The Future of Children, Princeton - Brookings: Providing research and analysis to promote effective policies and programs for children.

Answering questions about the nature and extent of juvenile violence depends on reliable sources of data. This section reviews the sources for and conclusions based on data that are often used to estimate frequency and prevalence of juvenile violence among various groups and in various settings.

There are four main sources of data on juvenile violence: the National Center for Health Statistics (NCHS), the National Crime Survey, the Uniform Crime Reports (UCRs), and the National Youth Survey (NYS). Some (for example, NCHS) profile victims of violence. Others (for example, UCRs and NYS) profile perpetrators of violent crimes. Most are imperfect, either because they are limited in scope or because they rely on reporting mechanisms that may under or overestimate true rates of violence. Nevertheless, these are the best data available and, when used in combination with one another, can help draw a picture of the extent of violence by juveniles in the United States.²

The National Center for Health Statistics

The National Center for Health Statistics compiles data annually on numbers and rates of homicide, based on reports of medical examiners. Although homicide represents only a small fraction of all violent crimes (slightly more than l%), these figures are quite accurate; therefore, these data are perhaps the most accurate available to gauge trends in the prevalence of violence across population groups, settings, and time. The major source of error in compiling these statistics is the variation in the judgments of medical examiners.

The National Crime Survey

The National Crime Survey, organized by the Bureau of Justice Statistics and conducted by the Census Bureau, gathers interview data on a probability sample of U.S. households to yield a reliable estimate of victimization for nonfatal violence on individuals 12 years of age and older. Excluded from these data are counts of the victimization of children younger than 12 years of age and robberies of businesses and other organizations.

Uniform Crime Reports

In contrast, Uniform Crime Reports (UCRs) are used to characterize perpetrators rather than victims of crimes. UCRs, the most commonly cited source of information on the frequency of crime, are annually compiled data that are based on counts of offenders supplied to the FBI by police agencies throughout the country. These data are prone to sampling biases, changes in police practices, and other sources of measurement error.

Self-Reported Survey Data

The National Youth Survey also provides data concerning perpetrators, but it relies on self-reports of violent behavior rather than on arrest reports.³ This survey, begun in 1976 by investigators at the Behavioral Research Institute in Boulder, Colorado, has followed a national sample of approximately 1,500 youths from adolescence to young adulthood (25 to 31 years of age in 1990). While the validity of these self-report data has been difficult to establish, the data do provide a useful contrast with official police reports. More recently, the Centers for Disease Control and Prevention has begun to collect self-report data on a wide range of health-compromising behavior, including fighting, suicidal actions, and use of weapons.⁴

Arrests

According to the Uniform Crime Report for 1990, 1,749,343 arrests during that year involved youth 17 years of age or younger. Of these, 95,677, or 5.5%, were for violent crimes. This number of juvenile arrests represents 16.3% of arrests for all age groups, a proportion that has increased by 27% since 1980. Most of those arrested in 1990 were male (88%), African-American or Hispanic (60%), and 14 through 17 years of age (80%).⁵

Incarcerations

Currently, more than 95,000 youths are incarcerated in the United States, an increase of more than 20% in less than a decade. Twenty-five percent of these incarcerated youths are held for crimes against persons. Although a relatively small proportion of youths arrested and convicted are responsible for violent offenses, violent incidents are increasing at a high rate for all racial/ethnic and socioeconomic groups of youths.⁶

Self-Reported Violence

According to findings of the National Youth Survey, a substantial minority of adolescents (21%) engage in at least one serious violent incident by age 18, and about 5% engage in multiple violent incidents.³ Based on these findings, one would estimate that 21% of the roughly 29 million youth aged 11 through 17 in the United States, or more than 6 million youths, will have committed a single serious violent offense, and nearly 1.5 million will have engaged in multiple violent offenses by the time they reach age 18. Because, as mentioned above, it is estimated that fewer than 5.5% of all arrests among juveniles in a single year are for violent offenses, these self-reported data may imply that only a fraction of all youth committing a violent offense are apprehended.

Homicides

Homicides are the most salient evidence of the problem of juvenile violence, and the strong quality of the data available concerning homicides makes it the best available candidate to judge trends in violence. Table 1 shows that the proportion of adolescent murder victims killed with a gun increased from 66.5% to 72.4% between 1976 and 1988. Although not listed in the table, the proportion of black victims who were killed by guns (80%) is much higher.⁷ The overall number of adolescent murder victims per year (1,432 in 1988) may not seem large, but it is important to note that homicide represents the leading cause of death for 11 through 17-year-olds.

Information about age, geographic location, minority overrepresentation, secular change, and community context is of particular importance to this discussion because it can provide clues to understanding the causes and solutions to juvenile violence. Key findings are presented here, but greater detail can be obtained from the 1993 report of the National Research Council's Panel on Understanding and Preventing Violence.⁸

Crime and Age

Most violent crimes are committed by offenders who are 17 to 25 years of age, but these violent acts are usually the culminations of long histories of nonviolent offenses.⁹ Once a juvenile has committed a violent crime, the probability increases significantly that he will continue criminal activity of some sort into adulthood.¹⁰

As indicated in Table 2, the arrest rate for violent crimes shows a clear relationship with age for each of four forms of violence: homicide, rape, robbery, and aggravated assault. The arrest rate for robbery is highest in adolescence, while the other types of violence all have a peak age of occurrence in young adulthood. Robbery is the only crime for which there is an overrepresentation of adolescents compared with young adults; while adolescents make up approximately 16% of the population, they account for 24% of robberies. For all four types of violence, the rates for children younger than 15 years of age are quite low, and there is a sharp and progressive decline in rates after age 30. Males represent about 90% of all persons arrested for all types of violent crimes. The proportion who are female ranges from less than 1% for forcible rape to 16% for simple assaults.

Crime Rates and Population Size

Figure 1 illustrates the variation in total violent crime rate by the population size of cities and towns. What is so striking about the figure is the uniformity with which increasing population size is associated with increasing levels of violence. Violent crime increases at each level of city/town size but is most marked for cities with more than one million residents. The relationship between population size and crime rate holds true for homicide, rape, and aggravated assault but is especially pronounced for robbery (not depicted in Figure 1). This implies that rates of robbery may be especially sensitive to the effects of the environment of youthful offenders.

As mentioned above, data on homicide are relatively reliable, so an examination of homicide rates affords a detailed view of the experience of large metropolitan areas with youth violence generally over the past three decades. Homicide has increased, but that increase has typically been limited to key areas within a given city. For example, Washington, D.C., had a homicide rate of 10.6/100,000 in 1960. By 1991, the homicide rate in the nation's capital had reached 79.6/100,000 (the highest in the nation).¹² Washington, D.C. police department records reveal that the proportion of all homicide victims under age 21 rose from 12% in 1986, when the city counted 194 victims overall, to 30% in 1991, when the number of victims was 489. During these same years, among all persons arrested for homicide, the proportion who were under age 21 rose from 21% to 51%.¹²

Despite the heavy concentration of violence in large urban areas, the variation among neighborhoods within a given city may be even more marked than the variation according to city size. Again for Washington, D.C., a few census tracts account for a large proportion of homicides.¹²

Crime Rates Among Different Racial and Ethnic Groups

Minority overrepresentation for all types of crime has become progressively more marked over the past two decades and is evident at each juncture in the criminal justice system, beginning with arrest rates and continuing through sentencing and incarceration.¹³ For example, from 1980 to 1990, the arrest rate for white males aged 15 to 24 has hovered around 12 to 13/100,000, while the rate for black males in the same age range has been nearly 10 times greater.¹⁴ Although black youth make up 12% of the population, they accounted for 24% of arrests for burglaries and 67% of arrests for robberies in a 1987 study of 14 states.¹⁴ Another study indicated that the proportion of youths held in detention among 13 states increased by 15% over the decade of the 1980s, but this increase was only 1% for whites compared with 30% for blacks.¹⁵ Black males aged 10 to 17 make up 4.5% of the general population of California but 34.4% of the juveniles in custody.¹⁵ The rates for Hispanics are usually intermediate between the high rates represented by blacks and the low rates represented by whites at each point in the system. Data comparing Spanish-speaking groups of different national origin, Asians, and Native Americans are either unavailable or imprecise.

The Settings Where Juvenile Violence Occurs

Most data sources do not record the settings where violent incidents among youths occur. This information should be recorded, however, because it can suggest possible interventions. It is unlikely, for example, that all violence by youth occurs in favorite neighborhood gathering places or in protection of gang-designated "turfs." The distribution and selling of illicit drugs constitutes one likely source of violent confrontations, although more generally it may be the recruitment of young people into illegal economies of various types that increases their risk for committing violent acts as well as being victims of violence.

Schools have traditionally been considered safe havens, but the insularity of school settings from violence is eroding in many locales. In a recently conducted survey of public schools in Illinois, for example, nearly 8% of students said they had been physically attacked in or near their school, and another 15% reported only narrowly escaping an attack.¹⁶ One-third of the students in the state's public high schools reported carrying a weapon to school, and 5% reported carrying a gun.

These figures for Illinois are consistent with figures reported in a 1991 survey of public schools in Seattle.¹⁷ In the Seattle survey, 47% of males reported easy access to a gun, 6.4% said that they owned a handgun, and 6% reported that they had carried a gun to school. These handgun owners were more likely to have been members of a gang, charged with assault and battery, suspended or expelled from school, and sentenced by a judge than other students, and one-third reported that they had used the gun to shoot at someone. The rate of gun ownership among female students was only 1.5%. The authors of this study suggest that these rates of access to and use of handguns are underestimates due to sampling limitations and the fact that many of the most deviant children may have already dropped out of school. Because the homicide rate for Seattle is not one of the highest among the larger cities in the country and because it has a relatively small minority population, it is likely that rates of gun possession and use by students are higher in cities where the homicide rate is higher.