The two types of comparisons made between the data collection systems are intrasystem and intersystem. Intrasystem comparisons focus on similar variables collected at different times by various organizational levels within the same reporting system. Intersystem comparisons, on the other hand, examine the extent of agreement between two different homicide reporting systems. The assessment of the systems concludes that both the UCR and the NCHS provide, consistent with their definitions, a reasonably accurate count of the annual number of homicides for approximately the past 25 years. When questions of variable completeness are considered for SHR's, there is a substantial amount of disagreement between national and city police sources. Whether a similar pattern of disagreement exists for national and local NCHS sources is unknown due to a lack of research. The single greatest problem for both reporting systems has been an unwillingness to focus research on how the homicide is initially reported at the level of reporting officer or medical examiner.