In: Statistics and Probability
The residuals for 15 consecutive time periods from a simple linear regression with one independent variable are given in the following table.
Time_Period Residual
1 +4
2 -6
3 -1
4 -5
5 +3
6 +6
7 -3
8 +7
9 +7
10 -3
11 +2
12 +3
13 0
14 -5
15 -7
B) Compute the Durbin-Watson statistic. At the 0.05 level ofsignificance, is there evidence of positive autocorrelation among the residuals?
The Durbin-Watson statistic is D=
(Round to three decimal places as needed.)
What are the critical values for this test?
dL=
dU=
(Round to three decimal places as needed)
At the 0.05 level of significance, is there evidence of positive autocorrelation among the residuals?
A.Yes comma because the value of Upper D is less than nbspYes, because the value of D is less than d Subscript Upper LdL.
B.No, because the value of D is greater than d Subscript Upper LdL.
C.Yes, because the value of D is less than d Subscript Upper UdU.
D.No comma because the value of Upper D is greater than nbspNo, because the value of D is greater than d Subscript Upper UdU.
E.The test is inconclusive.
c. Based on (a) and (b), what conclusion can you reach about the autocorrelation of the residuals?
A.There appears to be strong negative autocorrelation among the residuals.
B.There appears to be positive autocorrelation among the residuals.
C.There appears to be positive and negative autocorrelation among the residuals.
D.There does not appear to be autocorrelation among the residuals.
t | e | (et- e_t-1)^2 | e_t^2 |
1 | 4 | 16 | |
2 | -6 | 100 | 36 |
3 | -1 | 25 | 1 |
4 | -5 | 16 | 25 |
5 | 3 | 64 | 9 |
6 | 6 | 9 | 36 |
7 | -3 | 81 | 9 |
8 | 7 | 100 | 49 |
9 | 7 | 0 | 49 |
10 | -3 | 100 | 9 |
11 | 2 | 25 | 4 |
12 | 3 | 1 | 9 |
13 | 0 | 9 | 0 |
14 | -5 | 25 | 25 |
15 | -7 | 4 | 49 |
559 | 326 |
d = 559/326
= 1.715
k = 1
n = 15
dL = 1.08
dU = 1.36
d = 1.715 > dU(1.36)
hence there is no statistical evidence that the error terms are positively autocorrelated.
option D) is correct
c)
4-d = 4-1.715 = 2.285
since 4-d > dU
there is no statistical evidence that the error terms are negatively autocorrelated.
D.There does not appear to be autocorrelation among the residuals.