NCJ Number
188126
Journal
Journal of Quantitative Criminology Volume: 17 Issue: 1 Dated: March 2001 Pages: 3-32
Date Published
2001
Length
30 pages
Annotation
The research reported in this article used the 1995 National Crime Victimization Survey (NCVS) and tract-level census data to examine how the risk of violence is distributed across persons and places in the United States and whether empirical findings are sensitive to the operationalization of violence.
Abstract
The NCVS data have several desirable properties that permit a more comprehensive set of descriptive analyses than are possible with other data sets that contain self-reports of victimization. Several key features include the sample's size and coverage, participation rates, and an extensive history of methodological analyses. The tract-level measures that were extracted from the decennial census and linked to the NCVS were chosen because they were shown to be important predictors of community crime rates. Although numerous indicators were examined for their relationship to risk of violence, the findings focus on three summary indices of community context based on the results of principal components analyses: socioeconomic disadvantage, immigrant concentration, and residential instability. Individual characteristics that were shown in previous research to be associated with risk for violence included age (measured in years), race (black vs. white), gender, and marital status (married versus not married). Those factors, along with measures of family income, number of evenings at home in a typical week, and number of years at the current residence were considered in this study. Two dependent variables were the focus of this research: any violent victimization in the previous 6 months, and violent victimization that occurred in the respondent's neighborhood. Violent victimization was defined as "an incident or attempted incident of robbery, rape, or assault." Events were classified as occurring in the neighborhood if they were reported to have happened either at or near the respondent's home or within one mile of their home. Findings showed that some individual-level predictors (e.g., gender and race) were sensitive to the operationalization of violence; whereas, others (e.g., age and marital status) were not. In addition, the impact of community characteristics on violence depended on central-city residence. In central cities, persons most at risk were in disadvantaged tracts, with lower proportions of immigrants. Outside central cities, the proportion of immigrants in an area increased risk, and community disadvantage had no independent influence. The importance of an empirical foundation for the development of theories of risk is discussed. 6 tables and 22 references