| Re: merging the DV module with men's recode [message #25603 is a reply to message #25568] |
Wed, 16 November 2022 09:11   |
Janet-DHS
Messages: 888
Registered: April 2022
|
Senior Member |
|
|
Following is a response from DHS staff member Tom Pullum:
I should have looked at the DV variables in the CR file. I do see, now, that they are empty. I just went to the PR file and entered "lookfor select" because that will often identify the variable(s) used for subsampling. Sure enough, hv027 and hv044 indicate selection for the DV module and for the male interview, respectively. If you do "tab hv027 hv044" you will see that the variables are mutually exclusive.
This kind of subsampling is done at the household level. Looking at the HR file, a crosstab of hv027 and hv044 gives this:
. tab hv027 hv044
| household selected
household | for domestic violence
selected for | module
male interview | household household | Total
-----------------+----------------------+----------
not selected | 0 7,093 | 7,093
men's survey | 7,063 0 | 7,063
-----------------+----------------------+----------
Total | 7,063 7,093 | 14,156
Half the households were selected for the DV module and half for the men's survey, with no overlap. This should be described in the final report, but it could be buried in the sampling appendix. I agree with you that this is an unfortunate design. We on the analysis team have often recommended against this type of subsampling.
|
|
|
|
The DHS Program User Forum
Members
Search
Help
Register
Login
Home
