I have a problem that my feable datastep-skills simply can't seem to overcome. I have data that looks like: Iden change1 change2 change3 change4 change5 .... change44 1 a x x z 2 a x x x x x 3 b x x 4 c x x y 5 d z z z x x 6 a x x x . . 1786 What i would like to do, is to identify all the IDEN's that have "a" or "b" in change1, and then "z", "y" or "empty" three weeks later. That is in change4. In the above example, that would pull out IDEN 1,3,4 and 5. Then calculate the percentage of the ones who had either "a" or "b" in change1, and then "z", "y" or "empty" three weeks later, out of everyone who had "a" or "b" in change1. Exactly this, is what i would like to do in change1-change44. So the of the IDEN's in change2, who has "a" or "b", has "z", "y" or "empty" in change5. And that i would like to do all the way to change40. I hope that it's understandable.. /T
... View more