Is it possible to test more than one outcome at a time? What revisions in the testing approach would need to be made to do this?