| Re: svyset [message #22586 is a reply to message #22554] |
Wed, 07 April 2021 14:03  |
Bridgette-DHS
Messages: 3199
Registered: February 2013
|
Senior Member |
|
|
Following is a response from Senior DHS Specialist, Kerry MacQuarrie:
Yes, using svyset is appropriate for analyzing intimate partner violence data, and you identify the correct weight variable as well. You may want to divide the weight variable (d005) by 1,000,000 as DHS weights have 6 implicit decimal places.
It is always wise to check the questionnaires that are published in the appendix of the survey's Final Reports. As far as the Colombia surveys go, they often implement a modified domestic violence module rather than the DHS standard domestic violence module. I checked the 2015 Colombia DHS Final Report and it indicates that for most forms of spousal intimate partner violence (e.g. physical), the survey did not ask frequency in the last 12 months. It does ask about whether women experienced each violence item ever, and in the last 12 months (yes/no), as well as whether this occurred in public or private. Other idiosyncrasies are questions about economic violence, and an expanded set of emotional/psychological violence items, which are absent from the DHS standard module.
Because these are non-standard variables, they will not have the same variable name. The variables reserved for frequency of IPV in the last 12 months will be empty. Data on whether the violence occurred in the last 12 months are contained in the d103 (emotional) and d105 (physical and sexual) variable series, as they are in surveys that use the DHS standard module. Survey-specific variables are indicated with a s. For example, s1204 variables refer to place of physical violence items.
[Updated on: Wed, 07 April 2021 14:05] Report message to a moderator |
|
|
|
The DHS Program User Forum
Members
Search
Help
Register
Login
Home
