Create a new variable based on observations within a group

[Marcel Weil]

Hello,

my dataset is structered the following:

clear
input float (group_id won_close_dummy loss_close_dummy)
1 0 1
2 1 0
2 1 0
3 1 0
6 1 0
7 0 1
9 0 0
13 0 1
13 1 0
13 0 1
13 0 1
17 1 0
17 0 1
18 0 1
20 1 0
21 1 0
24 1 0
24 1 0
25 0 1
28 0 1
33 1 0
34 1 0
35 1 0
38 0 0
42 1 0
43 1 0
43 1 0
43 1 0
43 1 0
43 1 0
45 1 0
45 0 1
46 0 1
46 1 0
47 1 0
47 0 1

I now want to create a new binary variable "donated" which takes the value of 0 if within a group there is at least one 1 for both "won_close_dummy" and "loss_close_dummy". This means if within a group (eg. group_id ==13 or group_id==17) both variables "won_close_dummy" and "loss_close_dummy" have at least one 1 the new variable "donated" should get a 0 for the whole group.
If within a group (eg. group_id==2 or group_id==43) only one variable has a 1 the new variable "donated" should get a 1 for the hole group.
This means it should look like this in the end:

input float(group_id won_close_dummy loss_close_dummy donated)
1 0 1 1
2 1 0 1
2 1 0 1
3 1 0 1
6 1 0 1
7 0 1 1
9 0 0 1
13 0 1 0
13 1 0 0
13 0 1 0
13 0 1 0
17 1 0 0
17 0 1 0
18 0 1 1
20 1 0 1
21 1 0 1
24 1 0 1
24 1 0 1
25 0 1 1
28 0 1 1
33 1 0 1
34 1 0 1
35 1 0 1
38 0 0 1
42 1 0 1
43 1 0 1
43 1 0 1
43 1 0 1
43 1 0 1
43 1 0 1
45 1 0 0
45 0 1 0
46 0 1 0
46 1 0 0
47 1 0 0
47 0 1 0

I hope you can help me solving this problem. Thanks in advance!

Greetings
Marcel

[Nick Cox]

Thanks for your data example. This should help.

Each indicator (dummy in your terminology, which I don't favour) being 1 at least once corresponds to a total being at least 1, and so non-zero. So you can combine criteria to get what you want by negating the logical combination (criterion1) & (criterion2) evaluated as 1 (true) if and only if both arguments are true (non-zero).

clear 
input float(group_id won_close_dummy loss_close_dummy donated)
1 0 1 1
2 1 0 1
2 1 0 1
3 1 0 1
6 1 0 1
7 0 1 1
9 0 0 1
13 0 1 0
13 1 0 0
13 0 1 0
13 0 1 0
17 1 0 0
17 0 1 0
18 0 1 1
20 1 0 1
21 1 0 1
24 1 0 1
24 1 0 1
25 0 1 1
28 0 1 1
33 1 0 1
34 1 0 1
35 1 0 1
38 0 0 1
42 1 0 1
43 1 0 1
43 1 0 1
43 1 0 1
43 1 0 1
43 1 0 1
45 1 0 0
45 0 1 0
46 0 1 0
46 1 0 0
47 1 0 0
47 0 1 0
end 

egen criterion1 = total( won_close_dummy), by(group_id) 
egen criterion2 = total(loss_close_dummy), by(group_id) 

gen wanted = !(criterion1 & criterion2) 
assert wanted == donated

For more on true and false see https://www.stata.com/support/faqs/d...rue-and-false/

For why indicator is a better term see https://www.stata-journal.com/articl...article=dm0099