| Re: About Ethiopian DHS 2011 sample design section (pp.7) and Parent raw data set inconsistence [message #11872 is a reply to message #11858] |
Wed, 22 February 2017 12:36  |
Bridgette-DHS
Messages: 3230
Registered: February 2013
|
Senior Member |
|
|
Following is a response by Senior DHS Sampling Specialist, Ruilin Ren:
The Ethiopia DHS 2011 was originally designed for 650 clusters. After the selection of the sample clusters, the survey was redesigned for some reasons, and 26 clusters were dropped. So the Ethiopia DHS 2011 actually included 624 clusters as stated in Appendix A. However, during the survey, another 28 clusters could not be interviewed due to several reasons, so the actual and final number of clusters included in the survey was 596.
You are right in stating that you found 54 missing cluster numbers. The clusters were numbered 1 through 650 at the onset of the survey, but the cluster numbers were not re-assigned after exclusion of clusters before data collection. As a result, the cluster numbers in Ethiopia DHS 2011 were not continuous.
We may need to improve the accuracy of Appendix A, by indicating changes made, and also reporting the survey results (number of clusters not interviewed), in the write up of Appendix A.
|
|
|
|
The DHS Program User Forum
Members
Search
Help
Register
Login
Home
