Gen newvar = 1 if var1==1 and all subsequent observations (long data)

[Nate Fitzpatrick]

Hello, I am trying to determine how to generate a new variable when the criteria is met, and all following observations for that study id in a long formatted dataset. Essentially censoring the data beyond the set time point, but as a 0/1 dummy variable.

I have tried codes along the lines of:

bysort id: gen newvar = 1 if var1==1 & var1[_n+1]==0

or just

bysort id: gen newvar = 1 if var1==1 & var1[_n+1]

However, I cannot figure out how to get stata to identify subsequent observations for that ID.

Thank you so much.



EDIT:

I THINK I was able to find the answer on another forum. A user suggested:

bysort id (visit): gen newvar = sum(var1[_n-1]) > 0  

replace newvar = 1 if var1==1

Which seems to have worked, but I don't really understand why. If anyone could explain this, or have a different suggestion, I'd greatly appreciate it.

Thanks again!

[Nick Cox]

Your set-up appears to be that you have an indicator variable visit in panel data and your single criterion appears to be that you want a marker that switches from 0 to 1 when you pass the first value of 1.

bysort id (visit) : gen wanted = sum(visit) > 0

would do that as a the cumulative sum of a series of 0s remains zero but the cumulative sum jumps up at the first occurrence of 1.

When cross-referencing please give the URL of your source.

See also https://www.stata.com/support/faqs/d...t-occurrences/ for more explanation.