| Armenia 2015-2016 DHS womens weights [message #19569] |
Mon, 13 July 2020 11:43  |
Mary
Messages: 2
Registered: July 2020
|
Member |
|
|
|
The womens weights v005 in the childrens file: AMKR72DT sum to 1672 when divided by 1,000,000 not the number of cases 1724. I would like to know if it is common that the normalized weights do not sum to the number of cases in DHS surveys.
|
|
|
|
| Re: Armenia 2015-2016 DHS womens weights [message #19620 is a reply to message #19569] |
Wed, 22 July 2020 07:34  |
Bridgette-DHS
Messages: 3199
Registered: February 2013
|
Senior Member |
|
|
Following is a response from DHS Research & Data Analysis Director, Tom Pullum:
The total weight should equal the sample size in the IR file (except for the factor of 1000000, but not necessarily in the KR and BR files. In the KR file, which I assume is the child file you are talking about, the units are children, not women, and each child is given the weight of the mother. Some women in the IR file did not have a child in the past five years and therefore they do not appear in the KR file at all. Some women had more than one child in the past five years and appear more than once. The total of v005 in the KR file will always be at least a little different from the number of children in the Births file.
If you are working with Stata and you use "pweight=v005", the analysis will automatically normalize v005 in whichever file you are using. You don't have to worry about doing any kind of re-normalization yourself.
|
|
|
|
The DHS Program User Forum
Members
Search
Help
Register
Login
Home
