The DHS Program User Forum
Discussions regarding The DHS Program data and results
Home » Countries » Ethiopia » Weighting and Multiple Imputation
Weighting and Multiple Imputation [message #15033] Sat, 26 May 2018 09:49 Go to next message
Hassen
Messages: 121
Registered: April 2018
Location: Ethiopia,Africa
Senior Member
Dear DHS Experts,I have no clear understanding regarding weighting and Multiple imputation.
1)In this forum Sometimes I have seen "iweight", sometimes "pweight" etc. Is there any difference between them?
2)Is it possible to make multiple imputation for DHS Data? If so,what assumptions should be fullfiled?
In SPSS I have tried,but it is impossible to run multiple imputation,if the missing value for one variable <10%.
How can we know whether the missing value is "Missing Completely At Random (MCAR),Missing At Random (MAR) or Missing None At Random (MNAR) for DHS Data?
All my Hero,I need your response to improve my thesis!!
Respectfully,Hassen


Hassen Ali(Chief Public Health Professional Specialist)

[Updated on: Sat, 26 May 2018 09:52]

Report message to a moderator

Re: Weighting and Multiple Imputation [message #15057 is a reply to message #15033] Tue, 29 May 2018 08:22 Go to previous messageGo to next message
Hassen
Messages: 121
Registered: April 2018
Location: Ethiopia,Africa
Senior Member
Dear all DHS experts,I am waiting your constructive response for the above Question (message#15033).
With best wishes,Hassen


Hassen Ali(Chief Public Health Professional Specialist)
Re: Weighting and Multiple Imputation [message #15072 is a reply to message #15057] Wed, 30 May 2018 07:33 Go to previous messageGo to next message
Bridgette-DHS is currently offline  Bridgette-DHS
Messages: 3017
Registered: February 2013
Senior Member

Following is a response from Senior DHS Stata Specialist, Tom Pullum:

In Stata, there are 4 kinds of weights--aweight, fweight, pweight, and iweight. You can get a description with "help weight". In statistical models and svyset you would only use pweight. There have been many postings on weights. I don't know the weight options in SPSS, but perhaps another user can help with them.

In DHS data there are usually not enough "missing" (as distinct from "not applicable") cases to worry about. If you just drop those cases, the results will be equivalent to using multiple imputation with the MCAR assumption. In general, of course, you should not make the MCAR assumption; there are tests for that assumption. I would advise against using multiple imputation unless you are doing some pretty advanced multivariate analysis. We hardly ever use it here at DHS and cannot advise you on how to use it.
Re: Weighting and Multiple Imputation [message #15079 is a reply to message #15072] Wed, 30 May 2018 12:54 Go to previous message
Hassen
Messages: 121
Registered: April 2018
Location: Ethiopia,Africa
Senior Member
Dear Bridgette and Dr.Tom Pullum,Thank you very much for your attractive response!! It is extremely helpful.
I learned a one week lecture of STATA software,which was given for post-graduate students in our University for the last 1(one)week. Thus,I have shifted from SPSS to STATA Version 14 for my MPH Thesis Analysis.
Again,Thank you very much for your assistance and Guidance.
With Best Wishes,Hassen


Hassen Ali(Chief Public Health Professional Specialist)
Previous Topic: Women Empowerment Index
Next Topic: Q123 on household bank account
Goto Forum:
  


Current Time: Fri Mar 29 06:28:03 Coordinated Universal Time 2024