I’m analysing the contribution of environmental variables to characterize the community composition in different stream sites (I use a RDA in R). However, I’m also interested to know the contribution of these environmental variables in two groups of the data (e.g. forested and not-forested rivers).
It is possible to analyse the variance partitioning of this groups separately using the RDA of all the data? And…is it correct?
Is better if I do two separates RDA one for each group? (I was reticent to these option because I will have to remove spp due to I’m going to have more spp than sites)
Best regards,
Jordi
You must always analyse all of your data together so that you include all of the interactions :)
Thanks for your advice! It is also my first option. However, I don’t know how to know the % of explanation of the environmental variables for each “community type” in that case…
The ANOVA will do it for you by partitioning the sum of the squares I think :)
Could you get around this issue by including forested and non-forested rivers as a binary variable in the RDA? Can RDA handle discontinuous variables?
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Statistical Modeling
- Analyses of Variance