The DHS Program User Forum
Discussions regarding The DHS Program data and results
Home » Topics » Reproductive Health » Paired weighted analysis in SAS
Paired weighted analysis in SAS [message #12009] Sun, 19 March 2017 10:38 Go to next message
kdiclem1 is currently offline  kdiclem1
Messages: 8
Registered: March 2017
Location: New Haven, CT
I am using Tanzania DHS VII 2015-2016 data. I am using SAS to look at the relationship between hunger score and desire for more children. In order to do this, I need to do a weighted analysis of course, however I also need to do a paired analysis, since some of the women respondents are in the same household and are therefore not independent.

There is no "proc survey genmod" which would allow for this, so as I read about how to handle this, it looks as though I need to calculate my own individual sample weights (ISW) for each woman and use that in a regular proc genmod. Does anyone have any experience with calculating these ISWs and then using them in proc genmod for a paired analysis? Any insight on whether or not the ISW already exist in the DHS recode?

Thank you!
Re: Paired weighted analysis in SAS [message #12045 is a reply to message #12009] Wed, 22 March 2017 14:57 Go to previous message
Messages: 1384
Registered: February 2013
Senior Member
A response from senior sampling expert, Dr. Ruilin Ren,

If "ISW" means the women individual sampling weight, then we already calculated it and it is named V005, but diving it by 1000000 when using it because it was multiplied by 1000000 in data processing. If "ISW" does not mean the women individual survey weight, please clarify.

Previous Topic: women's age
Next Topic: About the 'CALENDAR' variables
Goto Forum:

Current Time: Thu Dec 13 09:08:55 Eastern Standard Time 2018