The group must be large enough to yield stable rates. However, two potential disadvantages were addressed. First, the comparison work cohort might not be large enough to produce stable mortality or morbidity rates. Second, presence of differences in other extraneous variables such as demographic characteristics.