The DHS Program User Forum
Discussions regarding The DHS Program data and results
Home » Data » Weighting data » is re-weighting possible if strata have changed?
is re-weighting possible if strata have changed? [message #12380] Mon, 08 May 2017 10:48 Go to next message
AnnM
Messages: 3
Registered: March 2013
Location: Boston
Member
I'm working with a survey in which one EU was placed in the wrong stratum (new District boundaries were drawn after sample was selected). The sampling weights were stratum specific. Is there a way to re-allocate the EU from stratum 1 to stratum 2 and recalculate the sampling weights so that the new District boundaries are used? I've searched the threads on this and it may be that the answer to this question is out there, but I'm searching for the wrong terms. Many thanks for any advice!
Re: is re-weighting possible if strata have changed? [message #13370 is a reply to message #12380] Sun, 29 October 2017 07:43 Go to previous messageGo to next message
Bridgette-DHS is currently offline  Bridgette-DHS
Messages: 2548
Registered: February 2013
Senior Member
Email sent to user.
Re: is re-weighting possible if strata have changed? [message #13553 is a reply to message #13370] Fri, 17 November 2017 13:56 Go to previous message
Bridgette-DHS is currently offline  Bridgette-DHS
Messages: 2548
Registered: February 2013
Senior Member
Following is a response from Senior DHS Stata Specialist, Tom Pullum:

Hi Ann--Sorry for the delayed response to this forum question!

If you are sure about this change in classification, then you could do a recode of the stratum variable. You should never change the original data file, so you should save the changed file with a different name. Say that the cluster or enumeration area (did you mean EA rather than EU?) is given by v001=v021 and the stratum is given by v023 (which is not always the case). Say that EA #12 was put into stratum 20 but should be in stratum 22.

I would construct a revised stratum variable called v23rev like this, in Stata:

gen v023rev=v023
replace v023rev=22 if v023==20 & v001==12

It would be a little simpler to just say

gen v023rev=v023
replace v023rev=22 if v001==12

but I consider it to be a little safer to use the first version.

Please let me know if this does not answer the question. With regards, Tom
Previous Topic: Merged Weights: Couple & HIV Recode (Stata)
Next Topic: Weighting across surveys when only including youth in analysis
Goto Forum:
  


Current Time: Mon Aug 8 12:29:51 Coordinated Universal Time 2022