| Combined men/women/hiv dataset [message #2359] |
Sun, 08 June 2014 19:18  |
carolina07
Messages: 1
Registered: June 2014
|
Member |
|
|
Hello, we are analyzing behavioral characteristics among HIV positives in several countries separately. We have merged the IR (women), MR (men), and the HIV dataset into a single dataset for each country.
Is it correct to analyze men and women combined while applying the HIV weight, as in the below example?
gen float hiv_weight = hiv05 / 1000000
tab x1 x2 [iw=hiv_weight]
egen strata = group(v024 v025) // Note: this depends on the country
svyset [pweight=hiv_weight], psu(v021) strata(strata)
svy: logistic y x1 x2 // Note: adjusting for sex
|
|
|
|
| Re: Combined men/women/hiv dataset [message #2366 is a reply to message #2359] |
Mon, 09 June 2014 13:40  |
Bridgette-DHS
Messages: 3199
Registered: February 2013
|
Senior Member |
|
|
Following is a response from Senior Data Processing Specialist, Noureddine Abderrahim:
When you combine men and women(matching couples), and need to get the HIV prevalence, you need to use the man's weight: MV005 divided by 1000000.
If you concatenated men and women without matching them, you should use the HIV weight: HIV05 divided by 1000000, to get the HIV prevalence.
|
|
|
|
The DHS Program User Forum
Members
Search
Help
Register
Login
Home
