Partial / Rank correlation documentation

mattansb commented 4 years ago

I think it should be made more explicit (especially for the Bayesian methods) how Partial and Rank correlation are actually obtained.

These are not "true" Bayesian rank correlations (which have some weird priors that I don't understand), but x & y a rank transformed and then these are tested for a Pearson correlation.
And these are not "true" Bayesian partial correlations (no prior for the correlation between X<->Z or Z<->Y), but Z is partial-ed out of x & y using OLS, and then these residual-scores are tested for a Pearson correlation.

DominiqueMakowski commented 4 years ago

yes! that package definitely needs a proper documentation and articles...

Regarding these two points, should we call them "pseudo-Bayesian" correlations?

mattansb commented 4 years ago

Hmm.. The correlation is Bayesian, it's the "method" isn't.. Maybe "Bayesian pseudo-Spearman / pseudo-partial correlation (see details)"?

DominiqueMakowski commented 4 years ago

haha good point

easystats / correlation