Although victim surveys can generate information about characteristics of victims, precursors of victimization, effectiveness of resistance, victim‐offender interaction, effects of crime on the victim, and police response, current research has hardly begun to mine their potential. Much of the underutilization and misinterpretation of victim surveys can be understood in light of their original objective: to measure the “dark figure” of crime. Victimization estimates were originally intended to index all occurrences of crime, both crimes that become known to the police and those that do not. Because victim‐survey data and police data are products of numerous definitional decisions made by victims, other citizens, and the police, some types of crime are likely to be captured by both data sources, some by only one, and some by neither. Thus, victim surveys do not simply reveal more crimes than police statistics; they capturedifferentcrimes. This very difference makes victim‐survey data especially