This is probably a really simple question for those of you who have used this test before but I am getting rather confused. This is using eviews 7, I have asked it to perform whites test and these are the results I recieved.

Heteroskedasticity Test: White

F-statistic 12.53629 Prob. F(35,4) 0.012

Obs*R-squared 39.63864 Prob. Chi-Square(35) 0.2707

The first probability is for the f statistic test for ommitted variables which comes out significant at the 5% level suggesting that I have omitted variables, as in by combining a number of my current variables I could find new ones? Then the second probability is for the actual white test, and the probability is above the 5% level suggesting that my model does not contain heteroskedasticty.

I found that one of my variables was correlated with the square of the error terms and so I have taken this variable out which results in the following white test results:

Heteroskedasticity Test: White

F-statistic 1.479219 Prob. F(27,12) 0.2412

Obs*R-squared 30.75837 Prob. Chi-Square(27) 0.2812

So now it appears there is no heteroskedasticty and no omitted variable issues?

I would really appreciate a confirmation that what I have stated is correct, im not sure if I need to find the chi squared value to test against the obs*R-squared value or if I can just go by the probabilities.

Thanks!!