State Weight for Domestic Violence [message #16447] |
Wed, 16 January 2019 11:32 ![Go to next message Go to next message](/theme/default/images/down.png) |
Jasminc
Messages: 5 Registered: January 2019
|
Member |
|
|
Hello,
We are trying to conduct a prevalence analysis using the Domestic Violence module in NHFS-4 (2015-16) India survey dataset. We understand that we should use the special weight to account for DV respondents, but we need further clarification on sampling weights.
We are trying to compare the prevalence across states using the nationally representative sample.
Should we use the SD005 (State Weight for DV) or D005 (Weight for DV)?
Is D005 national DV weight? or household weight for DV subsample?
Thank you for your help!
Sincerely,
Jasmin
[Updated on: Wed, 16 January 2019 14:17] Report message to a moderator
|
|
|
Re: State Weight for Domestic Violence [message #16450 is a reply to message #16447] |
Thu, 17 January 2019 08:49 ![Go to previous message Go to previous message](/theme/default/images/up.png) |
Bridgette-DHS
Messages: 3139 Registered: February 2013
|
Senior Member |
|
|
Following is a response from Senior DHS Stata Specialist, Tom Pullum:
You would use the state weight for an analysis restricted to a single state. For a national analysis, or an analysis with a group of states, use the national weight. The SD or D weights would be used for an outcome in the DV module. When you are confronted with an issue like this, I recommend that you calculate a mean or proportion first with, say, the SD weights and then repeat with the D weights, and see if there is a difference. I believe you would find that the state estimates will be the same if you use either SD or D, but the national estimates will NOT be the same, and you do need to use the national weights for the national estimates.
|
|
|