| Only 2 cases in a PSU in DHS 2014 data [message #9446] |
Tue, 29 March 2016 14:18  |
mmr-UMICH
Messages: 21
Registered: February 2015
Location: A2, MI
|
Member |
|
|
Hi DHS Experts,
Thank you for releasing the Bangladesh DHS 2014 survey data. I have a simple question as I found that the PSU=255 has only 2 respondents in individual women sample dataset. Other PSU sample sizes range from 15 to 44 respondents. Is it correct? (FYI, in HH dataset it (PSU=255) has 3 cases).
Thanking you.
Moshiur
|
|
|
|
| Re: Only 2 cases in a PSU in DHS 2014 data [message #9528 is a reply to message #9446] |
Fri, 08 April 2016 13:29  |
Bridgette-DHS
Messages: 3029
Registered: February 2013
|
Senior Member |
|
|
Following is a response from Senior DHS Stata Specialist, Tom Pullum:
That appears to be correct. Cluster 255, an urban neighborhood in Dhaka, has only 3 households, 16 household members, and 2 women in the IR file. It is very unusual to have such small numbers. These are by far the smallest counts in all 600 clusters in this survey. I cannot easily find the reasons for this. DHS saves field check tables, but the results for one cluster, without comparison with the other 599 clusters, would not be very helpful. Fortunately, even if this was a cluster with an extremely high refusal rate, which is a likely inference, the net effect should be negligible. Also note that the implementing agency did not discard the cluster or consolidate it with another cluster, strategies which would have made this strange outcome invisible. Something strange must have happened, but at least the cluster is there for all to see!
|
|
|
|
|
|
The DHS Program User Forum
Members
Search
Help
Register
Login
Home
