Partial correlation controlling one variable
Residualize a three-variable correlation summary algebraically.
Your entries stay in this page’s memory and are not submitted to us. Downloads are generated locally. Reloading or leaving the page clears your entries. Do not enter patient information.
All correlations must lie strictly inside −1 and 1 and jointly define a positive-definite correlation matrix. Use correlations from the same complete observations.
The starting values are hypothetical. Results update as you edit. Use aggregate or synthetic inputs only.
Calculation results
- Partial correlation XY given Z
- 0.54901081
How to interpret the output
The result is (rXY−rXZ×rYZ)/sqrt((1−rXZ²)(1−rYZ²)). It measures linear residual association after one covariate. Conditioning can introduce bias when the covariate is a collider or measured with error; this calculation does not establish causality.
Method referenceSource checked 2026-11-06. Read the reproducible calculator validation report.