Generating a new variable that includes the content of another variable from one panel to all panels

[Patrick Abi Nader]

Hello All,

My data are in long format, and I have multiple (up to 16) measurements per participant and they can be at different organizations (there are 25 organizations) at each measurement. I am trying to create a variable that populates the ID of the first organization (first_org) they were at when they first joined the study. For example, in the data example below I would like ID 11000 to have 1 populated for every cycle under a new variable first_org.

I came across the following Stata FAQ, however, I am failing at conceiving how I can adapt any of the code provided in there to my situation. Does anyone have a suggestion or recommendation?

Thank you for your time and assistance.

* Example generated by -dataex-. To install: ssc install dataex
clear
input double(ID cycle organization)
11000  1  1
11000  2  1
11000  3  1
11000  4 20
11000  5 20
11000  7 20
11000  8 20
11000  9 20
11000 10 20
11000 11 20
11000 12 20
11001  1  1
11001  2  1
11001  3  1
11001  4 20
11001  5 20
11001  7 20
11002  1  1
11002  2  1
11002  3  1
11002  5 20
11002 10 20
11002 13 23
11002 14 23
11002 15 23
11003  1  1
11003  3  1
11003  4 20
11003  5 20
11003  7 20
11004  1  1
11004  2  1
11004  3  1
11004  4 20
11004  5 20
11004  7 20
11004  8 20
11004  9 20
11004 10 20
11004 11 20
11004 12 20
11004 13 23
11004 14 23
11004 15 23
11005  2  1
11005  3  1
11005  5 20
11005  7 20
11005  8 20
11005  9 20
11005 10 20
11005 11 20
11005 12 20
11005 13 23
11005 14 23
11005 15 23
11005 16 23
11006  1  1
11006  2  1
11006  3  1
11006  5 20
11006  7 20
11006  8 20
11006  9 20
11006 10 20
11006 11 20
11006 12 20
11006 13 23
11006 14 23
11006 15 23
11006 16 23
11007  1  1
11007  2  1
11007  3  1
11007  4 20
11007  5 20
11007  8 20
11007  9 20
11007 10 20
11007 11 20
11007 12 20
11007 13 23
11007 14 23
11007 15 23
11007 16 23
11008  1  1
11008  2  1
11008  3  1
11008  4 20
11008  5 20
11008  7 20
11008  8 20
11008  9 20
11008 10 20
11008 11 20
11008 12 20
11008 13 23
11008 14 23
11008 15 23
11009  1  1
end

[Romalpa Akzo]

If I understand exactly what you need, this code should have.

What make me could not be sure is the noticing that ID 11005 has no cycle 1, but starting with cycle 2.

bys ID (cycle): gen first_org = organization[1]

[Patrick Abi Nader]

Hello Romalpa,

Yes not all participants joined at cycle 1. As you noticed they could have joined at any cycle from cycle 1 to 13.

I believe that the code that you provided was exactly what I was looking for. I am just double checking the sample that I get.

Best wishes and thank you,
Patrick