Bookmark and Share

Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: st: Regression diagnostics with panel data (-xtreg-)

From   "Tobias Pfaff" <[email protected]>
To   <[email protected]>
Subject   RE: st: Regression diagnostics with panel data (-xtreg-)
Date   Tue, 30 Aug 2011 12:12:11 +0200

Dear all,

I guess my original email was a bit too long, so I will sum up my questions
regarding regression diagnostics after fixed effects (-xtreg, fe-):

1.) How can you identify influential observations after a fixed effects
panel regression? Is it OK to use -regress- with the same equation and do
the diagnostics with DFITS etc. (as suggested here:

2.) -predict- after -xtreg- calculates either u_i (fixed-error component) or
e_it (overall error component). Concerning the assumption of normality of
residuals, is it correct that u_i needs to be checked for normality?

3.) I understand that the RESET (-ovtest-) after -regress- is a test for the
functional form, but not for omitted or irrelevant variables. Is there a
graphical or formal test for omitted or irrelevant variables after -xtreg-

My panel data set is unbalanced, covers 26 years, 40,000 individuals and
315,000 observations. Standard errors are clustered for region*year (-xtreg
..., fe vce(cluster region_svyyear)). I use Stata SE 11.2.

Any comments & help is appreciated!


*   For searches and help try:

© Copyright 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index