| Re: survey set preparation for Pakistan DHS 2012-13 [message #2793 is a reply to message #2770] |
Mon, 25 August 2014 19:01   |
Reduced-For(u)m
Messages: 292
Registered: March 2013
|
Senior Member |
|
|
This is a little tricky since I haven't used the Pakistan data, but the usual method, which seems to be applicable here, is to define strata as the combination of region and urban/rural (that is how the sample is stratified according to the Pakistan report (http://dhsprogram.com/pubs/pdf/FR290/FR290.pdf).
First, your PSU is defined wrong - use v021. v001 is something else
As for the strata, the usual method is to choose region, separated by urban/rural, which is done thusly:
egen strata = group(v024 v025), label
This tells Stata that the survey is stratified by region and urban/rural. That sounds like what the Pakistan report is saying in the Appendix: http://dhsprogram.com/pubs/pdf/FR290/FR290.pdf
Check out the DHS FAQ (http://dhsprogram.com/faq.cfm) under "using data files"
There is a small chance that the Pakistan survey works differently. You may want to investigate how v022 relates to the strata as I defined them above. But v023 is almost certainly the wrong variable.
|
|
|
|
The DHS Program User Forum
Members
Search
Help
Register
Login
Home
