| Re: IPV and Non-partner violence [message #17747 is a reply to message #17710] |
Tue, 21 May 2019 05:49   |
Bridgette-DHS
Messages: 3035
Registered: February 2013
|
Senior Member |
|
|
Following is a response from Senior DHS Specialist, Kerry MacQuarrie:
Sampling for the Domestic Violence module is a little different than the sampling for the individual woman's questionnaire. First, whereas EVERY eligible woman per household is selected for the individual woman's questionnaire, only ONE eligible woman per household is selected at random for the Domestic Violence. We do this for ethical reasons to protect respondents. Secondly, the Domestic Violence module is often implemented in a sub-sample of selected households. In the Haiti 2016-17 DHS (EMMUS-VI), this was a two-thirds sub-sample, i.e. two out of every three selected households included the module.
Women are selected for the Domestic Violence module regardless of marital status. Both unmarried and married women respond. That is why you see a mixture of marital statuses in your tabulation of v044. However, many questions are relevant only for currently married or formerly married women because they ask about marital control and emotional, physical, or sexual violence perpetrated by their current/most recent spouse. Other questions are relevant for never-married women and married women alike. If you review the questionnaire in Annexe E of the Final Report, you can follow the skip pattern to see which questions are asked of which women.
Finally, the numbers of women that you see in the Final Report are weighted numbers. They are weighted by either v005 if it is a variable for all women or weighted by d005 if it is a variable in the domestic violence module. This sample weight accounts for the different sampling probability and non-response rate in the Domestic Violence module. (Note: when using either v005 or d005, divide by 1000000).
|
|
|
|
The DHS Program User Forum
Members
Search
Help
Register
Login
Home
