3.5.5 Regression Statistic Test Analysis
After classical assumption test finishes and the result is free from classical assumption then regression statistic test analysis need to be done.
a. Hypothesis Test Method
If the result of the econometric model in this research is free from classical assumption, then hypothesis test needs to be done. Hypothesis test is
made through significant test to each independent variable to dependent variable which is done partially or together by through t test and F test.
b. Jointly Regression Coefficient Test F test
F test goal is to determine the significance of independent variable groups in influencing the dependent variable,
H :
β
1
= β
2
= β
3
= β
k
= 0 H
1
: β
1
≠ β
2
≠ β
3
≠ β
k
≠ 0 Suppose this research results in F
F
statistic
F
table
, it means that null hypothesis H
is accepted and the alternative hypothesis H
1
is rejected. If this condition happens, it means that the regression model
variation fails to explain the independent variable . Otherwise, if F F
statistic
F
table
the null hypothesis H is rejected and the alternative
hypothesis H
1
is accepted. If this condition happens it the model regression variation successfully explains the independent variable Gujarati, 2003.
F statistic equation, is;
� =
�
2
�−1 1
−�
2
�−�
3.18
In which: K = total estimation of parameters including the constant inside
N = total observation In the significant level of 5 , the requirement criteria are below;
1. Suppose from this research results in F
F
statistic
F
table
, the that null hypothesis H
is accepted and the alternative hypothesis H
1
is rejected. If this condition happens, the regression model variation fails to explain the independent
variables. 2.
Otherwise, if F F
statistic
F
table
the null hypothesis H is
rejected and the alternative hypothesis H
1
is accepted. If this condition happens the model regression variation successfully
explains the independent variables.
c. Individuality Coefficient Regression Test t test
t test goals is to determine the significance of independent variable individual influence an dependent variable. The hypotheses are below:
H :
β
1
= 0 H
1
: β
1
≠ 0 Suppose t
t
statistic
t
table
the null hypothesis H is accepted
and the alternative hypothesis H
1
is rejected. If this condition happens the model that is used is not good, because the independent variable could not be
explained by the dependent variable or, it’s not significant. Otherwise, if t t
statistic
t
table
, the independent variable is success fully explains the dependent variable perfectly, or it’s significant Gujarati, 2003.
d. Determination Coefficient Test of R