Hi, Stata users, thank you all for sharing your knowledge and help others!
I have panel data, and tests showed me NO individual effects that are statisitically signifacant. On the other side, I tested for time fixed effects (N dimension is > T), and F test showed that I should include them in the model. So I need to apply LSDV method, since time effects can be included by dummy variables - am I right?
My question is, is it correct to use BP test to test heteroskedasticity? And what test is suitable in this case to test autocorrelation?
Thank you all for help.
King regards,
Jo
Related Posts with Including only time effects in model - how to test specification
Margin effects with multinomial logistic regression- Moderation effect
- Choice models -cmxtmixlogit- command for including random parameter for one of the alternatives (the base/status quo alternative)
- Estimation using only one out of many categorical variables without dropping the rest
- define column width in xl() "open" depending on values (reals & strings)?
Subscribe to:
Post Comments (Atom)
0 Response to Including only time effects in model - how to test specification
Post a Comment