Partial Correlation Using Covariance vs OLS methods?

[adam2392]

For the partial correlation function in pingouin, I am wondering what the differences are wrt to the method of estimating partial correlation by:

• fitting OLS of Z on X
• fitting OLS of Z on Y
• comparing the two residuals to compute the partial correlation of X and Y given Z

? The function seems to not do this, and I'm wondering if we lose anything or gain anything for doing either procedure.

[raphaelvallat]

Hi @adam2392,

Pingouin's partial correlation used to rely on the residuals of linear regression. We changed it in Pingouin 0.4.0. Copy-pasting the changelog below:

Major refactoring of pingouin.partial_corr(), which now uses the same method as the R ppcor package, i.e. based on the inverse covariance matrix rather than the residuals of a linear regression. This new approach is faster and works better in some cases (such as Spearman partial correlation with binary variables, see issue 147). One caveat is that only the Pearson and Spearman correlation methods are now supported in partial/semi-partial correlation.