| Comparing DHS6 and DHS7 district wise data [message #30124] |
Sat, 28 September 2024 11:59  |
HS
Messages: 4
Registered: September 2024
|
Member |
|
|
Hello, I have 3 questions:
1)I need to calculate and analyse Female Labour force Participation rate (as per DHS6 data) in each district of Pakistan with FLFPR as per DHS7 data using STATA. Is there any efficient way I can ensure that the districts are the same for both datasets?
2) Moreover, I want to merge the LFPR rates with CR file. Which variable can be used as an identifier?
3) Similarly, how to merge data about household head with CR file?
Thanks
|
|
|
|
| Re: Comparing DHS6 and DHS7 district wise data [message #30147 is a reply to message #30124] |
Thu, 03 October 2024 14:16  |
Janet-DHS
Messages: 895
Registered: April 2022
|
Senior Member |
|
|
Following is a response from DHS staff member, Tom Pullum:
There have been earlier forum questions on this issue. Unfortunately, the implementing agency in Pakistan has not released the label for the district variable in the Pakistan 2017-18 (DHS-7) survey. Therefore we cannot confirm that the numerical codes match with the same names as in the 2012-13 (DHS-6) survey. We do not know why this information was restricted for the 2017-18 survey, especially since it was freely provided for the 2012-13 survey. It is possible that it will be released eventually, but the most recent request was made about a year ago. If any Pakistani researchers can help with this effort, it will be appreciated.
For your second question, a rate is a characteristic of a group. Perhaps what you want to do is to merge a woman in the IR or PR file with the same woman in the CR file? Please clarify.
I also do not understand your third question. You can construct a file of household heads from the PR file by selecting the PR file members who have hv101=1. The CR file is a file of couples (a woman and a man) who name each other as partners. In most of these couples, the man will be the head of the household; in some of them the woman will be; in some of them neither will be (the household head is someone else). Are you trying to identify couples which include the head? Couples which do not include the dead? Are you trying just to attach some household-level variables (from the HR or PR file) to the CR records? If so, you can get household-level variables for the woman and man themselves, because both of them appear in the HR and PR files, and you do not need to merge with the household head. Please clarify.
|
|
|
|
| Re: Comparing DHS6 and DHS7 district wise data [message #30152 is a reply to message #30147] |
Sat, 05 October 2024 04:51 |
HS
Messages: 4
Registered: September 2024
|
Member |
|
|
Thank you so much for your reply.
Regarding my third question, I want to use some household head characteristics (like education ) with couples data. Not all households have reported husband/wife as the head. I will therefore need to merge PR file with CR file.
What I gathered from the discussions on the forum, hv044==1 & hv101==1 will give us household heads for households that were selected for dom violence module. This gives us data for 5,224 individuals. CR file has data for only 3,334 couples. The number of heads cannot be higher than the number of couples. Is the difference due to non-response? Or is there an error in my method?
I am planning to then concat cluster number and household id in CR to create a unique identifier for merging with PR. Can you pls confirm if that is the correct approach?
Thanks
|
|
|
|
| Re: Comparing DHS6 and DHS7 district wise data [message #30193 is a reply to message #30152] |
Fri, 11 October 2024 16:11 |
Janet-DHS
Messages: 895
Registered: April 2022
|
Senior Member |
|
|
Following is a response from DHS staff member, Tom Pullum:
The following lines will do this merge. I did not restrict to household heads with hv044=1, but you could do that. These lines will attach the household head's data to the record for each of the 3,334 couples in the CR file.
Many households do not have a couple in the CR file. There are actually many more household heads than couples. What you were doing was correct; I have only confirmed that.
* Stata lines to merge household head data with with couples, Pakistan 2017-18 survey
* Specify a workspace
cd e:\DHS\DHS_data\scratch
* Prepare a file of household heads
use "...PKPR71FL.DTA", clear
keep hv*
keep if hv101==1
gen cluster=hv001
gen hh=hv002
gen in_PR=1
save PK7A_head.dta, replace
* Prepare CR file
use "...PKCR71FL.DTA", clear
gen cluster=v001
gen hh=v002
gen in_CR=1
merge m:1 cluster hh using PK7A_head.dta
tab _merge
tab in_PR in_CR,m |
|
|
|
The DHS Program User Forum
Members
Search
Help
Register
Login
Home
