H0: 
H1: 
, 
for some j=0,1,2,3,4
We need to test this model make sense or not by
doing an analysis of variance.
If H0 is
accepted, all 
are equal to
zero, then there are no relationship between y, X1, X2, X3
and X4. If H0 is rejected, then the model makes sense. Therefore, it is meaningful to do the forward selection,
backward elimination and stepwise selection to get the best model.
Analysis of Variance( ANOVA table)11
|
Source |
DF |
Sum of square |
Mean square |
F-value |
Pr>F |
|
Model |
4 |
2158.73 |
539.68 |
4.07 |
0.0184 |
|
Error |
16 |
2122.22 |
132.64 |
|
|
|
Corrected Total |
20 |
4280.95 |
|
|
|
From the above table, we see that the model with all four variables is acceptable.
Since the p-value from ANOVA table is 0.0184 which is smaller than 0.05