| Wealth index variable without water/sanitation (Tanzania 2015-2016) [message #13988] |
Wed, 31 January 2018 22:25  |
ec2017
Messages: 4
Registered: October 2017
|
Member |
|
|
|
Since my independent variable is related to water/sanitation, I am exploring options for creating a new wealth index variable that does not include these characteristics. I have two questions. 1) Does your data team have any examples of how this has been done before? I am concerned that simply removing the water/sanitation questions will not result in a valid measure of household wealth. 2) Is the code that was used to create the wealth index for Tanzania's 2015-2016 survey available yet? It looks like it is not yet posted and I am guessing it is not available, but I wanted to check. Thank you.
|
|
|
|
|
|
| Re: Wealth index variable without water/sanitation (Tanzania 2015-2016) [message #15438 is a reply to message #13988] |
Wed, 18 July 2018 17:31  |
scinnema
Messages: 1
Registered: July 2018
|
Member |
|
|
ec2017 wrote on Wed, 31 January 2018 22:25Since my independent variable is related to water/sanitation, I am exploring options for creating a new wealth index variable that does not include these characteristics. I have two questions. 1) Does your data team have any examples of how this has been done before? I am concerned that simply removing the water/sanitation questions will not result in a valid measure of household wealth. 2) Is the code that was used to create the wealth index for Tanzania's 2015-2016 survey available yet? It looks like it is not yet posted and I am guessing it is not available, but I wanted to check. Thank you.
Hello!
Before I express my questions directly, let me explain about the context of my doubts that are a little bit related to the previous post.
I am doing an investigation about sanitation and child health, it will use a logit model in which I choose sanitation variables (source of drinking water and type of toilet facility) as principal independent variables. Then, in order to control the model I will use the wealth index as control variable. I understand that the wealth index is a composed variable where sanitation variables are present. So, when I run the model, it seems to me that sanitation variables will explain the dependent model twice.
- Is this a real problem in the sense that the wealth index (control variable) could affect sharply the results of dependent variable?
- Does anyone or any study consider this issue?
- Just like "ec2017" said, an idea that could solve this point is to create a new wealth index that excludes sanitation variables. This idea I have not tried yet since I did not find a clearly/complete guide to code programmation in Stata software. But I was wondering if this new index could be different from original or be the same (scores and quintiles)?
I hope that someone can give me hand
Thanks in advance
|
|
|
|
The DHS Program User Forum
Members
Search
Help
Register
Login
Home
