Wealth index for non de jure resident respondents [message #9114] |
Fri, 12 February 2016 10:20 |
mmr-UMICH
Messages: 21 Registered: February 2015 Location: A2, MI
|
Member |
|
|
Dear DHS experts,
For individual women data, HH level characteristic variables have a category for Non De Jure Resident respondents, e.g., "Source of drinking water", has a value 97 ="Not a de jure resident". We also know that wealth index is a HH level (composite, relative) characteristic variable for individual women data. For Bangladesh DHS 2011 women data has more than 1300 non de jure resident respondents, but wealth index variables (v190 and v191) have valid score/quintiles corresponding to those non de jure residents. My confusion is that because wealth index is calculated based on interviewed HHs (that is non de jure residents usual HHs were not part of this HH survey). If so how wealth index then has values for those non de jure residents. Can you please explain the intuition of the matter at your conveniences?
A DHS expert said how to treat them as missing in an analysis if we include those HH level variables having non de jure resident category in them:
http://userforum.dhsprogram.com/index.php?t=msg&th=112&a mp;S=780148f5e8474d4add9f09d14463f456#msg_289
I am using individual women data in analysis that includes wealth index as an independent variable, do you suggest me to discard those cases correspond to non de jure resident respondents from my analysis? If so, please also let me know if any changes/adjustment needed for normalized weight in analysis?
Thanking you so much.
Moshiur
|
|
|
Re: Wealth index for non de jure resident respondents [message #9136 is a reply to message #9114] |
Mon, 15 February 2016 10:57 |
Liz-DHS
Messages: 1516 Registered: February 2013
|
Senior Member |
|
|
Dear User,
Here is a response from Dr. Shea Rutstein:
Quote:Non de-jure respondents are assigned the wealth index of the household where they spent the night before interview, since it is likely to be at approximately the same level of wealth as their own (for which we have no information). Note that all respondents in the household get the households wealth index. Additionally, most missing data is assigned the does not have category. For continuous variables (such as members per sleeping room), the mean is assigned before calculating the index for that household but after running the PCA for the score coefficients.
Thank you!
|
|
|
|