Where:
SSE1 : Sum Square Error from model Ordinary Least Square
SSE2 : Sum Square Error from Fixed Effect Model
n : Number of companies (cross section)
nt : Total cross section x total of time series
k : The number of independent variables
To compare with the value of F table, can use the following formula :
α : The significance level used (alfa)
n : Number of companies (cross section)
nt : Total cross section x total time series
k : The number of independent variables
b. Hausman-Test
After conducting the Chow test then the next we will examine which model the Fixed Effects Model or Random
Effects Model the most appropriate, this test is referred to as the Hausman test.
Tests conducted by the Hausman test the following hypotheses :
H0 : Random Effect Model
H1 : Fixed Effect Model
The Hausman test statistic follows the Chi Square statistic distribution with degree of freedom as k-1, where k is
the number of variables of the study overall. If the value of the Hausman statistic is greater than the critical value
then H0 is rejected and the appropriate model is the Fixed Effects Model while the opposite when the value of
the Hausman statistic is smaller than the critical value then the appropriate model is the Random Effects Model.
c. Lagrange Multiplier
Lagrange Multiplier (LM) is a test to determine whether the Random Effects Model or Ordinary Least Square
model is most appropriate.
Hipotesis yang digunakan adalah :
H0 : Ordinary Least Square Model
H1 : Random Effect Model
LM test is based on the chi-squares distribution with degree of freedom for the number of independent variables.
If the value of the LM statistic greater than the critical value of chi-squares statistic we reject the null hypothesis,
which means that a precise estimate for the panel data regression model is a model of Random Effects Model of
the Model Ordinary Least Square. Conversely, if the value of the LM statistic is smaller than the value of chisquares
as a critical value, then we accept the null hypothesis, which means that the estimates used in the panel
data regression model of Ordinary Least Square is not Random (Random Effects Model).
4.2.Hypothesis testing
4.2.1.Hypothesis Test Using the t test (partial)
After making the overall regression coefficient test, then the next step is to calculate the individual regression
coefficients (partial), using a test known as the t test. The hypothesis in this test is as follows :
H0 : βj = 0
H1 : βj ≠ 0; j = 0,1,2 ….,k
k is the slope coefficient
From the hypothesis, it can be seen whether the independent variables have a significant influence on
the dependent variable. T values resulting from the processing will be compared with the value of the t table. If it
turns out after │t count│> t table, the t values are in the rejection region, so that the null hypothesis is rejected at
confidence level (1-α) × 100%. In this case it can be said that the statistically significant independent variables
on the dependent variable.
4.2.2.Hypothesis Testing Using the F test (simultaneous)
F test is used to determine whether all the independent variables together can influence the dependent variable
(the goodness of fit model). F test is done by comparing the F count and F table with a predetermined degree of