Liz C.S.--I have provided Malayaka some example code that I wrote using the Ethiopia 2011 DHS Final report table 4.6 (I didn't have access to the Malawi data) via private message, which I reproduce below.

---


Malayaka, I was able to replicate the 2011 Ethiopia DHS Final Report using the code below after calling in the dataset:

gen wt=v005/1000000

svyset [pw=wt] /*you can specify cluster and strata if you want for central tendency, but cluster and strata affect standard errors, not proportions*/

*create indicator variables of "married by" for a specific age (15, 18, 20, 22, 25)
gen mb15=(v511<15)
label var mb15 "married by 15"
gen mb18=(v511<18)
gen mb20=(v511<20)
gen mb22=(v511<22)
gen mb25=(v511<25)

svy: tab v013 mb15,row count
svy: tab v013 mb18,row count
svy: tab v013 mb20,row count
svy: tab v013 mb22,row count
svy: tab v013 mb25,row count

Table from DHS report (png) and my output (txt) are attached. As you can see for the row proportions where "mb15==1," my proportions are off only by some small decimal points from what was reported in the report. You can use this same formula to do your fertility and domestic and sexual violence variables. Unfortunately, I have never done domestic and sexual abuse analysis, so I can't tell you which variables you should use, but the code given here will give you a headstart.

HTH,
RHS