The partial correlation between the dependent variable and thek-th independent variable in a regression can be calculated simply from thek-th regression coefficient and its standard error. Partial correlations are useful in determining relations among the estimates obtained under alternative assumptions about which variables in a regression are independent of the disturbances.