| IPV -Domestic Violence [message #28800] |
Sun, 10 March 2024 21:45  |
Deidi
Messages: 12
Registered: March 2021
|
Member |
|
|
Dear DHS Forum,
I'm working on the topic of intimate partner violence using the Colombia, Guatemala, Dominican Republic, Peru and Bolivia datasets, and I would like to confirm that I am using the right variables and set up for my svyset.
Colombia 2015, Guatemala (2014-15) and Dominican Republic (2013)
gen weight=d005/1000000
gen psu = v021
gen strata = v023 //ta v023 stratification used in sample design
svyset psu [pw = weight], strata(strata)
Peru (2012) & Bolivia (2008)
gen weight=v005/1000000 //d005 is missing." A prior response to this question had suggested to use v005 (sample weight) for the weight variable.
gen psu = v021 //v021 primary sample unit
gen strata = v022
svyset psu [pw = weight], strata(strata)
Thank you in advance for your guidance on this. I really appreciate it.
Best regards,
Deidi
|
|
|
|
| Re: IPV -Domestic Violence [message #28951 is a reply to message #28800] |
Tue, 02 April 2024 12:35  |
Bridgette-DHS
Messages: 3043
Registered: February 2013
|
Senior Member |
|
|
Following is a response from Senior DHS Stata Member, Tom Pullum:
Here is a link to a file that gives the stratum ID for all surveys: HERE. In the recent surveys (the past 10+ years) stratum is correctly given by v022 or v023 (and they are the same) but in some of the older surveys v022 and v023 are mislabelled. In most surveys, the strata are combinations of v024 and v025 (region x residence), whether or not that matches with v022 and v023.
The weight variable for DV is d005, as you say, but if it is missing you have to fall back on v005. The PSU or cluster is v001 or v021--they are the same.
For pweights, as in the svyset command, it is not necessary to divide d005 (or v005, etc.) by 1000000. You can do it, but Stata automatically normalizes pweights to have a mean of 1.
|
|
|
|
The DHS Program User Forum
Members
Search
Help
Register
Login
Home
