0
$\begingroup$

I try to estimate fixed effects model in Stata and I have some problems with interpretation of my results.

My panel data consists of 700 observations. Five variables have significant impact on dependent variable.

What are the desirable values for: a) R-squared: within, between, overall b) corr(u_i, Xb) c) rho?

Could someone help me and give some explanations? I read a lot of presentations or Stata documentation but the examples of fe models which I found have very different parameters. I enclosed two links: https://www.princeton.edu/~otorres/Panel101.pdf http://www.svt.ntnu.no/iss/Arild.Blekesaune/SOS8515-filer/Panel_Lab2.pdf

$\endgroup$

1 Answer 1

0
$\begingroup$

a/ One cannot answer this question without more details on your research design. Having a high $R^2$ seems quite important to make predictions after you did the regressions. But for other research (when one is interested on the marginal impact of a variable, which is mostly what is done in economics research for instance), it is quite possible to find regressions with low $R^2$.

b/ and c/ which Stata command do you use ?

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.