[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

st: Pooled Data tests and Quantile Regressions

From   Sachin Chintawar <>
Subject   st: Pooled Data tests and Quantile Regressions
Date   Thu, 21 Aug 2008 11:07:41 -0500

Dear Statausers
I have a question on the poolability of data and tests associated with it. I
am stacking data from 2004-2006 to run Censored Qunatile regressions and
control for the years using year dummies. Following are some of the
1. One reason to stack data was to increase the number of observations
(roughly 14000) and thus the robustness of the estimates. The data is
collected for same variables but different individuals (this is a farm level
data). Do I still need to do poolability tests.
2. I checked the archives for any answers that I may get for poolability
tests and found a discussion about it earlier. Literature (Baltagi, 1981,
Journal of Econometrics)  suggests that Chow tests that was recommended to
test for poolability of data assumes homosckadasticity of data. In cases of
heterosckdasticity what would be the best way to do a poolability tests in
censored quantile regressions.

Any help would be greatly appreciated.


Sachin Chintawar
Research Assistant
Dept. of Agricultural Economics
Louisiana State University

-----Original Message-----
[] On Behalf Of Sebastian E. Wenz
Sent: Wednesday, August 20, 2008 2:01 PM
Subject: st: Your Stata list posting


Seriously: Please read/follow the Statalist FAQs before/when posting:

Repost with an informative subject line, your version of Stata, 
characteristics of your data, your code, the results, et cetera. (Again: 
check out the FAQs before reposting.)



Nishant Dass schrieb:
> Hi list-members,
> I have a question about unreported standard errors.  I include several
dummy variables in my regression but Stata doesn't report standard errors
for these dummy variables as well as for the constant term. 
> I thought it might be because the number of observations across the
different groups (each represented by the dummy variable) is not unevenly
distributed (i.e., there might be too many 1's or 0's in some groups) so I
tried making my groupings more coarse so as to reduce the number of dummy
variables.  And yet, with the dummy variables down from 50 to 10, I find the
standard errors missing.
> Could someone please help with an explanation/suggestion?
> Thanks a lot,
> Nishant
> *
> *   For searches and help try:
> *
> *
> *

*   For searches and help try:

*   For searches and help try:

© Copyright 1996–2017 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index