| Children Ever Born [message #23585] |
Thu, 14 October 2021 20:55  |
Tesfay
Messages: 7
Registered: April 2021
|
Member |
|
|
Dear DHS support group, to see the association between women empowerment and fertility in Ethiopia, I want to use children ever born (CEB) as an count outcome variable in Poisson regression. I have two questions: 1 should I estimate the CEB this for all women or "for currently married women" or for "ever married women" using the all women factor? If I have to use the 2nd one, how can I compute in the stata. Would you help me with the syntax of estimating CEB for ever married women?
Data source = EDHS 2016
Thank you for prompt response.
|
|
|
|
| Re: Children Ever Born [message #23600 is a reply to message #23585] |
Tue, 19 October 2021 10:47  |
Bridgette-DHS
Messages: 3199
Registered: February 2013
|
Senior Member |
|
|
Following is a response from DHS Research & Data Analysis Director, Tom Pullum:
The all-women factors are only relevant for surveys that are limited to ever-married women, such as the Jordan surveys. The surveys in Ethiopia have included all women, regardless of marital status. You can of course limit your analysis to ever-married women if you want to, or any other subpopulation, but there is no need to use awfact. The individual-level variable for children ever born is v201. It and the variables related to empowerment are in the IR file (ETIR71FL.dta).
You would use the poisson command, for example "poisson v201 i.v106" to include education as a categorical predictor. I strongly recommend including controls for age and some other things such as region. Also include the svyset adjustment. You should be able to find relevant earlier postings on the forum.
|
|
|
|
The DHS Program User Forum
Members
Search
Help
Register
Login
Home
