Significant changes

[Zilvinas Kvedaravicius]

Hello,

I am running regressions on panel data by using fixed effects model, I want to compare changes in betas in my subsamples, for example 2000-2004 vs 2005-2009, how I can determine if change in beta coefficient was significant or not?

Best regard,
Zilvinas

[Nicole Boyle]

Question is a bit vague. Please provide your syntax and some sample output (monospaced).

[Oliver Elorreaga]

Maybe you can use a command called "suest", other option would be a dummy*beta to capture the effect of each subsample. If you find another solution, please share it.

[Maarten Buis]

My standard answer in such cases is to not estimate different models, but do it all at once in one model with interaction effects.