Is anyone aware of tools to perform power analyses on distance-based multivariate methods such as
Permanova
ANOSIM
dbRDA
etc...?
I was using the pwr library in R today and it struck me that I dont know how to do a multivariate PA.
The earth is round: P<0.05
Now that I think of it, are there functions that deal with factorial and unbalanced designs?
the function:
deals with balanced designs. There was a brief discussion about this in stackoverflow but nothing came of it. Has anyone had to deal with this?Code:pwr.anova.test(k=2,sig.level=0.05,f=0.25,power=0.8)
The earth is round: P<0.05
Well you can always do simulations to do your power analysis. I like that approach too because it makes it easier to assess what kind of sample size you would need even in the face of a broken assumption.
"His programming is malfunctioning. It begins! Get your weapons, he's going to become a killbot!!!" - bryangoodrich
Dason, thanks. Can you perhaps point me towards some more info about simualtions? It was (briefly) metioned on stackoverflow.
The earth is round: P<0.05
Posting Permissions
|
|
Reply With Quote
