The DHS Program User Forum
Discussions regarding The DHS Program data and results
Home » Topics » Biomarkers » Anemia variable in MR (Men's Anemia)
Re: Anemia variable in MR [message #19336 is a reply to message #18925] Tue, 02 June 2020 00:46 Go to previous messageGo to previous message
nagar002 is currently offline  nagar002
Messages: 9
Registered: April 2018
Member
Hi all,

My analysis is for India 2005 and India 2015. I am running a regression with anemia as the dependent variable for the sample of men and women who can be matched to each other (i.e. from the CR file) and have non-missing anemia information for both husband and the wife (to be precise, the regressions are run separately for these men and women). Then,

1) Is this sample likely to be a selected sample of men and women, or is it safe to say that it is a random sample for both men and women?


2) Since this regression is for a sub-sample of men (women) who have non-missing anemia information AND must be matched to wives (husbands) with non-missing anemia information, I am wondering if mv005 is the correct weight to use in the two separate regressions for husbands and wives, and is there any adjustment required to the weight variable. I ask this because the analysis is not for the universe of men who were surveyed (and found in the MR files), but only those who have anemia information available in the PR file and who have wives with non-missing anemia information in the CR file. The results of my regressions differ substantially based on whether I use man's weights or not (mv005 in CR file). I am wondering if in this case weights should not be used at all.

Many thanks for any help. Much appreciated.


 
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Previous Topic: BMI FOR AGE
Next Topic: BMI
Goto Forum:
  


Current Time: Sun May 26 02:25:02 Coordinated Universal Time 2024