Home » Data » Sampling » Two or Three Sampling Stages
| Re: Two or Three Sampling Stages [message #2033 is a reply to message #2029] |
Wed, 16 April 2014 11:14   |
 |
user-rhs
Messages: 132
Registered: December 2013
|
Senior Member |
|
|
Quote:And as for a reliable source, I looked into it and think the respective censuses are the best approach since the DHS survey are mostly based on them
Ideally, we would use the census figures, but in my experience, the amount and quality of information on the respective Bureau of Statistics/Ministry of Health websites varies by country. Sometimes the age range breakdown is 15-64 instead of 15-49. Also, the census won't be available for all the years that you need. Even in the U.S., the census is done every 10 years with intercensal estimations done annually.
I wonder if DHS can make these figures available for all DHS surveys ever conducted, since the general advice is to de-normalize the sampling weight for pooled analysis. I haven't looked for census HH numbers (haven't had to), but my hunch is that it would be harder to come by than the # of individuals.
One limitation I can think of from relying on the census is that although the most recent census is used as the sampling frame, sometimes it is necessary to make adjustments in the field on the actual # of HH, especially if some time has passed since the census. In stable populations, this is probably OK, but where there is a lot of in and out-migration or population growth/loss, this might be an issue.
[Updated on: Wed, 16 April 2014 11:16] Report message to a moderator |
|
|
|
Current Time: Mon Feb 17 07:33:00 Coordinated Universal Time 2025
|
The DHS Program User Forum
Members
Search
Help
Register
Login
Home
