Discrepancies Between Crime Reports and Crime Surveys:

Abstract The absence of strong zero‐order associations between victimization and official crime rates for cities has been a puzzle for social scientists since the data for making such comparisons became available. Using the 26 large central cities for which data on both types of rates are available, we analyze the extent to which discrepancies between the rates can be accounted for by aspects of urban social structure that differ from city to city. After introducing such structural controls, we find a much closer correspondence between the two types of rates for motor vehicle theft, robbery, burglary, and forcible rape, but not for aggravated assault and larceny‐theft. These results are explained by citing evidence that we have identified some critical “suppressor” variables for the former crimes (i.e., variables that are positively associated with one type of rate and negatively associated with the other). By contrast, the heterogeneous nature of the phenomena subsumed by the latter two crime categories may preclude identification of a similarly parsimonious list of suppressors. One implication of these conclusions is that cross‐sectional analyses of intercity variation in official rates may produce results that are in reasonably close correspondence with what would be obtained with victimization rates for certain index crimes, provided that sufficient structural control variables are utilized.