Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
RE: st: RE: get robust SE and adjust for samling design
From
Timothy Mak <[email protected]>
To
"[email protected]" <[email protected]>
Subject
RE: st: RE: get robust SE and adjust for samling design
Date
Fri, 16 Aug 2013 13:00:46 +0800
Liling,
What we've been trying to say is that, if you use -svy-, the standard error is already robust, so you don't have to additionally specify vce(robust). In fact, Stata deliberately does not allow you to specify otherwise, because it does not know of another way to do the standard errors.
Tim
-----Original Message-----
From: [email protected] [mailto:[email protected]] On Behalf Of HUANG, LILING
Sent: 16 August 2013 11:56
To: <[email protected]>
Subject: Re: st: RE: get robust SE and adjust for samling design
Steve,
How do I get a robust SE with svy: poisson? The option vce(robust) is not allowed with the svy prefix.
Thanks,
Liling
On Aug 15, 2013, at 10:35 PM, "Steve Samuels" <[email protected]> wrote:
>
> -svyset- the data to properly account for strata, PSUs, and weights.
>
> . svyset psu1 [pwt = pwt], strata(stratid)
>
> . svy: poisson //_not_ . poisson
>
> Steve
>
> On Aug 15, 2013, at 10:37 AM, HUANG, LILING wrote:
>
> Hi Tim,
>
> Thank you very much for responding to my question. What I was trying to do is to obtain robust SE to control for mild violation of the assumption for the poisson model in my dataset. I cannot get robust SE when I ran svy:poisson x y z, vce(robust). I got the error message saying option vce() of poisson is not allowed with the svy prefix.
> So, I tried running a regular poisson with adjusting for sampling design to get robust SE. The syntax that I wrote is poisson x y z [pw=pwt], vce(cluster psu1). But this adjusts only for 61 clusters in psu in my dataset. I need to have it adjust for 6 strata too. Do you know how I can specify this adjustment for strata in my syntax, poisson x y z [pw=pwt], vce(cluster psu1).
> Thanks,
>
> Lilling
> ________________________________________
> From: [email protected] [[email protected]] on behalf of Timothy Mak [[email protected]]
> Sent: Thursday, August 15, 2013 12:03 AM
> To: [email protected]
> Subject: st: RE: get robust SE and adjust for samling design
>
> Hi Liling,
>
> When you've -svyset- your data, you don't need to use vce(cluster) anymore. That's already done.
>
> Tim
>
> -----Original Message-----
> From: [email protected] [mailto:[email protected]] On Behalf Of HUANG, LILING
> Sent: 15 August 2013 03:37
> To: [email protected]
> Subject: st: get robust SE and adjust for samling design
>
> Dear Statalist,
>
> Since option vce() of poisson is not allowed with the svy prefix, I tried running a regular poisson regression with [pw=sampwgt], vce(cluster psu) to get robust standard errors.
>
> I need to adjust for strata as well. Does anyone know how I can specify it in the syntax?
> . poisson x y z [pw=pwt], vce(cluster psu1)
>
> Here are my original svy syntax.
> svyset [pweight=pwt], strata(strata) psu(psu1)
> . svy:poisson x y z
> (running poisson on estimation sample)
> Survey: Poisson regression
> Number of strata = 6 Number of obs = 1986
> Number of PSUs = 61 Population size = 1934.5412
> Design df = 55
> F( 2, 54) = 86.34
> Prob > F = 0.0000
> ------------------------------------------------------------------------------
> | Linearized
> x | Coef. Std. Err. t P>|t| [95% Conf. Interval]
> -------------+----------------------------------------------------------------
> y | .2246702 .0172226 13.05 0.000 .1901554 .2591849
> z | .0017301 .0356631 0.05 0.961 -.0697404 .0732006
> _cons | .2436516 .0499279 4.88 0.000 .1435939 .3437094
> ------------------------------------------------------------------------------
>
> Thanks,
>
> Liling Huang
>
> *
> * For searches and help try:
> * http://www.stata.com/help.cgi?search
> * http://www.stata.com/support/faqs/resources/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/faqs/resources/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/faqs/resources/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/faqs/resources/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/faqs/resources/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/faqs/resources/statalist-faq/
* http://www.ats.ucla.edu/stat/stata/