Consider the following model:
vote A
=
33.2672 + 0.3425 prtystrA + 0.0349 expend A - 0.0349 expendB
(4.41678)
(0.08795)
(0.00337)
(0.00300)
n
=
173, R2
=
0.56866, SSR = 20901.67
We also have the following summary statistics for the variables
prtystrA
X Min. :22.00
X.1 1st Qu.:44.00
X.2 Median:50.00
X.3 Mean :49.76
X.4 3rd Qu.:56.00
X.5 Max. :71.00
expendA
Min. 0.302
1st Qu.: 81.634
Median: 242.782
Mean: 310.611
3rd Qu.: 457.410
Max. :1470.674
expendB
Min.: 0.93
1st Qu.: 60.05
Median: 221.53
Mean: 305.09
3rd Qu.: 450.72
Max. :1548.19
• Compute the average prediction when expendA = 310, expendB = 305, and prtystrA = 50.
Explain how you would transform the model to obtain the standard error. Once you have
the standard error, how would you contruct the confidence interval?
predict(res1, newdata=data.frame(expendA=310, expendB=305, prtystrA=50))
##
1
50.5678
• Do the same for expendA = 10, expendB = 1200, and prtystrA = 70
predict (res1, newdata=data.frame(expendA=10, expendB=1200, prtystrA=70))
##
1
15.68416
• Which prediction should have the smallest standard error? Explain.