Response is Win %

I.

1

2 44.1

3

50.8

4

61.3

5

63.2

64.4

65.4

65.5

II.

6

7

8

42.3

44.0

39.5

III.

IV.

V.

R

+ @ 20

34.9

30.4

Vars R-Sq R-Sq (adj) R-Sq (pred) Mallows Cp

21.4

18.7

12.3

24.4

40.3

32.5

11.2

45.5

34.5

8.8

55.5

3.8

0.13781 X

X

X

56.2

4.5 0.13680 X X

X

XX

55.9

5.7 0.13723 X

X

X

X X X

55.3

7.1 0.13812 X

X

X X X

X X

X X X X X X X

53.5

9.0

0.14093 X

a

t

2

4 SI

00 CN

+

Sek TRRR

2

0.18629 X

0.15969 X

0.15249 X

D+ON

X

X

0

4 A

0t

t

X

2. After a recent NFL season, we record for the 32 teams their Win % and eight other variables. Above is

Best Subsets Regression output from Minitab for regressions of the form: Win % vs. combinations of the

listed predictor variables. Based on this table and the criterion of Best Subsets regression, which of the

statements below is/are true?

X

Win% vs. Rate is the best simple regression model.

Win % vs. 20+ is the second-best simple regression model.

In the best 3-predictor model, 40 +R has a t score greater than 1 in absolute value.

The best model uses 5 predictors.

20 +R is used as a predictor in the best model.

Briefly explain your answer for each of the statements you choose as true.