strata and psu in 2005 dhs [message #15112] |
Tue, 05 June 2018 03:18 |
heleen.vermandere
Messages: 1 Registered: June 2018 Location: cairo
|
Member |
|
|
Hi,
I noticed that the number of psu (v021) per strata (v022) is very often (almost always) 1; this generates problems during analyses (no standard errors). I just wanted to verify that I am using the right variables and coding (stata).
gen weight = v005/1000000
svyset v021 [pweight= weight], strata(v022)
thx in advance,
Heleen
|
|
|
Re: strata and psu in 2005 dhs [message #15122 is a reply to message #15112] |
Wed, 06 June 2018 08:16 |
Bridgette-DHS
Messages: 3188 Registered: February 2013
|
Senior Member |
|
|
Following is a response from DHS Senior Sampling Specialist, Mahmoud Elkasabi:
In general, v022 is the correct stratification variable you should use. However, I would advise you to always compare the variable versus the stratification description in Appendix A.
Regarding the situation you described, on the national level, in most of the cases, you should not find single-cluster strata. It might happen though. On the sub-national level, where the analysis is conducted on a lower level for a sub-population or analysis domain, you should expect to see more single-cluster strata.
|
|
|