Statalist


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

st: AW: Robust standard errors and t-values with xtfevd


From   "Martin Weiss" <martin.weiss1@gmx.de>
To   <statalist@hsphsun2.harvard.edu>
Subject   st: AW: Robust standard errors and t-values with xtfevd
Date   Fri, 19 Jun 2009 17:18:47 +0200

<> 

This examples seems to confirm your complaint...


*************

/*get the command */ 
capture copy /* 
 */ http://www.polsci.org/pluemper/xtfevd.ado xtfevd.ado
di in red _rc
capture copy /* 
 */ http://www.polsci.org/pluemper/xtfevd.hlp xtfevd.hlp
di in red _rc

qui{
	webuse psidextract, clear
	
	/*HT Taylor as reference */ 
	xthtaylor lwage wks south smsa/* 
	 */  ms exp exp2 occ ind union fem blk ed,/* 
	 */  endog(exp exp2 occ ind union ed)/* 
	 */  constant(fem blk ed)
	estimates store HT
	
	/*xtfevd with standard se */ 
	xtfevd lwage wks south smsa/* 
	 */  ms exp exp2 occ ind union fem blk ed,/* 
	 */ invariant(fem blk ed)  s2iv_endog(ed)/* 
	 */  s2iv_exog(wks south smsa ms)
	estimates store xtfevd
	
	/*xtfevd with robust se */ 
	xtfevd lwage wks south smsa/* 
	 */  ms exp exp2 occ ind union fem blk ed,/* 
	 */ invariant(fem blk ed)  s2iv_endog(ed)/* 
	 */  s2iv_exog(wks south smsa ms) robust
	estimates store xtfevdrob
	
	/*xtfevd with robust se and depvar change */
	replace lwage=lwage/100
	xtfevd lwage wks south smsa/* 
	 */  ms exp exp2 occ ind union fem blk ed,/* 
	 */ invariant(fem blk ed)  s2iv_endog(ed)/* 
	 */  s2iv_exog(wks south smsa ms) robust
	estimates store xtfevdrobchan
}

estimates table HT xtfevd xtfevdrob xtfevdrobchan, /* 
 */ b(%4.2fc) se(%4.2fc) t(%4.2fc) /* 
 */ modelwidth(15) style(columns)
*************



HTH
Martin


-----Ursprüngliche Nachricht-----
Von: owner-statalist@hsphsun2.harvard.edu
[mailto:owner-statalist@hsphsun2.harvard.edu] Im Auftrag von Anders Dahl
Bakken
Gesendet: Freitag, 19. Juni 2009 15:24
An: statalist@hsphsun2.harvard.edu
Betreff: st: Robust standard errors and t-values with xtfevd

Im trying to use the robust option with xtfevd, but the problem for me is
that it seems that the t-values are
sensitive to the scale of my dependent variable. The results without the
robust option seems fine. But when i 
add the robust option, all the standard errors go trough the roof, (from
maybe .3 to 50000). But here's the thing, if 
a divide my dependent variable by for example 1000, my standard errors goes
down (naturally) but at the same time my t-values increases.

To me it almost seems like you can get whatever t-values you want by just
rescaling the left-hand side. As mentioned i dont have this problem 
without the robust option.

Have anybody else had this problem? And what can i do to fix it`?



Anders
 
_________________________________________________________________
Drag n? drop?Get easy photo sharing with Windows Live? Photos.

http://www.microsoft.com/windows/windowslive/products/photos.aspx
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/


*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/



© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index