Statalist


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

AW: AW: st: Re: estout and xtlogit


From   "Martin Weiss" <martin.weiss1@gmx.de>
To   <statalist@hsphsun2.harvard.edu>
Subject   AW: AW: st: Re: estout and xtlogit
Date   Wed, 11 Feb 2009 14:03:40 +0100

<> 




" it seems that standard errors are not saved at all (?)". 

Not so! The SEs are there for the coefficients, stored in the matrix e(V).
e(V) is a varcov, so to compare to the -xtlogit- output, take the square
root of the elements on the main diagonal...

*************
webuse union, clear
xtlogit union age grade not_smsa south southXt
matrix list e(V)
*************


HTH
Martin


-----Ursprüngliche Nachricht-----
Von: owner-statalist@hsphsun2.harvard.edu
[mailto:owner-statalist@hsphsun2.harvard.edu] Im Auftrag von Hilde Karlsen
Gesendet: Mittwoch, 11. Februar 2009 13:58
An: statalist@hsphsun2.harvard.edu
Betreff: Re: AW: st: Re: estout and xtlogit



"I do find it weird that the se for rho is not returned, as it is 
(probably)a prominent ingredient to the LR test at the bottom of 
-xtlogit- output..."


When I type: xtlogit movedout male, i(LPNR) quad(30)
...StataSE10.1 xtlogit output returns SE for all Coef., including 
sigma_u and rho. However, typing: ereturn list, it seems that standard 
errors are not saved at all (?). Storing the estimates typing: eststo... 
and then using the command estout/esttab then returns SE, but only for 
B-coeff, not for sigma_u and rho. I find it difficult understanding why 
these commands return SEs at all, when SEs did not appear to be saved 
for any of the coefficients in the ereturn list. I guess I've missed out 
on something.


Thank you for helping me; it is invaluable. I'm new to Stata and to the 
Statalist, but both the program and the list are great.


Regards,
Hilde


On 11.02.2009 11:01, Martin Weiss wrote:
> <> 
> 
> The manual is a little terse on this point. I got my info re -nlocm- from
> http://www.stata.com/statalist/archive/2008-06/msg00389.html
> It seems not many questions have been asked with respect to this issue,
> either. 
> 
> I do find it weird that the se for rho is not returned, as it is
(probably)
> a prominent ingredient to the LR test at the bottom of -xtlogit- output...
> 
> 
> 
> HTH
> Martin
> 
> 
> -----Ursprüngliche Nachricht-----
> Von: owner-statalist@hsphsun2.harvard.edu
> [mailto:owner-statalist@hsphsun2.harvard.edu] Im Auftrag von Ben Jann
> Gesendet: Mittwoch, 11. Februar 2009 10:55
> An: statalist@hsphsun2.harvard.edu
> Betreff: Re: st: Re: estout and xtlogit
> 
> To see a list of the returned results type
> 
>  . ereturn list
> 
> after running xtlogit. I seems xtlogit does not return the SEs of rho
> and sigma_u. You'd have to compute them yourself and add them to the
> results before being able to tabulate them. For example, the SE of
> sigma_u can be computed as follows:
> 
> . webuse union, clear
> . xtlogit union age grade not_smsa south southXt
> . estadd scalar se_sigma_u = [lnsig2u]_se[_cons] *
> exp([lnsig2u]_b[_cons]/2)/2
> . esttab, sca(rho sigma_u se_sigma_u)
> 
> Don't ask me about the SE of rho. Maybe check the manual.
> 
> ben
> 
> On Wed, Feb 11, 2009 at 9:59 AM, Hilde Karlsen <Hilde.Karlsen@hio.no>
wrote:
>> Thank you so much; it worked! However, it would be great if also standard
>> errors for rho and sigma_u could be presented in this table. I tried this
>> command:
>>
>> esttab, sca(rho sigma_u) se brackets
>>
>> ...but it leaves me with se only for B's and lnsig2u, - not for rho and
>> sigma_u. Any suggestions?
>>
>> Regards,
>> Hilde
>>
>>
>> On 10.02.2009 19:55, Martin Weiss wrote:
>>> <>
>>>
>>> *******
>>> webuse union, clear
>>> xtlogit union age grade not_smsa south southXt
>>> esttab, sca(rho sigma_u)
>>> *******
>>>
>>> HTH
>>> Martin
>>> _______________________
>>> ----- Original Message ----- From: "Hilde Karlsen"
<Hilde.Karlsen@hio.no>
>>> To: <statalist@hsphsun2.harvard.edu>
>>> Sent: Tuesday, February 10, 2009 7:28 PM
>>> Subject: st: estout and xtlogit
>>>
>>>
>>>> Hi, all!
>>>>
>>>> I could need some help regarding the estout command for xtlogit, and I
>>>> would really appreciate any suggestions on this issue. I have estimated
> a
>>>> multilevel model using xtlogit, and as I would like to have the results
> more
>>>> neatly presented, I used the estout-command. However, for some reason,
>>>> estimates and standard errors for sigma_u and rho are not present in
the
>>>> estout-output (only for the /lnsig2u)
>>>>
>>>> My question is therefore: how can I include estimates for sigma_u and
> rho
>>>> in the estout-table?
>>>>
>>>>
>>>> Regards,
>>>>
>>>>  Hilde Karlsen
>>>>
>>>>
>>>>
>>>> I don't know if this is helpful, but here is my estout-syntax:
>>>>
>>>>  estout using reging2.txt, eqlab("" "/lnsig2u ln of variance sigma_u "
>>>> "sigma_u L2 BS-variance" "rho Intraclass corr", none) replace
> style(fixed)
>>>> prehead("Model I: The probability of leaving nursing")
>>>>
>
posthead("------------------------------------------------------------------
> -")
>>>> cells("b(fmt(%9.3f) star label(Coef)) se(fmt(%9.3f) label (Std Err))")
> label
>>>> varwidth(30) varlabels(_cons "Constant")
>>>>
>
prefoot("-------------------------------------------------------------------
> -")
>>>> stats(N, fmt(%9.0g))
>>>> *
>>>> *   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/
>> --
>>
>> Hilde Karlsen
>> Høgskolen i Oslo
>> sg29/SG500
>> 22 45 27 47
>>
>>
>> *
>> *   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/
> 
> 
> *
> *   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/

-- 

Hilde Karlsen
Høgskolen i Oslo
sg29/SG500
22 45 27 47


*
*   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