Appendix

The rural violence research was conducted by the Institute for Families in Society at the University of South Carolina. The five components of this research include homicides committed by juveniles, patterns of gun ownership among nonmetropolitan middle school students, community factors affecting violence among rural youth, bullying and antisocial behavior among middle school students, and bullying prevention. The goal of this group of studies was to help fill a gap in the body of knowledge pertaining to the prevalence and nature of violence among youth in rural and nonmetropolitan communities and community-level predictors of youth violence.

The SC homicide study involved the examination of minors who committed homicides in a 3-year period in South Carolina. Computerized case record information was obtained from the State Department of Juvenile Justice for 98 youth referred to the State solicitor for homicide between 1992 and 1994. For the purposes of these analyses, only the data for the male youth (n=86, 88% of total sample) were reported. Case record information included limited demographic information about the youth and his family and a complete listing of referrals to the State solicitor (including dates for each offense, solicitor decisions, and dispositions). In order to obtain additional information pertaining to the circumstances surrounding the homicides, newspaper accounts were retrieved wherever possible (n=34).

For the purpose of comparison, computerized case record information was also obtained for two additional groups of youth who had been referred for serious and/or violent offenses: (1) 77 male youth who had been referred for assault and battery with intent to kill (assault and battery group) and (2) 87 male youth who had committed other serious offenses (other serious offense group), exclusive of homicide or assault and battery with intent to kill. Case records for the assault and battery group and the other serious offense group were randomly selected from the total sample of youth referred for these offenses between 1992 and 1994.

The SC gun study examined gun ownership in rural communities. Studies of youth in urban settings indicate youth who own guns for recreational purposes are less likely to engage in criminal behavior and less likely to carry guns regularly than are youth who own guns for protection or for engaging in dangerous or illegal activities. South Carolina researchers surveyed 6,263 students in 36 middle schools in nonmetropolitan counties to obtain information on gun ownership, reasons for gun ownership, and the relationship between patterns of gun ownership, antisocial behavior, and bullying.

In an effort to extend the study of community social disorganization and crime beyond its exclusive focus on large urban centers, the SC community study involved an analysis of structural correlates of arrest rates for juvenile violence in 264 nonmetropolitan counties in 4 States (Florida, Georgia, Nebraska, and South Carolina), where total populations ranged from 560 to 98,000. Delinquency was measured using the number of arrests for juveniles (ages 11 through 17) in each county, pooled over a 5-year period from 1989 to 1993. The primary dependent variables were arrests for homicide, forcible rape, aggravated assault, robbery, weapons offenses, and simple assault; arrests for crimes compiled for the Uniform Crime Reports violence index; and arrests for burglary, larceny, and motor vehicle theft. Explanatory variables, which were based primarily on 1990 census data, included mobility, unemployment rates, family disruption, ethnic heterogeneity, poverty, proximity to metropolitan counties, and proximity to an interstate highway.

The SC bullying study examined bullying and antisocial behavior among middle school students. The South Carolina researchers surveyed all fourth, fifth, and sixth grade students from six nonmetropolitan school districts in South Carolina. The survey instrument was an English language version of the Olweus Questionnaire for Students, which was revised for use with middle school students. This questionnaire was designed to assess the nature and frequency of bullying and related antisocial behavior. Bullying is commonly understood as repeated negative acts committed by one or more individuals against another individual (Olweus, 1993). These negative acts can be physical or verbal in nature, or they may involve indirect aggression such as social exclusion. Implicit in this definition is an imbalance in real or perceived power between the bully and the victim. Thus, fighting among peers of equal power, generally, is not considered bullying.

The SC bullying prevention study examined the impact of a bullying prevention program on middle school students. Participants included fourth through eighth grade students in six nonmetropolitan school districts in the Southeast. The districts were organized into matched pairs based on geographic location and student demographics. In each pair, one district was selected to receive the intervention for both years of the project (Group A). The other district served as a comparison group for the first year of the project and received the intervention during the second year (Group B). There were 11 Group A schools and 7 Group B schools.

Within each school district, all fourth, fifth, and sixth graders were given a baseline assessment of bullying and antisocial behaviors during the first 2 weeks of March 1995. To assess the effects of the intervention, similar surveys were conducted with the same cohort of students during the first 2 weeks of March for the next 2 years. At baseline, 6,389 students completed the survey. One year later, 6,263 students completed the survey. For the final survey, 3 schools in Group B elected not to participate, resulting in a sample of 4,928 students for the third year.

The Institute for Law and Justice (ILJ), in partnership with LINC in Alexandria, VA, and The Urban Institute in Washington, DC, conducted a study of juvenile violence in the District of Columbia. There were several components of the ILJ study, each focusing on specific objectives laid out in the legislation.

ILJ directed the survey and analyzed the data from a survey of 213 African-American boys, ages 13 to 17, randomly selected from 3 of the highest crime areas in the District. These interviews, undertaken with the assistance of faculty and graduate students from Howard University, provide a wealth of information about the attitudes, victimization, and offending behavior of the boys. Findings from the interviews reflect many of the problems these youth encounter in the District.

The Urban Institute focused on identifying where and when violence involving children was taking place. They used court records of juvenile cases and juvenile victimization records from the Metropolitan Police Department to identify current trends in juvenile offending and victimization, with a particular focus on violent offenses committed by or against juveniles. Juvenile court cases for the 3-year period from 1993 to 1995 were reviewed. The sample of juvenile violent offenders in DC includes all 2,686 juveniles charged with violent crimes in the District between 1993 and 1995. These charges include homicide (n=169), rape (n=66), robbery (n=801), and aggravated assault (n=1,650).

Juvenile homicide victimizations in DC were examined for the 1993-95 time period. There were 51 juvenile homicide victims in 1993, 38 in 1994, and 39 in 1995, for a total of 128 during the time period examined. It is important to note that these numbers are different from those that may be officially reported by the police department. The reason for this is that the age of the victim is often missing in the police records. The Urban Institute research team worked with the police department to develop more accurate information on the ages of the victims.

Nonfatal juvenile violent victimizations in DC were examined for 1993 and 1994. Nonfatal violent acts include rape, robbery, and assault. In 1993, there were 126 juvenile rape victims, 386 juvenile robbery victims, and 1,043 juvenile assault victims. In 1994, there were 140 juvenile rape victims, 374 juvenile robbery victims, and 902 juvenile assault victims.

Researchers at the Social Sciences Research Institute at the University of Southern California examined juvenile violence in the Los Angeles area, with special emphasis on gang violence. Two major components of this study are included in this report.

The first component looked at homicide incidents involving 12- to 17-year-old victims and/or offenders. A 50% sample of cases involving such incidents occurring in 1993-94 was selected from law enforcement agencies in each of three jurisdictions in Los Angeles County. The data were obtained from police department records.

The second component consisted of a household survey. Interviews with youth in neighborhoods with high rates of juvenile violence were undertaken to identify the characteristics and patterns of adolescent violence. Eight Los Angeles County neighborhoods (six in the City of Los Angeles) were included. Representative random samples of residential addresses yielded interviews with 349 boys ages 12 to 17. The age distribution of the sample is evenly distributed, with just 9 percentage points dividing the least frequent age category (17 years: 12% of the sample) to the most frequent (14 years: 21% of the sample). Consistent with the ethnic makeup of the selected neighborhoods, Hispanic (70%) and African-American (28%) youth comprise most of the interview sample.

The Milwaukee Homicide Study examined homicides involving juveniles (ages 13 to 17) and young adults (ages 18 to 24) that occurred in Milwaukee in 1992-93, a period during which homicides were peaking in that city. The study uses life history information obtained from incarcerated offenders and next of kin of victims, along with official record data to analyze differences across types of homicides and examine differences between victims and offenders. Official record data include court and police records, medical examiner records, social service records, and school records. The analysis is supplemented by an analysis of the spatial distribution of Milwaukee homicides over a longer period, from 1989 to 1993.

The sample was drawn from the total number of homicide incidents in Milwaukee in 1992 and 1993, which included a total of 417 offenders and 332 victims (next of kin). Eligible participants for the life history survey included a 40% sample of African-American male offenders and victims and the universe of victims and offenders in the other race/ethnic/gender groups. In order to increase the number of juveniles and young adults in the sample, the sample of eligible participants was weighted such that 50% of the pool of eligibles would represent 15- to 19-year-olds. The pool of eligible offenders was further restricted to include only incarcerated offenders. Of the 123 eligible incarcerated offenders, 86 agreed to be interviewed. Out of the 106 eligible victims, 57 next of kin agreed to be interviewed. The Milwaukee study report findings are based on a sample of juveniles and adults. The Milwaukee researchers provided OJJDP with followup results for some of the analyses, looking only at the juveniles in their sample. This sample comprised 30 juvenile offenders interviewed. Official record data were obtained for 48 juvenile homicide offenders. The findings used in this report to Congress are generally based on the followup analyses, thus limiting the Milwaukee findings in two ways: (1) the small sample size limits the generalizability of the findings, and (2) there are only a limited number of followup results to include in the report.

The Program of Research on the Causes and Correlates of Delinquency, initiated in 1986, includes three coordinated longitudinal projects: the Denver Youth Survey, directed by Dr. David Huizinga at the University of Colorado; the Pittsburgh Youth Study, directed by Dr. Rolf Loeber at the University of Pittsburgh; and the Rochester Youth Development Study, directed by Dr. Terence P. Thornberry at the University at Albany, State University of New York. The Causes and Correlates studies are designed to improve the understanding of serious delinquency, violence, and drug use through an examination of how individual youth develop within the context of family, school, peers, and community. While each of the three projects has unique features, they share several key elements. All of the projects are longitudinal investigations that involve repeated contacts with the same juveniles over a substantial portion of their developmental years.

In each project, researchers conducted face-to-face interviews with individual juveniles in a private setting to collect self-report information on the nature and frequency of serious violent behavior. The advantage of using self-report data, rather than juvenile justice records of arrests, is that researchers can more accurately measure actual violent behaviors and ascertain when a violent career began. Multiple perspectives on each child's development and behavior were obtained through interviews with the child's primary caretaker and, whenever possible, teachers. In addition to interview data, the studies have collected extensive data from official records, such as school, police, and juvenile court. This provides comparison data on the relationship between self-reported behavior and that which is officially detected and recorded.

The three longitudinal studies are prospective in nature. That is, subjects are repeatedly contacted to report on their current and recent violent activities. Deterioration of recall is minimized by avoiding lengthy gaps between interviews. Reporting periods were either 6 or 12 months. Sample retention has been excellent; as of 1997, at least 84% of the subjects had been retained at each of the sites, and the average rate of retention across all waves was 90%.

Samples were carefully drawn to capture inner-city youth considered at high risk for involvement in delinquency and drug abuse. The samples can be described as probability samples, in which youth at greater risk are oversampled.

	Denver's sample includes 1,527 youth (806 males and 721 females) who were 7, 9, 11, 13, and 15 years old when data collection commenced in 1988. This sample represents the general population of youth residing in 20,000 households in high-risk neighborhoods in Denver.
	Pittsburgh's sample consists of 1,517 males who ranged in age from 7 to 13 and attended grades 1, 4, and 7 when data collection began in 1987. This sample represents the general population of males attending Pittsburgh's public schools.
	Rochester's sample of 1,000 youth (729 males and 271 females) was drawn from students attending grades 7 and 8. This sample represents the entire range of seventh and eighth grade students attending Rochester's public schools.