Sources of Data
Data on serious and violent juvenile offending are primarily of two types: official data from records generated by criminal and juvenile justice agencies and self-report data produced by delinquency and victimization surveys conducted independently of these criminal and juvenile justice agencies. Each source has strengths and weaknesses in the study of racial and ethnic differences in rates of serious offending.
Official Data Traditionally, arrest data have been used to study differences in rates of offending. The primary weakness of arrest data is that the data are collected only for those criminal and delinquent events that come to the attention of the police and result in an arrest. If ethnic and racial groups differ in their inclination to report crime to the authorities, or if crimes committed by certain groups are more likely to result in an arrest, these factors can bias estimates of racial differences in offending rates. Police themselves may be biased in their arrest policies and may handle offenders differently (e.g., arresting rather than warning) depending on the offender's racial or ethnic background (Hagan and Peterson, 1995; Mann, 1993). Other limitations include the fact that official crime statistics are incident based rather than person based. In other words, these data do not provide information about the chronicity of individual offending or the length of time the offender has been involved in crime. Also, while race-specific arrest rates can be calculated from official crime statistics, reliable comparisons of differences within racial groups (e.g., Caribbean blacks versus native U.S. blacks) cannot be drawn because such information is not recorded consistently in police data. Nevertheless, arrest statistics are often used to measure the level of juvenile involvement in crime.| Race and Ethnicity The data discussed in this Bulletin rely on race and ethnicity categories defined by the U.S. Department of Commerce, Bureau of the Census. Race is defined as one of five categories—white; black; American Indian, Eskimo, or Aleut; Asian or Pacific Islander; or other. The term "black" includes African Americans and people of this racial designation whose ancestral origin lies outside the United States (e.g., Haitians). Ethnicity usually indicates a person's country or countries of origin. In most data collections, however, this is limited to the designation Hispanic, which is based on people's identification of themselves as persons of Spanish-speaking origin, although they may be members of any one of the above race groups. This Bulletin focuses on racial rather than ethnic comparisons, because the data contrasting Hispanics and non-Hispanics are limited and lack the consistency needed for comparisons. Similar problems affect the collection of crime data for other racial groups such as Asians and Native Americans. |
Despite these limitations, several researchers have shown that data from the Federal Bureau of Investigation's (FBI's) Uniform Crime Reports (UCR) can be used reliably to assess differences in rates of serious offending (Hindelang, 1978; Gove, Hughes, and Geerken, 1985; Sampson and Lauritsen, 1997). In crimes such as robbery and assault, the distribution of arrestees by race has been compared with victims' reports of the perceived race of the offender in order to estimate the extent to which arrest data might misrepresent involvement in offending. These comparisons suggest that much of the race difference in arrests for violence is due to greater involvement in offending on the part of blacks. In cases of homicide, where victim reports are impossible, arrest data also are believed to accurately reflect race differences in offending because arrests are typically made on the basis of physical evidence and witness reports. It is important to note that most violent crime is intraracial and that blacks are disproportionately the victims of homicide and other forms of violence (Sampson and Lauritsen, 1997). But because it is impossible to compare victim reports to arrest data for the majority of crimes in which the victim does not see the offender, conclusions about race differences for other types of offending are difficult to make. Thus, the overall evidence suggests that UCR data in general should be used cautiously and, when possible, should be supplemented with data from other sources. Self-Report Offending Data Self-report surveys provide a useful alternative source of data because they eliminate some of the weaknesses inherent in official records. These surveys collect data directly from juveniles who report to researchers about their own conduct, even conduct not detected by authorities. The data provide another view of the overall distribution of crime and delinquency. The disadvantage of self-report surveys in the study of serious and violent offending is that the sample size is typically too small to generate enough information about serious and violent offenders to assess racial and ethnic differences reliably (Cernkovich, Giordano, and Pugh, 1985). Another disadvantage is that there may be a difference in the validity of self-reports for blacks and whites (Hindelang, 1981; Huizinga and Elliott, 1986). Early studies found that blacks were more likely to underreport serious misconduct, but a more recent study found no racial differences in predictive validity based on these self-reports (Farrington et al., 1996). In addition, with few exceptions, these early studies were based on local rather than national population samples. Self-Report Victimization Data The National Crime Victimization Survey (NCVS), another alternative to UCR data, is an ongoing survey conducted by the Bureau of Justice Statistics that measures the extent of personal and household victimization in the United States. The NCVS provides data about the characteristics of offenders as perceived by victims, regardless of whether the victim reported the crime to the police. Like other data sources, NCVS data have limitations in making racial and ethnic comparisons. As a survey of victims, no data on homicide are provided, no data are collected on the victimization experiences of persons under the age of 12, and victims can only identify an offender's race or ethnicity in crimes involving personal contact. Also, victims of crimes committed by family members or nonstrangers underreport victimization to interviewers. Other limitations include the fact that victims' reports of offender attributes can be difficult to corroborate and that reports of crimes involving multiple offenders do not identify the characteristics of each offender (Laub, 1987). In summary, the exclusive use of any one source of data can produce an underestimate of the volume of violent crime. None of these sources alone provides sufficient information about the characteristics of offenders and victims and the nature of the violence committed. Data on the social contexts of violence are especially lacking in the UCR and NCVS (Loftin and Mercy, 1995). However, these sources of data can be used in tandem to develop reliable estimates of racial and ethnic differences in serious juvenile offending.
| Previous | Contents | Next |
 |
|
| Race, Ethnicity, and Serious and Violent Juvenile Offending | Juvenile Justice Bulletin • June 2000 |