Comment on “A New Measure of Predictor Variable Importance in Multiple Regression”