Multiple forms

[Olga Oguaah]

Hi everyone,
Am doing a study and my outcome variable is Multiple forms of undernutrition. I have calculated my z scores and gotten variables for stunting, wasting, underweight, and anemia. I need to generate a variable ( Coexisted) that will mean a child that experiences at least two or more of the forms of undernutrition concurrently so I can continue with my analysis.
I tried this syntax but it didn't work
//generate Coexisted=1. if stunting=1|wasting=1|underweight=1|anemia=1
replace Coexisted=0 if stunting=0|wasting=0|underweight=0|anemia=0
replace Coexisted=. if stunting==.|wasting==.|underweight==.|anemia=.
label define 1 “Coexisted “ 0 “Not Coexisted”, replace
label values Coexisted Coexisted //

[Nick Cox]

The code does three main things.

generate Coexisted=1. if stunting=1|wasting=1|underweight=1|anemia=1

replace Coexisted=0 if stunting=0|wasting=0|underweight=0|anemia=0

replace Coexisted=. if stunting==.|wasting==.|underweight==.|anemia=.

The new variable is 1 if any of four variables is 1, but then overwritten by 0 if any is 0, but then overwritten by missing if any is missing.

Again: Watch the way the voting works. A "vote" of missing anywhere overwrites a vote of 0 anywhere and one of 0 overwrites a vote of 1 anywhere.

Nothing in the code counts instances of 1. You can do that directly.

egen wanted = rowtotal(stunting wasting underweight anemia)

adds rowwise, ignoring missings. Then

gen Coexisted = wanted >= 2

identifies observations with counts of 2, 3 or 4.

See also https://www.stata.com/support/faqs/d...ummy-variables

https://www.stata.com/support/faqs/d...rue-and-false/

[Øyvind Snilsberg]

* Example generated by -dataex-. To install: ssc install dataex
clear
input float(id stunting wasting underweight anemia)
1 1 0 1 0
2 0 0 0 1
3 . 0 1 .
end

egen forms = rowtotal(stunting wasting underweight anemia)
gen wanted = forms>1

[Olga Oguaah]

Thank you works perfectly now

[Nick Cox]

This is for any future readers. The code in #1 has expressions like this

stunting=1|wasting=1|underweight=1|anemia=1

where each equals sign should be doubled

stunting==1|wasting==1|underweight==1|anemia==1

My bad for not noticing this in #2. but that oversight doesn't undermine the suggestion there.