| Unweighted woman-years of exposure for the fertility rates (TFR) [message #29382] |
Tue, 11 June 2024 04:33  |
kclakhara
Messages: 30
Registered: June 2016
Location: New Delhi
|
Member |
|
|
Dear Sir/Madam,
I am working on the estimation of total fertility rate (TFR) by background characteristics using India NFHS-3,4 and 5. I want to know about the generating unweighted base numbers to identify which TFR values should be not shown and put in the parenthesis.
Kindly help me to generate unweighted cases for "* Not shown; based on fewer than 125 unweighted woman-years of exposure for the fertility rates and ( ) Based on 125-249 unweighted woman-years of exposure for the fertility rates".
For your kind information, please find attached weighted figures of TFR and base numbers (exposure).
With Regards,
Kailash Chandra Lakhara
Kailash Chandra Lakhara
|
|
|
|
| Re: Unweighted woman-years of exposure for the fertility rates (TFR) [message #29419 is a reply to message #29382] |
Fri, 14 June 2024 09:15  |
Janet-DHS
Messages: 888
Registered: April 2022
|
Senior Member |
|
|
Following is a response from DHS staff member, Tom Pullum:
The conditions for bracketing the TFR in these tables are different in the India reports than in other DHS reports. Someone, presumably on the staff of IIPS, created them.
The age-specific rates, which go into the construction of the TFR, are based on WEIGHTED woman-years of exposure and WEIGHTED numbers of births within five-year age intervals. These numbers are added up within each age group and sub-population. The weighted numbers of births are divided by the weighted woman-years of exposure to get the asfrs. Then the asfrs are added and multiplied by 5 to get the TFR.
Normally the UNWEIGHTED numbers are NOT added up. To save them and add them up (with a collapse command in Stata, for example) you would have to go into the program that produces the asfrs and TFR and add lines of code. You would have to include the specified totals in the draft of the table. Then, either with a program or with editing, you would insert the brackets or mask the estimates depending on the values of the specified totals.
The usual CSPro programs to calculate the rates, and the Stata (etc.) programs on our GitHub site do not produce the unweighted totals. For virtually all tables in DHS final reports, the n's are weighted, not unweighted.
DHS staff cannot take the time to revise the TFR programs for this purpose.
|
|
|
|
The DHS Program User Forum
Members
Search
Help
Register
Login
Home
