The DHS Program User Forum
Discussions regarding The DHS Program data and results
Home » Topics » Wealth Index » Wealth index with pooled datasets
Wealth index with pooled datasets [message #11903] Tue, 28 February 2017 07:32 Go to next message
clarapd is currently offline  clarapd
Messages: 8
Registered: November 2015
Member
Hi,

I'm trying to assess the association between a variable X with wealth (controlling for other factors). My study population is composed by three different DHS countries, so I have a pooled dataset.

I would like to know if it's better to use the wealth index as it is in each dataset or to compute a new one with the pooled population of the 3 countries. I supose that if I use the wealth indices given by the DHS, I will be observing the association between my variable X and relative wealth within each country, and if I use the pooled wealth index, I will observe the association between my variable X and relative wealth across the 3 countries.

I am trying the second option, with a new wealth index caculated for the pool, and almost all of the observations of country1 appear in the same wealth quintile, so maybe it is better to use the wealth indices provided by DHS to observe something more interesting.

In the case I decide to use the wealth indices provided by the DHS, is there any problem if they are not computed using the same variables? So, can I use the three variables merged, as if they are the same variable, without problems of comparability? Maybe it is a good idea to use just wealth quintiles to allow comparability.

Thanks for your help,

Clara


Clara Pons Duran

[Updated on: Tue, 28 February 2017 07:35]

Report message to a moderator

Re: Wealth index with pooled datasets [message #12072 is a reply to message #11903] Sat, 25 March 2017 09:10 Go to previous messageGo to next message
Liz-DHS
Messages: 1516
Registered: February 2013
Senior Member
Dear User, A response from Dr. Shea Rutstein:
Quote:

Using the wealth index for each country in the pooled regression would indicate relative position in that country. Calculating a wealth index from the pooled data set is problematic in a couple of areas: A common lower set of variables would need to be used losing some information. Yes, one country could be at the low end and other at the high end. An alternative would be to use the Comparative Wealth Index but the second problem would still remain if the countries are much different from each other. The Comparative Wealth Index allows the wealth for each household to be comparable to that of households in other countries that have the same score.


Re: Wealth index with pooled datasets [message #12075 is a reply to message #12072] Mon, 27 March 2017 02:58 Go to previous message
clarapd is currently offline  clarapd
Messages: 8
Registered: November 2015
Member
Many thanks!

Clara Pons Duran
Previous Topic: Recalculating wealth index for rural subset
Next Topic: Computing sub-national wealth indices for within-state inequality analysis
Goto Forum:
  


Current Time: Tue Mar 19 00:30:23 Coordinated Universal Time 2024