The DHS Program User Forum
Discussions regarding The DHS Program data and results
Home » Data » Sampling » Stratification
Re: Stratification [message #1467 is a reply to message #1464] Fri, 28 February 2014 14:35 Go to previous messageGo to previous message
hlyons is currently offline  hlyons
Messages: 5
Registered: February 2014
Location: Seattle
Member
Thanks Trevor, this explanation does help. In terms of prescription, though, should one use the strata as coded or the strata as described in the report? I interpret implicit strata to mean an organization and classification of PSU's undergone sometime after selection/sampling of PSU's (correct me if I'm wrong). I would want to use the strata that best reflects the true sampling design and my suspicion is that these should be generally region + urban/rural regardless of what strata are actually listed, which might have been an ex post facto creation.

Thanks again!
Hil

P.S. What this also implies to me is that reports that use strata not originally in the design may also have incorrect or inadvisable standard errors (though maybe not far off)? I can't recall off the top of my head whether with-replacement standard error estimation will be somewhat robust to overdoing the strata for typical estimators (totals, ratios).
 
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Previous Topic: Specific description of the data
Next Topic: Two or Three Sampling Stages
Goto Forum:
  


Current Time: Wed Nov 27 08:51:19 Coordinated Universal Time 2024