| Re: Correct weight for a sub-sample [message #491 is a reply to message #486] |
Wed, 29 May 2013 15:13   |
adi.greif@yale.edu
Messages: 5
Registered: May 2013
|
Member |
|
|
I would also like an answer to this question! I am interested in looking at rural women who rank poorest on the wealth index for various datasets.
I know that if you are using STATA, you can use svyset (and set the sample weight for the overall sample to v005/100000), and then the subpop option to specify the group you are interested in looking at. STATA will handle the weighing of the subpopulation. However, in your case, if you are looking at all non-missing answers to certain variables, I think you don't need to specify a subpopulation. I am basing this answer on the examples given here:
http://www.stata.com/features/survey/svy-survey.pdf
Does anyone know how to handle weighing sub-samples without using STATA's subpop command?
[Updated on: Wed, 29 May 2013 15:17] Report message to a moderator |
|
|
|
The DHS Program User Forum
Members
Search
Help
Register
Login
Home
