The two systems differ in several respects: the varieties of victims from whom reports come, the way that offenses are identified, the way information is stored and retrieved. Comparing official crime statistics with survey crime statistics in the United States and Great Britain shows that the current data collection procedures do not permit pooled estimates of the amount of crime. Instead, the two sources provide crude and somewhat misleading methods for making such estimates. Police statistics permit analyses that are not possible based on surveys of victims and offenders, however. Although police statistics and victim surveys should be designed to complement one another, victim surveys should be used less for estimating crime rates than for conducting other types of studies. Notes and 27 references.