(statistics) To remove the effect of a third variable from a correlation or regression analysis so the pure relationship between two other variables can be measured.
"The researchers partialled out the effect of age before examining the link between stress and memory."