Re: st: suest after svy:clogit

[William Buchanan <william@williambuchanan.net>]

The help file for -suest- states that it cannot be used after estimations with robust/cluster robust standard errors.  Instead the models are fitted with normal standard errors and the correct standard errors are requested when you run -suest- as an option.  Since you're using -svyset- data, the standard errors are going to be adjusted for the weighting.  So -suest- will continue to give you the same error.  If you've already fitted the model with the interaction term, what is the addition problem you are attempting to solve?  

HTH,
Billy

Sent from my iPhone

On May 17, 2013, at 9:04, Klaus Nowotny <klaus.nowotny@sbg.ac.at> wrote:

> As that was the first thing I did, could you be more specific?
> 
> > If your interest is only in whether the coefficient of interest varies by groups why not include the appropriate interaction in the model and interpret the coefficient from the interaction directly?
> 
> Did that too, as mentioned.
> 
> Anyway, thanks for your time.
> Klaus
> 
> 
> Am 17.05.2013 17:44, schrieb William Buchanan:
> 
>> Read the help file for -suest-.  If your interest is only in whether the coefficient of interest varies by groups why not include the appropriate interaction in the model and interpret the coefficient from the interaction directly?
>> 
>> HTH,
>> Billy
>> 
>> Sent from my iPhone
>> 
>> On May 17, 2013, at 8:32, Klaus Nowotny <klaus.nowotny@sbg.ac.at> wrote:
>> 
>>> Dear Maarten,
>>> 
>>> thanks for the answer and the hint comcerning the iweights. But changing svyset to
>>> 
>>> svyset id [pw=weight]
>>> 
>>> does not solve my initial problem. I've already estimated the model using interactions, but it would have been interesting to know why I get the "option cluster() is not allowed"-error.
>>> 
>>> In any case, thanks a lot!
>>> Klaus
>>> 
>>> Am 17.05.2013 13:14, schrieb Maarten Buis:
>>> 
>>>> On Thu, May 16, 2013 at 6:31 PM, Klaus Nowotny wrote:
>>>>> I am having troubles using suest to test the equality of coefficients across
>>>>> two groups after svy:clogit. My survey data is set up as follows:
>>>>> 
>>>>> svyset id [iw=weight]
>>>> To quote -help weight-: "iweights, or importance weights, are weights
>>>> that indicate the "importance" of the observation in some vague sense.
>>>>  iweights have no formal statistical definition; any command that
>>>> supports iweights will define exactly how they are treated.  Usually,
>>>> they are intended for use by programmers who want to produce a certain
>>>> computation."
>>>> 
>>>> So using iweights for a statistical analysis is almost certainly a mistake.
>>>> 
>>>>> I then estimate the model for two subgroups:
>>>>> 
>>>>> svy, subpop(if group==1): clogit choice x1 x2, group(id)
>>>>> est sto group1
>>>>> svy, subpop(if group==2): clogit choice x1 x2, group(id)
>>>>> est sto group2
>>>>> 
>>>>> To test for the equality of the coefficients of x1 across the two groups I
>>>>> follow the Stata reference manual for suest and enter
>>>>> 
>>>>> suest group1 group2
>>>>> 
>>>>> but Stata returns "r(198); option cluster() is not allowed with svy results"
>>>>> although I never entered the cluster option anywhere.
>>>> Why not just estimate the entire model in one go?
>>>> 
>>>> svy : clogit choice i.group##(c.x1 c.x2), group(id)
>>>> 
>>>> Hope this helps,
>>>> Maarten
>>>> 
>>>>  The same occurs if I
>>>>> try the alternative procedure suggested in the reference manual:
>>>>> 
>>>>> clogit choice x1 x2 if group==1 [iw=weight], group(id)
>>>>> est sto group1
>>>>> clogit choice x1 x2 if group==2 [iw=weight], group(id)
>>>>> suest group1 group2, svy
>>>>> 
>>>>> Why do I keep getting this error message? Any help would be greatly
>>>>> appreciated!
>>>>> 
>>>>> Thanks,
>>>>> Klaus
>>>>> 
>>>>> --
>>>>> Klaus Nowotny
>>>>> University of Salzburg
>>>>> 
>>>>> *
>>>>> *   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/
>>>> 
>>>> --
>>>> ---------------------------------
>>>> Maarten L. Buis
>>>> WZB
>>>> Reichpietschufer 50
>>>> 10785 Berlin
>>>> Germany
>>>> 
>>>> http://www.maartenbuis.nl
>>>> ---------------------------------
>>>> *
>>>> *   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/