| Wealth deciles [message #11568] |
Wed, 11 January 2017 07:01  |
Mercysh
Messages: 35
Registered: April 2014
|
Member |
|
|
I have checked posts on wealth deciles and saw the Stata code below
xtile dec_rural = hv271 [pweight = weightvariable] if rural==1, nq(10)
xtile dec_urban = hv271 [pweight = weightvariable] if rural==0, nq(10)
gen decile = dec_rural
replace decile = dec_urban if rural==0.
A quick question- do I include the weight, even if I am going to use svy, for complex survey design later
Mercy
|
|
|
|
| Re: Wealth deciles [message #11752 is a reply to message #11568] |
Mon, 06 February 2017 14:15  |
Liz-DHS
Messages: 1516
Registered: February 2013
|
Senior Member |
|
|
A response from Dr. Sarah Staveteig,
Quote:
Yes, the percentiles need to be weighted. SVY is not needed to weight percentiles. To produce the household member weight, multiply the number of dejure household members by the household sampling weight divided by 1,000,000. If there are no dejure members, use the number of defacto household members.
|
|
|
|
| Re: Wealth deciles [message #12006 is a reply to message #11568] |
Sat, 18 March 2017 10:42 |
Mercysh
Messages: 35
Registered: April 2014
|
Member |
|
|
Thank you for the response.
One more, please help me construct separate wealth variables for urban and rural areas to include in a regression model.
I have two wealth variables using the following code but do not work in a model:
gen weight=hv005*hv012
drop if weight==0
xtile wealth_rur=hv271 [fweight=weight] if hv025==2, nq(5)
xtile wealth_urb=hv271 [fweight=weight] if hv025==1, nq(5)
When I add the folowing code, I get one variable which seem to works for older surveys which did not account for urban-rural differences and to be unnecessary for those which already consider different assets for rural and urban areas:
gen wealth_quintile = wealth_rur
replace wealth_quintile = wealth_urb
Mercy
[Updated on: Sat, 18 March 2017 10:42] Report message to a moderator |
|
|
|
The DHS Program User Forum
Members
Search
Help
Register
Login
Home
