 0.5889 100 4
>
> ----------
>
> Part (semi-partial) R2:
> Predictor(s) R2 CI_lower CI_upper nboot ndf
> Model 0.5133 0.4136 0.5889 100 4
> SpeciesDiversity 0.1729 0.0322 0.2792 100 3
> Temperature 0.3058 0.1842 0.3966 100 3
> Year 0.0140 0.0000 0.1362 100 3
> SpeciesDiversity+Temperature 0.4916 0.3918 0.5663 100 2
> SpeciesDiversity+Year 0.1862 0.0475 0.2913 100 2
> Temperature+Year 0.3276 0.2091 0.4155 100 2
> SpeciesDiversity+Temperature+Year 0.5133 0.4136 0.5889 100 1
To compute commonalities, here what I did:
Commonalities (SpeciesDiversity & Temperature): Part (SpeciesDiversity+Temperature) – Part (SpeciesDiversity) – Part (Temperature) = 0.4916 - 0.1729 - 0.3058 = 0.0129
Commonalities (SpeciesDiversity & Year): Part (SpeciesDiversity+Year) – Part (SpeciesDiversity) – Part (Year) = 0.1862 - 0.1729 - 0.0140 = -0.0007
Commonalities (Temperature & Year): Part (Temperature+Year) - Part(Temperature) – Part(Year) = 0.3276 - 0.3058 - 0.0140 = 0.0078
Commonalities (SpeciesDiversity & Temperature & Year): Part (SpeciesDiversity+Temperature+Year) – Part (SpeciesDiversity+Temperature) - Part (SpeciesDiversity+Year) - Part (Temperature+Year) + Part (SpeciesDiversity) + Part (Temperature) + Part (Year) = 0.5133 - 0.4916 - 0.1862 - 0.3276 + 0.1729 + 0.3058 + 0.0140 = 0.0006
Does it sound reasonable?
If so, I have two questions: 1) How do we obtain confidence intervals for the commonalities? 2) What does it mean when we have a negative commonality.
Thank you again,
Narun