| Re: Weighting Data for Pooled Indonesian DHS Dataset [message #10056 is a reply to message #10054] |
Mon, 20 June 2016 10:34   |
adelia
Messages: 4
Registered: May 2016
|
Member |
|
|
Thank you for your reply.
So what I am trying to do is to decompose the change in inequality (measured using Erreygers Index) of infant mortality and under-5 mortality, whether the change is due to the change in the association (the coefficient) of wealth and mortality or due to the change of wealth (i.e. more wealthy people in 2012 and 2007). This analysis are usually done using income instead of wealth index, thus the change in income can be analyzed easily since income has absolute value. However, wealth index is based on rank, has no absolute value, and the distribution of its value will always be the same as well so I cannot really examine the change in wealth of the population. I try to analyze the change in wealth by computing pooled wealth index using 2007 and 2012 dataset and then see whether there are more people with higher wealth index in 2012 than 2007. In order to calculate this pooled wealth index, I think I need to use the re-normalize sample weight is it correct? Do the weighted sample sizes need to be proportional to the number of women in the population or the same weighted number of cases in each survey is enough for this?
Many thanks
|
|
|
|
The DHS Program User Forum
Members
Search
Help
Register
Login
Home
