Re-normalizing HIV weights for pooled couples data [message #12925] |
Fri, 11 August 2017 14:50 ![Go to next message Go to next message](/theme/default/images/down.png) |
DaniD
Messages: 13 Registered: November 2014
|
Member |
|
|
Hi all,
First, I always have to start out by thanking everyone, especially Tom Pullum, for your help with the DHS data--I would be lost without this forum!
Ok so I have a quick question about re-normalizing the HIV weights. I have merged HIV and Couples datasets that I am pooling for multiple time points within single countries in order to assess change over time using confidence intervals. I am using the men's HIV sampling weight and would like to re-normalize the weight to be population based for each time point and country. My question is about the total population ("TOTWT" in Tom's immensely helpful code) number I should be using. Should this total population number be based on the men's population numbers for that country and year or should it be based on both men and women? And should that population number be narrowed to the ages included in the specific survey?
I am confused about this because the weight is for men but the description of the HIV weight on the DHS website indicates that it is normalized based on the population of both men and women. The website also uses the terminology of "total population" so I am uncertain about the age part.
Thank you!
Dani
|
|
|
|