The DHS Program User Forum
Discussions regarding The DHS Program data and results
Home » Countries » Ethiopia » Rural Addis PSUs in 2005 Ethiopia Sample
Rural Addis PSUs in 2005 Ethiopia Sample [message #3440] Thu, 11 December 2014 11:25 Go to next message
samanthaski is currently offline  samanthaski
Messages: 2
Registered: December 2014
Location: United States
Member
Hello,

In the 2005 Ethiopia data there are 2 psus that are both in the region of Addis and are rural (both are in stratum 19). Is this accurate? In the 2011 data all Addis households were urban.

As a follow up question, I am having an issue where in the model I am specifying using births data (the model includes ~4900 observations), there is a stratum with just one sampling unit (which is one of the two rural addis ababa clusters). As a result, Stata won't calculate standard errors and p values, etc. with the svy command. Is there a recommended way to deal with this by lumping this cluster into another similar cluster, and should it be a similar cluster in addis ababa (but there are only urban options) or a similar rural cluster in another region? I can't use the other rural addis cluster because there are no observations from that cluster in my model. Or is it better to drop these observations?

Thank you.
Re: Rural Addis PSUs in 2005 Ethiopia Sample [message #3444 is a reply to message #3440] Thu, 11 December 2014 14:09 Go to previous message
Liz-DHS
Messages: 1350
Registered: February 2013
Senior Member
Dear User,
Here is a response from one of our samplers, Dr. Ruilin Ren. Dr. Ren was not involved with the Ethiopia DHS 2005 but offers the following:
Quote:
It is possible that Addis Ababa had a portion of rural areas according to the previous census. Actually there are quite a few capital cities have both urban and rural areas.
As for the single-psu per stratum the data user encountered, I will advise to combine it with the other Addis Ababa clusters, it is better than to combine it with any other clusters outside of Addis. The other option is to drop it if there is not a shortage in sample size.


Previous Topic: Biomarkers / Malaria
Next Topic: Identifying Woredas
Goto Forum:
  


Current Time: Sun Oct 21 22:02:54 Eastern Daylight Time 2018