Hypothesis testing

[monica88]

Hi,

I have done OLS estimation with panel data (in a panel workfile) and would like to do some hypothesis testing (I am using EViews 7).

A) I need the t-ratio for the hypothesis that the difference between two coefficients is zero (against the hypothesis that the two coefficients differ). Do I use the Wald test for this? If the two coefficients are C2 and C4 would I just write in the box "C(2)=C(4)"?

B) I need to also test the hypothesis that one coefficient exceeds the other. How would I do this?

Thank you for your help!

[EViews Glenn]

A) yes
B) http://forums.eviews.com/viewtopic.php? ... =one+sided

[monica88]

Thank you very much, you are very helpful!

This might be a stupid question, but do I use the same degrees of freedom as reported for the two-tailed test (wald test) when I look up the critical value in the table for the one-tailed test?

I really appreciate your quick answers :)

[EViews Glenn]

Yes. It's the same statistic, you're just looking at one-tail of the distribution. Note that given symmetry, the p-value should be half of the quoted p-value.