NCJ Number
86700
Journal
Annual Review of Sociology Volume: 7 Dated: (1981) Pages: 107-128
Date Published
1981
Length
22 pages
Annotation
Because the social-indicator interest in victimization survey data is likely to take precedence over interests in the etiology and dynamics of victimization, the National Crime Survey might be improved by using two or more separate survey designs.
Abstract
To sociologists, the National Crime Survey (NCS), a United States victimization survey, has two broad appeals. One is as a social indicator to be used primarily on the national or regional level to track serious crime in the Nation independently of the criminal justice system. For this purpose, only a minimal amount of information about the respondents and their victimizations is required, e.g., information sufficient to categorize the crimes and perhaps to ascertain whether or not the victims claim to have contacted the police. The second broad appeal is as a reservoir of details about respondents, the dynamics of their victimizations, the offenders who committed the crimes, and the long-term physical and mental consequences of these events for the victims. A single survey may not be the best vehicle to achieve these disparate goals. For example, as Fienberg (1977) implies, 'time-in-sample' biases would probably be smaller if the detailed followup questions were fewer. Owing to the mobility of the American population, time-in-sample bias may be a serious threat to the validity of the social-indicator uses of the NCS. Although such biases also have consequences for sociologists more interested in the patterns of victimization, they are not as detrimental. Thus, the several uses of the NCS data may best be served by using two or more separate designs.