st: RE: Eureka! Bootstrapping and Interval Regressions for Panels

["jyotsna puri" <pjyotsna@hotmail.com>]

Apologies!
I had my eureka moment immediately after sending my email off.
(Should have sent it off a day earlier!)
For those interested, it was the '-e(sample) -' option within the inverted 
commas
in the bootstrap command, that was reducing my sample size...
-bs " " -

Another incredible use of this list:
...Epiphanies!

-jo


Jyotsna (Jo) Puri
pjyotsna@hotmail.com
I am supporting the education of underprivileged kids. You can too:
http://www.ashanet.org/nycnj/marathon/runnernet/publicmypage.php?runnerid=2005TA486




----Original Message Follows----
From: "jyotsna puri" <pjyotsna@hotmail.com>
Reply-To: statalist@hsphsun2.harvard.edu
To: statalist@hsphsun2.harvard.edu
Subject: st: Re: Bootstrapping and Interval Regressions for Panels
Date: Sun, 05 Feb 2006 21:33:27 +0000

Hi all,

I have to bootstrap an Interval regression model for an unbalanced panel 
dataset.
(I have an endogenous variable - have to correct the standard errors in the 
interval regression.)
The Interval regression runs fine. (I use -xtintreg-)
The bootstrap does not. (-bs "xtintreg ...." - )

Things I have done to troubleshoot:
I have removed the cluster option (from bs), so that the bootstrap sample 
size
is not limited to the no. of clusters.
Put a trace on bootstrap (more on this below).
Have let the number of quadrature points to the default i.e. 12 for 
xtintreg.
I have set a -size - option in the bootstrap and set my size (of bootstrap 
sample)
to 100, which should be greater than any number of quadrature points, if 
that were
the bottleneck.

None of this in turns and together helps.
My output (and error message) below - which I have tried to deal with.

Will be extremely grateful for any clues.
-Jo

[The last line is the error message ]

-xtprobit crbadu1 year Lpopn Ltrmnew agwsc2 agwlc2 adltrt acidsoildum core 
compedrt Ltrdmin if
e(sample), re i(id)
[OUTPUT OMITTED]
-predict crhat if e(sample), pu0
[OUTPUT OMITTED]
-xtintreg useint1 useint2 year Lpopn Ltrmnew agwsc2 agwlc2 adltrt core 
acidsoildum crhat if e(sample), i(id) quad(26)
[OUTPUT OMITTED]
-quadchk, nooutput
[OUTPUT OMITTED]

-bs "xtintreg useint1 useint2 year Lpopn Ltrmnew agwsc2 agwlc2 core adltrt 
acidsoildum crhat if e(sample), i(id) " _se, reps(50) saving(bsout1) replace 
bca trace

. bs "xtintreg useint1 useint2 year Lpopn Ltrmnew agwsc2 agwlc2 core adltrt 
acidsoildum crhat if e(sample), i(i

> d) " _se, reps(50) saving(bsout1) bca trace
bootstrap: First call to (xtintreg useint1 useint2 year Lpopn Ltrmnew agwsc2 
agwlc2 core adltrt acidsoildum crh

> at if e(sample) , i(id)) with data as is:
. xtintreg useint1 useint2 year Lpopn Ltrmnew agwsc2 agwlc2 core adltrt 
acidsoildum crhat if e(sample) , i(id)
number of quadrature points must be less than or equal to number of obs




Jyotsna (Jo) Puri
Cell and Voicemail: 201-805-1690
pjyotsna@hotmail.com
I am supporting the education of underprivileged kids. You can too:
http://www.ashanet.org/nycnj/marathon/runnernet/publicmypage.php?runnerid=2005TA486


*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/


*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/