Home » Countries » Kenya » KDHS 2022 Gender Based Violence Indicators
| Re: KDHS 2022 Gender Based Violence Indicators [message #28093 is a reply to message #28052] |
Fri, 10 November 2023 14:12   |
Bridgette-DHS
Messages: 3172
Registered: February 2013
|
Senior Member |
|
|
Following is a response from Senior DHS staff member, Jeff Edmeades
Question 1:
I am trying to generate the GBV indicators (most importantly Table 17.9.1, Table 17.9.2, Table 17.11.1, Table 17.11.2 and Table 17.12.1) using the attached do files from DHS GitHub repository and datasets for KDHS 2022. There are couples of areas I will appreciate guidance:
It seems to me the indicator Prevalence of intimate partner violence (Experience of any of the specified acts of physical, sexual, psychological/emotional, or economic violence committed by ones current or most recent husband/wife/intimate partner) may not have been captured in any of the do files. I would have expected this to be in the DV_PRTNR.do but this seems not to be the case.
Response: The questions on economic violence are specific to the KDHS and are not asked in the standard set of questions typically administered by the DHS. As a result, the do-file on github only includes the code for the combination of physical, sexual and emotional violence. In order to construct the combined measure that includes economic violence, as in Table 17.9.1, the code for the other forms of violence should be replicated for economic violence. I am attaching a do-file that creates the basic variables needed that you can adapt as you need. The variables in question for economic violence are d103d ('restricted, exploited, or sabotage ability to acquire/access economic resources') and sdv16d ('any previous husband/partner restrict, exploit, or sabotage ability to acquire/access economic resources'). Please note that any survey-specific variables have the 's' prefix.
Another important thing to note is related to a change that was made to the questionnaire between DHS7 and DHS8 that is not yet reflected in the GitHub code. In DHS7, the domestic violence questions were asked only for those who had ever been married (v501>0). In DHS8, this was expanded to include those who had never been married but who did have an intimate partner at some point (i.e. someone with whom they were sexually active but never married nor lived with as if married). This is captured by the d100 variable. In order to include these individuals and replicate table 17.9.1 in particular, the code for the construction of the variables needs to be adapted from (using physical violence as an example):
• OLD: gen dv_prtnr_phy = 0 if v044==1 & v502>0
• NEW: gen dv_prtnr_phy = 0 if v044==1 & (v502>0 | inlist(d100,1,2))
I have modified the code in the attached do-file to match this for most of the variables and any additional code can be adapted from these examples. When applying the corrections for survey design (in the do-file these are shown using the svy command in the do-file), I was able to replicate Table 17.9.1.
Question 2:
2. Similarly, I have used the do files and am able to reproduce most of the report estimates for women using KDHS 2022 IR dataset. However, I am facing challenges when trying to replicate the same for men. For instance, variable d105 seems not to exist in the MR dataset. Isn't the do file applicable to the MR file or variables are named differently?
Response: The variables in the MR file are named differently to help distinguish between women and men's responses should the files be merged. All the variables in the MR file have an 'm' prefix (e.g. mv005 for men, v005 for women), though for the most part correspond directly to the variables in the IR dataset (i.e. they have the same name and label, with the exception of the leading 'm' in the variable name. The DHS Guide to Statistics https://dhsprogram.com/pubs/pdf/DHSG1/Guide_to_DHS_Statistic s_DHS-7_v2.pdf
provides more information (note that this is for DHS7 as the updated version for DHS8 hasn't been released yet). Because the men's domestic violence module is rarely used, the variables are coded in the way that survey-specific variables are, with the 's' prefix. So the equivalent of the d105 variable in the IR dataset is sdm105a-k and the weight that should be used to weight the analyses of the data from the men's domestic violence module is sdm005.
[Updated on: Fri, 10 November 2023 14:12] Report message to a moderator |
|
|
|
Current Time: Mon Oct 14 08:49:05 Coordinated Universal Time 2024
|
The DHS Program User Forum
Members
Search
Help
Register
Login
Home
