Home » Topics » Domestic Violence » Determining 'Ever-Married' sample
| Re: Determining 'Ever-Married' sample [message #28266 is a reply to message #28182] |
Wed, 06 December 2023 16:54  |
Janet-DHS
Messages: 888
Registered: April 2022
|
Senior Member |
|
|
Following is a response from DHS staff member, Tom Pullum:
For the most part, the 2014 and 2022 Cambodia surveys had very similar designs with respect to selection for the DV module. In the household file, both of them include a variable hv044 (v044 in the IR file) that is 1 if the household was selected for the DV module and 0 if it was not. (Another variable, hv027 or v027, indicates whether the household was selected for the male interview; exactly the same households were selected or deselected for both.) In the 2022 survey, it was a 50% subsample, consisting of households with an odd household number (hv002=1, 3, 5, etc.) In the 2014 survey, it was a 1/3 subsample (of households with hv002=1, 4, 7, etc.). Then, in both surveys, one eligible woman (if there was more than one) was selected at random. As is usual, eligibility for the women's survey was based on de facto residence (hv103=1) and age 15-49 and is given in the PR file with hv117=1.
The selection on marital status for the DV module in the 2014 survey that you are talking about was apparently made during the women's interview. It was not done as part of the household survey and was not based on the report of marital status in the household survey, but on v502 in the IR file. To identify which women in the IR file actually were given the DV module, you can select any of the DV variables and find which cases are not NA. I arbitrarily picked d104 for this purpose. Then "tab d104 v502,m" in the IR file for the 2022 survey gives this:
experience |
d any |
emotional |
violence | currently/formerly/never in
(d103x | union
series) | never in currently formerly | Total
-----------+---------------------------------+----------
no | 133 4,560 264 | 4,957
yes | 9 1,080 158 | 1,247
. | 4,474 8,115 703 | 13,292
-----------+---------------------------------+----------
Total | 4,616 13,755 1,125 | 19,496
In the 2014 survey you get this:
experience |
d any |
emotional |
violence | currently/formerly/never in
(d103x | union
series)*** | never in currently formerly | Total
-----------+---------------------------------+----------
no | 0 2,530 178 | 2,708
yes | 0 693 97 | 790
. | 4,651 8,445 984 | 14,080
-----------+---------------------------------+----------
Total | 4,651 11,668 1,259 | 17,578
Therefore, I infer that you will get a subsample in the 2022 survey that is comparable to the 2014 survey if you drop the cases with v502=0 in the 2022 survey. As you can see, it only amounts to dropping a few (142 unweighted) cases. I'll add that I believe the DV module is usually restricted to women who have been in a union and I don't quite understand how the reference to a spouse or partner could even apply to women who have never been in a union. However, the current version of the module includes potential abusers beyond the spouse or partner, and for them a restriction based on v502 would not be appropriate.
|
|
|
|
Current Time: Tue Nov 26 02:54:35 Coordinated Universal Time 2024
|
The DHS Program User Forum
Members
Search
Help
Register
Login
Home
