| Re: Missing values for child weight and height [message #10225 is a reply to message #6686] |
Mon, 11 July 2016 10:45   |
RenaM
Messages: 6
Registered: May 2015
|
Member |
|
|
Dear DHS users,
Related to the previous posts, I have a question with respect to the code "no measurement found in household". In my study on associations between nutrition and experience of IPV (intimate partner violence), I would like to present an overview of missing values per indicator (for women and children, i.e. height/weight, anemia level, questions on IPV) to discuss the issue of (selection) bias in my estimation.
You mention:
The height and weight information however is collected at the Household level (so as to capture orphans, etc), and thus the "No measurement found in Household" represent Birth History cases for which there were no matching case found in the Height/Weight section.
Could you please elaborate a little bit more? What does this mean exactly for data on women and their children in the IR files?
The problem is that in some country surveys (e.g. Malawi 2010, Mali 2012), a large share of anthropometry and hemoglobin data for women AND their children are missing for this reason (around 50%-60%, evermarried women sample). Since I cannot properly explain why these data are missing, my arguments are weak.
Should I just ignore these missing cases or should I include them in my calculations for "Missing observations as percent of eligible women sample"?
Thank you very much.
|
|
|
|
The DHS Program User Forum
Members
Search
Help
Register
Login
Home
