The protest
function in vegan returns a "Correlation in a symmetric Procrustes rotation". My assumption is that it evaluates the linear relationship between two sets of continuous variables (i.e. coordinates), and is therefore a Pearson correlation (rather than Spearman). Is this correct, or is it something more nuanced?
Although the documentation does not explicitly state whether vegan::protest
uses Pearson or Spearman correlation, by doing some testing we can see that vegan::protest
uses Pearson correlation
cor(iris[c("Sepal.Length", "Sepal.Width")], method="pearson" ) # gives -0.1176
cor(iris[c("Sepal.Length", "Sepal.Width")], method="spearman" ) # gives -0.1668
library(vegan)
protest(iris[,c("Sepal.Length")], iris[,c("Sepal.Width")]) # gives 0.1176