I would like to test the null hypothesis that there is no difference between the outcomes of the Spearmann correlation test run on pairs of data (x,y) from the original dataset I have , vs one where the y value may have been permuted or pairs of data (xi, yi) has been sampled with replacement (bootstrapping).
I have come across examples of calculating p values from permutating
Two-sided permutation test vs. two one-sided
And then also bootstrapping p values Computing p-value using bootstrap with R
and Get p-value of coefficients in regression models using bootstrap (for regression not correlation).
Presume I need to calculate a two-sided p value - but permutation using the correlation coefficient and bootstrapping examples above use the test statistic. Not sure which (if any ) is the most statistically valid way to test the null hypothesis above.
Any advice or pointers appreciated.
