EViews.com

Hello E-views forum and thanks a lot for the help you provide to starters like me.

I have a question regarding correlation matrix.
I have panel data ( 14 countries, 11 years) and I want to produce a correlation matrix for my thesis.

I open the variables as a group, and then I go view-Covariance analysis-Correlation.

Afterwards I estimate a fixed effects model (country) and the coefficient signs are different with those on the correlation matrix.
Is it ok? Or that indicates a problem?

Thank you very much in advance for any help provided.

Perfectly okay. The correlation simply looks at the pair-wise relations. The regression looks at relations controlling for other factors.

Thank you very much for the reply.
If I get it right the correlation matrix does not take into account the fixed effects and that's why the negative correlation (on the matrix) turns to a positive one in the fixed effects model. Is this right?

Thank you

Ya got it.

Perfect! Thanks and keep up the good work!!