Output changes when code is rerun

[Mary Burckhette]

Dear Statalist Forum,

I have the following issue: When I run my whole code (beginning with clear, use, and data cleaning), I get the result that the variable newport_sa == 1 frequency is 0. Once I only rerun the specific line

replace newport_sa =1 if ((original_sa < changed_sa) & (changed_sa < endw_sa)) | ((original_sa > changed_sa) & (changed_sa > endw_sa)) //changed allocation is between original and endweight

and then rerun

 count if newport_sa == 1

, the result suddenly changes to 7, how it should be.

Could anyone please explain what is happening and why my results are inconsistent? How can I prevent these changes in output?

Thank you very much in advance!

[Mary Burckhette]

I figured it out; I had some other variables with fewer restrictions coded as 2 and 3. Therefore, the output for 1 changed.