Dear Dr. Kelly,
I suggest you e-mail Dr. Brogan for insight into this matter. She is an
expert in complex survey data analysis and has knowledge of Stata software.
Her e-mail:
[email protected]
Good Luck,
Suzy
Jana Morgan Kelly wrote:
Dear Colleagues,
I am currently working on a project in which I am using STATA to
analyze survey data collected using a multi-stage stratified sampling
method with clustering, where the clusters are selected using a
weighted sampling scheme according to their relative size and each
cluster has an N of 30. The data do include a variable denoting which
cluster each case belongs to. However unfortunately, the data set does
not seem to include variable(s) indicating the different strata (here
five regions and then within each region urban or rural). Furthermore
I have not been able to generate a strata variable for region based on
the information that is available in the data set, although I do have
a variable that distinguishes between the capital city, urban, and
rural, which is an aggregation of the regions.
The way I see it I have three basic options:
1. Use svyset to set the weights, the PSUs, and the strata, using the
capital city vs. urban vs. rural as the strata variable.
2. Use svyset to set the weights and PSUs but leave the strata
unspecified
3. Use svyset to set the weights and leave both the PSUs and the
strata unspecified.
Any suggestions about which of these strategies would be most
satisfactory given the constraints of my data would be most welcome.
Thanks in advance,
Jana
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
