Question on analyzing multiple answers question

[Njoki Karinja]

I have a data set of 987 respondents. I asked them on there most common sources of treatment and they could choose more than one source out of a possible 10 choices. I have created the following loop and it works perfectly as i can now report frequency of each type of source of care. I would like to collapse/combine district hospital, sub district hospital, health center and dispensary into a category called "government facility" and private hospital, private clinic and mission hospital into a category called "private facility" etc. How do i go about this bearing in mind this was a multiple answers question?

Thanks in advance.

forvalues i=1/10 {
generate treatment`i' = regexm(trtmntsrc,"`i'")
replace treatment`i'=. if Trttype!=1
tab treatment`i',miss
recode treatment`i' (1=1 "Yes") (0=0 "No"), gen(treatment_new`i')
tab treatment_new`i'
}

rename treatment_new1 district_hospital
rename treatment_new2 sub_district_hospital
rename treatment_new3 health_center
rename treatment_new4 dispensary
rename treatment_new5 community_health
rename treatment_new6 mission_hospital
rename treatment_new7 private_hospital
rename treatment_new8 private_clinic
rename treatment_new9 pharmacy
rename treatment_new10 shop

[Nick Cox]

Looks buggy as the string "10" is a match for pattern "1".

. display regexm("10", "1")
1

[Njoki Karinja]

Looks buggy as the string "10" is a match for pattern "1".

. display regexm("10", "1")
1

Oh my, you are right. How else can i analyse multiple answers in this case?

[Andrew Musau]

How is your data structured? You can, for example, present the results of

ssc install dataex
dataex in 1/10