The DHS Program User Forum
Discussions regarding The DHS Program data and results
Home » Topics » Wealth Index » Ethiopian Wealth Index (How to use it for 2000, 2005, and 2011)
Ethiopian Wealth Index [message #24927] Wed, 03 August 2022 16:54 Go to next message
Daniel Pérez is currently offline  Daniel Pérez
Messages: 10
Registered: May 2022
Hi there,

I am working with the Ethiopian DHS (2000, 2005, 2011) and I have just realized that the wealth index was not constructed in the same way throughout the years. For instance, the WI of DHS 2000 ranges from -1 to 4 while DHS 2005 and DHS 2011 have a much broader range of values. How can I compare them? Are the WI of DHS 2005 and DHS 2011 comparable/constructed following the same method?


Re: Ethiopian Wealth Index [message #24961 is a reply to message #24927] Wed, 10 August 2022 16:32 Go to previous message
Janet-DHS is currently offline  Janet-DHS
Messages: 737
Registered: April 2022
Senior Member
Following is a response from DHS Research & Data Analysis Director, Tom Pullum:

For the 2000 survey there is a file (ETWI41FL.dta) that has the wealth index. Maybe you have already used that and merged it with the other ET41 files. In the WI file, "wlthindf" is the continuous index (the first principal component of a principal component analysis) and is equivalent to v191 or hv271. "wlthind5" is the 5-category index, with values 1,2,3,4,5, and is equivalent to v190 or hv270. I see that the standard indexes are included in the 2005 (ET51) and 2011 (ET61) data.

Data files with the wealth index include both the continuous version and the 5-category version. The 5-category version is derived from the continuous version by identifying 4 cut points on the continuous version that lead to 5 categories with equal numbers of de jure household members (see the Guide to DHS Statistics for more). The continuous version is included so users can try alternative constructions, for example terciles (3 categories), but otherwise the continuous version is hardly ever used.

The procedures to construct the wealth index are very similar across surveys, but are not exactly the same. The indicators of household assets can vary slightly and the loadings will vary. Details for each survey can be accessed from the website.
Previous Topic: Uganda MIS 2018-19 urban/rural sample size discrepancies
Next Topic: Replicating the computation of the wealth index using STATA
Goto Forum:

Current Time: Sun Jun 16 14:26:38 Coordinated Universal Time 2024