This is my first post, so I hope I’ve posted within the correct category. Please let me know if my post is more suited for a different board.
I’m still very new to statistic, in fact I’m still a student, so my question is probably very much surface level for many of you.
I took a test, and one of the questions asked:
Suppose you deleted several observations with large residual values. If you re-estimate the regression equation using this reduced sample, then what would likely happen to the standard error of the estimate (SEE) and the R-squared?
The correct answer was that both decrease. I, however, though that R-squared would increase because fewer residual values would mean that the remaining observation would be fitted closer to the regression line. Can someone help me understand why I was wrong?