Re: st: using -mixed- with clustered data that includes probability weights

[Alfonso Sánchez-Peñalver <alfonso.statalist@gmail.com>]

If doctors were not sampled, then they basically are a consequence of the sampling of the patients. Since we know that some patients have been over-sampled and others under-sampled, the question really is what proportion of each type of patients does each doctor have. Because a doctor would then be over-sampled or under-sampled depending on the over-sampling and under-sampling of the patients. Wouldn't they? Thus it may be appropriate to estimate a weighted average of the patients' weights for a doctor, where the weights for this weighted average could be the illness severity score, since it's the basis for the over-sampling. Don't you agree?

Alfonso Sanchez-Penalver, PhD

On Feb 27, 2014, at 9:48 PM, Steve Samuels <sjsamuels@gmail.com> wrote:

> To elaborate on Stas's post: if doctors were not sampled, then you can
> define the doctor-level weight with pweight(1).
> 
> Steve
> sjsamuels@gmail.com
> 
> On Feb 27, 2014, at 8:21 PM, Stas Kolenikov <skolenik@gmail.com> wrote:
> 
> Whether your results are biased really depends on your study design.
> -mixed- knows nothing about your design (Stata 14 request: a command
> to read the pdf file and extract the design information from the
> narrative -- may I suggest -pdf2svyset- as a prospective name? I am
> sure a lot of researchers would find this handy), and just warns you
> in case you had sampling at several levels. If there were no sampling
> at the doctor's level, then weighting only at the patient level that
> you have is appropriate.
> 
> -- Stas Kolenikov, PhD, PStat (ASA, SSC)
> -- Principal Survey Scientist, Abt SRBI
> -- Opinions stated in this email are mine only, and do not reflect the
> position of my employer
> -- http://stas.kolenikov.name
> 
> 
> 
> On Thu, Feb 27, 2014 at 7:55 PM, Stephen Henry <sghenry@ucdavis.edu> wrote:
>> Hi,
>> 
>> I want to know whether I can used -mixed- in Stata 13.1 to analyze
>> clustered data that include probability weights.
>> 
>> My data were collected to study patients during clinic visits.
>> Each patient is unique, and patients are clustered within doctors.
>> In addition, patients were sampled based on an illness severity score.
>> Patients with more severe symptoms were over-sampled.
>> Patient sampling was done independent of which doctor was seeing the
>> patient.
>> 
>> I have been analyzing data using the -reg- command and cluster option as
>> follows:
>> 
>> reg v1 v2 v3 [pweight=weight], cluster(doctor_id)
>> 
>> However, I'd like to use -mixed- instead to take advantage of the
>> additional postestimation commands.
>> 
>> Stata will run the following command:
>> 
>> mixed v1 v2 v3 [pweight=weight] || doctor_id:
>> 
>> but warns me that "Sampling weights were specified only at the first level
>> in a multilevel model."
>> Are my results with the -mixed- command potentially biased?  If so, is
>> there an easy way to fix this?
>> 
>> Thanks in advance,
>> 
>> Stephen Henry
>> University of California Davis
>> Sacramento, California
>> *
>> *   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/