The DHS Program User Forum
Discussions regarding The DHS Program data and results
Home » Data » Dataset use in Stata » Using the "SVY" module and sampling weights of STATA GUATEMALA
Using the "SVY" module and sampling weights of STATA GUATEMALA [message #13566] Mon, 20 November 2017 11:47 Go to next message
MMazariegos is currently offline  MMazariegos
Messages: 1
Registered: November 2017
Location: Mexico
Member
Good morning,
I am using the DHS 2014-2015 Guatemala Database: GUIR71FL
I just want to be sure if I am using the correct variables to specify the SVY, can you please check the code above?

use "C:\Users\mmazariegos\Desktop\ENSMI 2015\GUKR71FL.DTA"
svyset v021 [pweight = v022], strata(v023) singleunit(centered)

Also, I do not know if I have to do it anything else to be able to calculate proportions of some variable over other. For example, the proportion of stunting over wealth status.

First, I calculate stunting:

*To generate STUNTING*
sum hw70
gen haz= hw70/100
sum haz
replace haz=. if haz<-6
replace haz=. if haz>6
gen stunting= 1 if haz <-2 & (hw1<60 & hw1!=.)
replace stunting= 0 if haz >=-2 & (hw1<60 & hw1!=.)
tab stunting, gen(stunting_cat)

*Stunting by wealth Index*
tab v190
svy:proportion stunting, over(v190) *This is working*


But I want to calculate not only the proportion by wealth status but be able to assess if there are statistical differences using lincom:
svy, subpop(keeptag): mean stuntingcat, over(v190)
lincom [_subpop_1] - [_subpop_2]
lincom [_subpop_1] - [_subpop_3]
lincom [_subpop_2] - [_subpop_3]

With this, I get an error saying "invalid subpop() option"

Thanks in advance,
Monica



Re: Using the "SVY" module and sampling weights of STATA GUATEMALA [message #13830 is a reply to message #13566] Mon, 08 January 2018 15:31 Go to previous message
Bridgette-DHS is currently offline  Bridgette-DHS
Messages: 1510
Registered: February 2013
Senior Member

Following is a response from Senior DHS Stata Specialist, Tom Pullum:

We apologize for the long delay with this response. You have probably already found an answer.

You do not include your definition of "keeptag", as in "subpop(keeptag)". When I have had the error statement you mention, it has been because of an improper specification of the argument of subpop. Here, the data will be reduced to the cases with keeptag=1. You must define the argument so that it takes the value 1 for the cases you want to include. For other cases, you assign keeptag=0 or keeptag=. At least that's how I always specify the subpopulation.

I have never used subpop followed by lincom. It is possible that you also need to explicitly limit the cases for which lincom is applied to the cases with keeptag=1. However, I doubt that that is the problem.
Previous Topic: Percent of postpartum women with unmet need for contraception
Next Topic: creating strata and svyset of SPA datasets of Nepal
Goto Forum:
  


Current Time: Mon Dec 17 00:53:18 Eastern Standard Time 2018