The DHS Program User Forum
Discussions regarding The DHS Program data and results
Home » Data » Weighting data » Couple's weights and domestic violence (Couple's weights while looking at domestic violence)
Couple's weights and domestic violence [message #25216] Sat, 17 September 2022 17:07 Go to next message
salome is currently offline  salome
Messages: 19
Registered: August 2021
Member
Good morning

I've read through lots of forum discussions as well as Dr. Stan Becker's paper and I'm still unclear about the use of weights in a couple's dataset. I have read that mv005 is the best way to go with when working with the couples dataset. How about if I am working with a couple's dataset and DV outcomes within the couple? If I use the DV weights, this would not account for male partners' non-response/lack of presence in the couple's dataset correct?

Thank you so much in advance. Any help would be much appreciated.

kindly
S
Re: Couple's weights and domestic violence [message #25228 is a reply to message #25216] Mon, 19 September 2022 09:20 Go to previous messageGo to next message
Bridgette-DHS is currently offline  Bridgette-DHS
Messages: 3016
Registered: February 2013
Senior Member

Following is a response from DHS staff member, Tom Pullum:

You are right --if you wanted to use the DV variables with couples (in the CR file) you would like to be able to take into account both women's nonresponse to the DV module and men's nonresponse. Another way of putting it is that you would like to adjust the couple weight (if it were included) for nonresponse to the DV module. Theoretically, it would be possible to do this, but I don't think anyone has tackled it.

I suggest that you do a sensitivity analysis, for whatever you would consider to be the most important estimates in your research. Produce them once with v005 as the weight variable, as a kind of baseline, and then with mv005 and d005. To be conservative, you could select whichever tends to give fewer statistically significant results. (Or you could have some other criteria.) But I don't think you will find that the results are really very sensitive to the choice of weights.
Re: Couple's weights and domestic violence [message #25229 is a reply to message #25228] Mon, 19 September 2022 09:32 Go to previous message
salome is currently offline  salome
Messages: 19
Registered: August 2021
Member
Dear Tom and Bridgette

Thank you so much for your response. What you suggest makes sense. So far, qualitatively the results when using no weights vs. DV weights are very similar. I will attempt what you propose as well.
I suppose one could also do a very crude sensitivity analysis by postulating selection (into the couples dataset) probabilities per Lash, Fox and Fink 2009.

S
Previous Topic: What is the difference between v005 and sweight DHS India?
Next Topic: Weighting for the Pakistan Special 2019 DHS on Stata
Goto Forum:
  


Current Time: Thu Mar 28 07:43:32 Coordinated Universal Time 2024