The DHS Program User Forum: Weighting data » Why I am getting different total observations when using iweight for tabulating a variable

Home » Data » Weighting data » Why I am getting different total observations when using iweight for tabulating a variable

Show: Today's Messages :: Show Polls :: Message Navigator

Re: Why I am getting different total observations when using iweight for tabulating a variable [message #26121 is a reply to message #26119]

Thu, 09 February 2023 13:12

Bridgette-DHS
Messages: 3230
Registered: February 2013

Senior Member

Following is a response from Senior DHS staff member, Tom Pullum:

I am pretty sure that this is another manifestation of the fact that hv005 is normalized (so that total units = total weight, except for the factor of 1000000) for households, not individuals. That is, if you did the same thing in the HR file, you would get the consistency you expect, even though you do not get it in the PR file. You are ok--the discrepancy is small and we (the DHS research team) ignore it. If you want, you can re-normalize hv005 in the PR file by multiplying by a factor. The following lines will construct a normalized weight for the PR file.

gen unwtd=1000000
total unwtd hv005
matrix B=e(b)
matrix list B
scalar list sfactor
gen hv005_PR=round(sfactor*hv005)

I usually advise against using aweight. You can read what Stata says ("help weight") about aweight.

Report message to a moderator

[Message index]

		Why I am getting different total observations when using iweight for tabulating a variable By: sujata on Tue, 07 February 2023 12:48
		Re: Why I am getting different total observations when using iweight for tabulating a variable By: Bridgette-DHS on Thu, 09 February 2023 09:11
		Re: Why I am getting different total observations when using iweight for tabulating a variable By: sujata on Thu, 09 February 2023 11:45
		Re: Why I am getting different total observations when using iweight for tabulating a variable By: Bridgette-DHS on Thu, 09 February 2023 13:12
		Re: Why I am getting different total observations when using iweight for tabulating a variable By: sujata on Fri, 10 February 2023 01:31
		Re: Why I am getting different total observations when using iweight for tabulating a variable By: Bridgette-DHS on Fri, 10 February 2023 07:58
		Re: Why I am getting different total observations when using iweight for tabulating a variable By: sujata on Fri, 10 February 2023 10:49
		Re: Why I am getting different total observations when using iweight for tabulating a variable By: sujata on Thu, 30 March 2023 07:02
		Re: Why I am getting different total observations when using iweight for tabulating a variable By: Bridgette-DHS on Thu, 30 March 2023 11:04
		Re: Why I am getting different total observations when using iweight for tabulating a variable By: sujata on Fri, 31 March 2023 07:56
		Re: Why I am getting different total observations when using iweight for tabulating a variable By: Bridgette-DHS on Fri, 31 March 2023 12:03
		Re: Why I am getting different total observations when using iweight for tabulating a variable By: sujata on Sat, 01 April 2023 02:23
		Re: Why I am getting different total observations when using iweight for tabulating a variable By: Bridgette-DHS on Mon, 03 April 2023 08:35

Previous Topic:	Multilevel model with region random effects
Next Topic:	Response rate and weights

Goto Forum:

-=] Back to Top [=-

[ Syndicate this forum (XML) ] [

] [

]

Current Time: Thu Dec 18 05:09:20 Coordinated Universal Time 2025