A test case identification technique which produces enough test cases such that each decision has a true and a false outcome at least once.