EViews.com

Dear all,

I am currently working with a panel data set and want to run an OLS regression on dependent variable y with independent variable x.
The data set contains ~100 different groups (i.e. countries) and each has 10 years of observations.

I want to control for cross sectional and period effects. Í dont have significant results if I use "fixed" for both, cross sectional and period effects, but I get significant results if I use "cross sectional weighting" and add "@year" in order to control for period effects.

However, since Eviews forbids me to tick "fixed" effects for cross-sectional and period AND weighting, I am thinking if my results are biased?

I hope the question was clear!!!

Kind regards,

Chris


edit: I am using eviews 7

We don't allow the option because the memory efficient algorithms that we employ break down in the weighted case. You can always add the effects yourself, though there can be issues if you have small sample sizes.

Thank you for your answer Glenn!