Guilherme,
The sampling weight only accounts for the unequal selection probabilities and gives unbiased point estimates. You would use the weighting variable (I assume this is just a weight with post-stratification adjustments) provided for the combined analysis, and then use your stratum and cluster identifiers to adjust standard errors. Are there no bootstrap weights provided?
Cam
----------------------------------------
> From: [email protected]
> To: [email protected]
> Subject: Re: st: svyset for survey with oversampling
> Date: Sat, 22 Aug 2009 02:22:34 +0900
>
> Hi Cam,
>
> Thank you for your reply. They do provide a weight variable to use
> when analysing the samples together to correct for urban population
> oversampling, but I believe that it is not taking in to account the
> other biases introduced by the multistage sampling.
>
> It would be nice if there was a solution that allowed me both to
> correct the biases of multistage sampling and of oversampling of the
> urban population.
>
>
> Guilherme
>
> On 2009/08/22, at 2:12, Cameron McIntosh wrote:
>
>> Hi Guilherme,
>> They don't provide some adjustment to the sampling weights to allow
>> for combined analysis?
>> Cam
>>
>> ----------------------------------------
>>> From: [email protected]
>>> To: [email protected]
>>> Subject: st: svyset for survey with oversampling
>>> Date: Sat, 22 Aug 2009 02:00:35 +0900
>>>
>>> Dear Statalist Members
>>>
>>> It is my first post to this list. My name is Guilherme Kenji Chihaya
>>> and I am currently a graduate student in Tohoku University, Japan. I
>>> am using a data set called Life Histories and Social Change in
>>> Contemporary China. It is a survey in which the rural and urban
>>> populations of China were treated as different populations and each
>>> one was sampled for about 4000 samples using multistage sampling with
>>> clustering.
>>>
>>> The proportion of China's rural population was about 70% by the time
>>> of the survey, however, this survey is designed so that it accounts
>>> for 50% of the whole sample if you try to analyse the two samples
>>> together. Besides this problem, it still has all the other biases
>>> caused by multistage clustering.
>>>
>>> I wonder if it is possible to use svyset to correct all these biases
>>> so that I can analyse the two samples as one single dataset. The
>>> codebook for the survey is incomplete, it states that the data is
>>> supposed to be analysed using Stata complex samples functionality but
>>> the part about how to use it is unfinished in the available version.
>>> There are variables identifying the primary sample unit and the
>>> stratum used for sampling. However, I don't know how to deal with the
>>> rural-urban disproportionality.
>>>
>>> I would appreciate if someone could help me with this, especially if
>>> there is anybody in this list that is familiar with this dataset.
>>>
>>> Guilherme Kenji Chihaya
>>>
>>> *
>>> * 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/
>> _________________________________________________________________
>> More storage. Better anti-spam and antivirus protection. Hotmail
>> makes it simple.
>> http://go.microsoft.com/?linkid=9671357
>> *
>> * 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/
_________________________________________________________________
More storage. Better anti-spam and antivirus protection. Hotmail makes it simple.
http://go.microsoft.com/?linkid=9671357
*
* 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/
