The DHS Program User Forum
Discussions regarding The DHS Program data and results
Home » Data » Sampling » Oaxaca decomposition and stratification.
Oaxaca decomposition and stratification. [message #15364] Fri, 06 July 2018 08:50 Go to next message
Mayank_Ag is currently offline  Mayank_Ag
Messages: 19
Registered: May 2018
Location: Hyderabad
Hello all,

I am using DHS 15-16 for India. I am doing an Oaxaca-Blinder decomposition for my research project. I wanted to know is it necessary to take into account stratification and clustering while doing such a decomposition. If yes, how to do it. I noticed that the Decompose and Oaxaca commands in Stata do not have a svy option.

Thanks in advance!!

Re: Oaxaca decomposition and stratification. [message #15937 is a reply to message #15364] Mon, 08 October 2018 12:47 Go to previous message
Bridgette-DHS is currently offline  Bridgette-DHS
Messages: 1510
Registered: February 2013
Senior Member
Please see the following response from Senior DHS Stata Specialist, Tom Pullum and Senior DHS Sampling Specialist, Mahmoud Elkasabi:

We apologize for the delay with this response to your July 6 email. You may have found an answer elsewhere. You could enter "help mvdcmp" in Stata. That will give you a link to a package written by Daniel Powers and colleagues. That package (you can also get to it with "help oaxaca") may be friendlier than oaxaca. If a an estimation procedure does not accept svy, it may accept weights and may accept the cluster() option, so you are only missing the stratum specification. Our standard recommendation would be to use the features in the current version of a procedure, and not decide against the procedure just because you cannot do a full svyset. We used mvdcmp a few years ago in a further analysis of neonatal mortality in Rwanda and you may want to look at that report (
Previous Topic: Sampling weights for Senegal 1986
Next Topic: Calculating Sampling Errors at the State Level
Goto Forum:

Current Time: Sat Dec 15 07:25:45 Eastern Standard Time 2018