Creating new variables based on information from three other variables

[William Lobao]

I’m having some difficulty creating a new variable from three others.
All variables are of binary type (values 0 or 1).

They are a result of an analysis of a set of variables (Like an indicator). For instance:
gen alerta_rec = a1maoseoupesfrios+ a3noudornoestmago+ a2bocaseca+ a4aumentodesudorese+ a5tensomuscular+ a6apertonamandbula+ a7diarreiapassageira+ a8inosnia+ a9taquicardia+ a10respiracaoofegante+ a11hipertensaosubita+ a12mudancadeapetite+ a13aumentosubitodemotivacao+ a14entusiasmosubito+ a15vontadesubita
gen alerta_id = alerta_rec
recode alerta_id 0/6=0 7/max=1
When I create a table with them, they show this:

alerta_id	n	%
0	101	89.4
1	12	10.6
	113

resistencia_id	n	%
0	70	61.9
1	43	38.1
	113

exaustao_id	n	%
0	102	90.3
1	11	9.7
	113

I’d like to create a variable which shows the total of individuals that agree with at least one of the three originals items (which would be 66), but it doesn't work!

gen lipp_status=0
replace lipp_status=1 if alerta_id==1 | resistencia_id==1 | exaustao_id==1
tab lipp_status

lipp_status	n	%
0	69	61.1
1	44	38.9
	113

How can I solve this?
Thank you

[William Lisowski]

Welcome to Statalist.

Your problem is that you have individuals who agree with more than one of the three original items, and for them, they will be counted only once in lipp_status.

list alerta_id esistencia_id exaustao_id if lipp_status==1