In this video we return to the FEV Example and show how the OUTPUT statement, in conjunction with a Sum statement, can be used to compute and store posttest-pretest differences even when the posttest and pretest values are on different observations (rows) of a data set. This example pulls together several techniques--conditional execution, the Sum statement, a Do-END group, and the OUTPUT statement--in a small data step that does task that is simple conceptually but difficult due to SAS' built-in loop that processes one observation at a time.