|Uganda 2011 domestic violence [message #13116]
||Fri, 22 September 2017 09:16
Registered: July 2017
Dear editor, |
I am using the 2011 Uganda data set to study the association between women´s exposure to IPV and HIV testing. I have several questions:
1. I am using the svy command in Stata to run the analysis taking int account the survey sampling design; however, key variables about the strata are empty in the survey (V022 V023). These variables are often used to run Svy so what can be done when these are empty.
. tab V023
2. The variables describing IPV exposures for women are confusing (i.e. D105A-J). For example, when you tab variable D105A (Ever been pushed, shook or had something thrown by husband/partner) you get five options: never, often, sometimes, Yes but not in the last 12 months, Yes but frequency in last 12 months.
How do you interpret this if you ant to create a variable that measures ever exposure to this behavior (yes, no). No must be never, but what about yes. Ever exposure (Yes), should include often, sometimes, Yes but not in the last 12 months or Yes but frequency in last 12 months?
This is an example of my stata dofile
generate EverPhy = .
replace EverPhy =1 if (D105A==1)
replace EverPhy =1 if (D105A==2)
replace EverPhy =1 if (D105A==3)
replace EverPhy =1 if (D105A==4)
replace EverPhy =0 if (EverPhy ==.)
This is more confusing as the questionnaire the questions are structured in a different way. First is a yes, no question for ever exposure. And then the 12 month exposure to IPV is evaluate uisng the following: often, sometimes, Yes but not in the last 12 months.
when i run my calculations my prevalence does not match the prevalence described in the report.
Can we get the dofile taht describes how the IPV exposure were created for the 2011 DHS report?
3. We are using variable V826A (months ago from recent HIV test) to create a binary variable describing HIV testing in the last twelve months. This is a continuous variable that goes from 0 months to two years or more. In this case what does 0 mean? The the test was done weeks or days before the interview but not up to a month? 0 months must be excluded from the calculation of this variable?