| Re: Clustering Respondents in Child Data for Analysis [message #2840 is a reply to message #2839] |
Fri, 29 August 2014 21:37   |
Reduced-For(u)m
Messages: 292
Registered: March 2013
|
Senior Member |
|
|
If you are accounting for survey design by clustering on the Primary Sampling Unit then you shouldn't need to worry about it. If you don't want to use the "svy" command described in the FAQs, then you could do it manually, using something like:
reg Y X [pweight=weight], cluster(PSU)
...this will work without the weighting, and clustering in this manner on the PSU subsumes the household/maternal clustering (meaning it also accounts for that clustering). Depending on the right-hand-side variables of interest, you may need to cluster at a "higher" level, but for standard regression approaches clustering on PSU is generally recommended by the DHS folks.
|
|
|
|
The DHS Program User Forum
Members
Search
Help
Register
Login
Home
