Linear and skip-branching programmes: a comparison study.