I am trying to create a new variable adv_outcome. I have 5 pre-existing variables I am interested in sad, angry, grief, disappoint, hope. These categories are not mutually exclusive (i.e. some patients will be sad AND angry). These categories are all binary 0=no, 1=yes.
I tried using
gen adv_outcome= 0 if sad==0 & angry==0 & grief==0 & disappoint==0 & hope==0
replace adv_outcome=1 if sad==1
replace adv_outcome=2 if angry==1
replace adv_outcome=3 if grief==1
replace adv_outcome=4 if disappoint==1
replace adv_outcome=5 if hope==1
The issue when I do this is that as it creates each new category it "takes" the observations from the previous category instead of adding them to the tally for both. The reason I want this adv_outcome variable is so that I can compare it to various other factors such as age, gender, rurality and have it show me what % of men end up experiencing each adverse outcome rather than creating 6 separate tables for each patient-factor variable. I do have an any_outcome variable that was created if any of the 5 (sad, angry, grief, disappoint, hope.) occurred which seems to be working fine, but I'd like to see which outcomes were most common.
This is what I would like the crosstabs to look like: (%'s made up for example purpose, obviously just want the numbers not the words but wanted to convey what I want as clearly as possible so put the words in for this)
Metro | Regional | Rural | |
none | 12% experienced none | 24% experienced none | 5% experienced none |
sad | 36% experienced sad | 47% experienced sad | 60% experienced sad |
angry | 40% experienced angry | 10% experienced angry | 40% experienced angry |
grief | 65% experienced grief | 60% experienced grief | 30% experienced grief |
dissapoint | 35% experienced disappointment | 5% experienced disappointment | 90% experienced disappointment |
hope | 80% experienced hope | 70% experienced hope | 67% experienced hope |
Is it easier to just run each table separately and manually type the numbers into a normal table? Am I missing something super obvious? I feel like I've been staring at it all day making zilch progress.
Thanks, J
0 Response to Create a new variable from existing variables where the categories are not mutually exclusive
Post a Comment