In: Statistics and Probability
onsider the following data:
UBI POH TLAM
120, 9, 21
60, 5, 16
18, 3, 12
21, 5, 12
85, 7, 17
60, 7, 16
(If you want to check data entry: sample covariance UBI, POH = 61.67; UBI, TLAM = 108.39)
a. What is the least squares linear regression equation when UBI is the dependent variable (Y) are X
variables two?
b. Which of the coefficients, if any, are significantly different from zero at the 90% level?
a.
The dependent and independent variables are as follows:
UBI (Y) | POH(x1) | TLAM(x2) |
120 | 9 | 21 |
60 | 5 | 16 |
18 | 3 | 12 |
21 | 5 | 12 |
85 | 7 | 17 |
60 | 7 | 16 |
The following matrix and vector are defined in order to conduct the matrix calculation required to compute the estimated multiple regression coefficients:
Now, the vector with the estimated regression coefficients is computed through the following matrix operation:
=
Now, for the values provided for and , we get that the vector with estimated regression coefficients is computed as follows:
=
Therefore, based on the data provided, the estimated multiple linear regression equation is:
b.
Here, we need to check which of the coefficients, if any, are significantly different from zero at the 90% level.
1 - = 0.90
= 0.90
And we know the p-value test.
p < . We reject the null hypothesis.
p > . We do not reject the null hypothesis.
Null Hypothesis :
Alternative Hypothesis :
The p-values corresponding to each coefficient are as follows:
Intercept 0.00566 < 0.01
POH 0.62132 > 0.01
TLAM 0.01226 > 0.01
Except intercept, all other coefficient are not significantly different from zero at the 90% level.