In: Statistics and Probability
Suppose for a multiple regression on just 5 observations you are given the following portion of an excel regression output:
RESIDUAL OUTPUT
Observation |
Predicted y(hat) |
Residuals |
1 |
73.61 |
0.39 |
2 |
93.03 |
-1.03 |
3 |
58.97 |
-0.97 |
4 |
85.21 |
-0.21 |
5 |
78.18 |
1.82 |
Test the model for autocorrelation at a 10% level of significance.
Test the model for heteroskedasticity using a level of significance of 5%
The details solution given below.
For autocorrelation
The Hypotheses for the Durbin Watson test are:
H0 = no first order autocorrelation.
H1 = first order correlation exists.
d = 6.7185 / 5.5104
d = 1.21924
From the value of d we can say that 0 to <2 is positive autocorrelation at given level of significance.
For Heteroskedasticiy
Test the model of Heteroskedasticity using the graphical method there is no prior information about the nature of heteroscedasticity.
We can say that there is no any pattern or not constant pattern between two variables we can't say at in present data.
By using the Breusch Pagan Test (B-P test) for Heterosckdasticiy
n=5
The test statistic is LM = ESS/2
=ei^2 /2
= 5.5104 / 5 = 1.10208
critical value chisq at 5% = 0.710723
Ho: There is no heteroskedasticiy in the data
H1: There is heteroskedasticity in the data
Conclusion: test statistic value is greater than critical value then Reject Ho
i.e. There is heteroscedasticity in above data at 5% of level of significance.