The DHS Program User Forum      
Discussions regarding The DHS Program data and results
Home » Topics » Wealth Index » Identifying items for wealth factor score calculation
Identifying items for wealth factor score calculation [message #14230] Wed, 07 March 2018 13:14 Go to next message
dhruschk is currently offline  dhruschk
Messages: 5
Registered: May 2015
As part of a current project, we have been working to replicate the DHS index calculations, and a few questions (only a few!) have come up that I was hoping you might be able to answer (or at least point me to the right person). I've listed them in order of importance.

1. A number of surveys have a variable for domestic laborer in the household that are not present in the Household recode datasets. The household roster section does not have a specific designation for this, though some surveys have the question included in the questionnaire, it appears there is no place to record it. Where did this variable come from?

2. PY 1990 and JO 1990 Both have a set of households that are coded as having the same value for the DHS Index but clearly have separate suites of assets.

3. NP 2001 It appears livestock and agricultural land variables are not included in the household dataset and we assume they come from the women's dataset. However how are households handled if they don't have women? Where are the livestock and land variables taken from?

Any advice you have on these would be greatly grateful.
Re: Identifying items for wealth factor score calculation [message #15331 is a reply to message #14230] Fri, 29 June 2018 13:20 Go to previous message
Messages: 1290
Registered: February 2013
Senior Member
A response from Dr. Shea Rutstein:

The answers are the following:

1. For the household to have a domestic (live-in) when not on the time sheet, is obtained from the individual respondent or spouse a) when the occupation is personal services (or household worker), and from the household schedule b) is a usual resident and c) is not related to the head of the household.
2. The early DHS surveys were not designed to obtain a wealth index. Both Paraguay and Jordan 1990 surveys have limited number of amenities and services available to classify households by wealth. Therefore, heaping on certain values is to be expected and is one of the reasons the questions about amenities and services was expanded in later surveys.
3. Individual women respondents were asked if they owned land or livestock by themselves or jointly. They get a 1 value if so and 0 otherwise. Note that these variables are different from those derived from a typical HH questionnaire for size of land area owned or number of animals by type. In Nepal 2001, there were no household questions on land size and number of livestock, so that only the individual responses can give any information. Therefore, for households with no individual respondents (only households with no women or only women 50+), the values given are 0 (no) to owned land and livestock by individuals. This calculation is imperfect but was considered the best that could be done with the available data. Another possibility would have been to assign the mean percentage of the values of the dichotomous variables to these households but that would also be imperfect.

Best regards,
Previous Topic: Ghana DHS 2014 wealth index
Next Topic: CWI
Goto Forum:

Current Time: Thu Aug 16 17:19:33 Eastern Daylight Time 2018