The DHS Program User Forum
Discussions regarding The DHS Program data and results
Home » Topics » Child Health » declaring child survey in Stata
Re: declaring child survey in Stata [message #13245 is a reply to message #13242] Sun, 08 October 2017 21:11 Go to previous messageGo to previous message
Reduced-For(u)m
Messages: 291
Registered: March 2013
Senior Member

You should just cluster by region. First you should make sure all the regions are the same in all 3 rounds...sometimes they change, and you'd have to adjust for that (just because two regions are "region 3" in two different surveys doesn't mean they are exactly the same region). Since no PSU spans two regions, you are implicitly clustering on PSU (you only have to cluster on the "higher" level). So set your clustering to region (with the caveat above about matching regions carefully), you don't need CGM stuff... (at least not for multi-way clustering....maybe for small number of clusters if you have less than, say, 40ish regions).

If the "svy:" prefix works, then it should work right. But I tend to use "xtreg" without the "svy:" prefix and set the clustering level and weighting myself. I haven't used the regxfe, but you should be able to set it all in the command itself you want (as options, not using 'svy').

 
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Previous Topic: Merging data sets: Namibia DHS 2013
Next Topic: Birth weight data for Mexico, 1987?
Goto Forum:
  


Current Time: Sat Apr 17 13:45:50 Coordinated Universal Time 2021