Final answer:
Deleting outliers is the best approach when there is multicollinearity and a high R2 in the estimated model.
Step-by-step explanation:
Outliers are data points that are located farther than two standard deviations above or below the best-fit line. By deleting these outliers and recalculating the best-fit line and correlation coefficient, we can determine if they are influential points that significantly affect the correlation and fit of the model.