You are viewing a Web site, archived on 07:28:45 Oct 16, 2004. It is now a Federal record managed by the National Archives and Records Administration.
External links, forms, and search boxes may not function within this collection.
Public Health Seal report title shim
contents search order resources links home curve end shim

Chapter 2:
The Magnitude of Youth Violence

Measuring Youth Violence

The Violence Epidemic

Arrests for Violent Crimes

Prevalence of Violent Behavior

Differences by Sex and Race/Ethnicity

Violence at School

Gangs and Violence



Appendix 2-A

Chapter 2
The Magnitude of Youth Violence

Headlines proclaim that the epidemic of youth violence that began in the early 1980s is over, but the reality behind this seemingly good news is far more complex and unsettling. Public health studies show that youth violence is an ongoing, startlingly pervasive problem. This chapter describes the magnitude of and trends in violent crime by young people, focusing on homicide, robbery, aggravated assault, and forcible rape (see Box 2-1 for definitions). A later chapter (Chapter 4) seeks to explain why young people become involved in violence in the first place.

Box 2-1. Definitions of the four violent crimes considered in this report

Criminal Homicide—Murder and Non-Negligent Manslaughter
The willful (non-negligent) killing of one human being by another.

The taking or attempting to take anything of value from the care, custody, or control of a person or persons by force or threat of force or violence and/or putting the victim in fear.

Aggravated Assault
An unlawful attack by one person upon another wherein the offender uses a weapon or displays it in a threatening manner, or the victim suffers obvious severe or aggravated bodily injury involving apparent broken bones, loss of teeth, possible internal injury, severe laceration, or loss of consciousness.

Forcible Rape
The carnal knowledge of a person, forcibly and/or against that person's will, or not forcibly or against the person's will where the victim is incapable of giving consent because of his/her temporary or permanent mental or physical incapacity (or because of his/her youth).

Source: U.S. Department of Justice, Federal Bureau of Investigation, 2000.


Surveillance is the backbone of the public health approach to youth violence or any other public health problem. It reveals the magnitude of a problem, tracks the magnitude over time, and uses the information gained from such monitoring to help shape actions to prevent or combat the problem.

Two approaches to measuring the magnitude of youth violence are commonly used. The first relies on official crime statistics compiled by law enforcement agencies, typically arrest reports. These statistics cannot answer questions about how many young people commit violent crimes or how many violent crimes were committed, but they can answer questions about the number of crimes reported to the police, the volume and types of arrests, and how the volume changes over time.

The second approach surveys young people and asks them in confidence about violent acts they have committed or have been victims of during a given period of time. Such reports can be obtained from the same group of people over a long period of time (a longitudinal survey) or from different groups of people at the same point in time (a cross-sectional survey). A prominent example of a repeated cross-sectional survey cited in this chapter is Monitoring the Future, a survey of high school seniors that has been conducted annually since 1975. Reports from young people themselves offer the best way to measure violent behavior that never reaches the attention of the justice system. In fact, evidence in this chapter makes it unmistakably clear that most crimes by young people do not reach the attention of the justice system.

Self-reports are well suited to answering such questions as: What proportion of youths are violent? What types of violent acts do they commit? Has the volume of violence changed over time? Are there differences by sex and race/ethnicity? When during development does violence arise, and what forms does it take? How do children's patterns of violence evolve over time, and how long do they last? These questions relate to the magnitude of violent behavior and to its developmental pathways, and they are addressed in this chapter and the next.

Both arrest reports and self-reports are reasonably valid and reliable ways of measuring the particular aspects of violence they were designed to measure (for general reviews see Blumstein et al., 1986; Cook & Laub, 1998; Elliott & Huizinga, 1989; Hindelang et al., 1981; Huizinga & Elliott, 1986). Arrests appear to be more objective, but they are not a good general measure of violent behavior, for several reasons. First, the majority of aggravated assaults, robberies, and rapes are never reported to the police; arrests are made in fewer than half of reported crimes (Cook & Laub, 1998; Maguire and Pastore, 1999; Snyder & Sickmund, 1999); and most youths involved in violent crimes are never arrested for a violent crime (Elliott et al., 1989; Loeber et al., 1998; Huizinga et al., 1995). Thus, arrests seriously underestimate the volume of violent crime and fail to distinguish accurately between those who are and are not involved in violence. Second, arrest records do not accurately reflect the distribution of reported violent crimes; that is, the offenses for which youths are arrested are not representative of the crimes reported to police (Cook & Laub, 1998). Nonetheless, arrest records are the best measure of the justice system's response to observed or reported crime.

Self-reports were designed specifically to overcome the limitations of violence measures based on official records of criminal behavior. They provide a more direct measure of criminal behavior, but they too have their limitations. Youths may fail to report their violent behavior accurately, either deliberately or because of memory problems, and they may exaggerate their involvement, reporting rather trivial events in response to questions about serious forms of violence. Research reveals that exaggeration (overreporting) is a greater problem than underreporting for reports that cover the previous year (Elliott & Huizinga, 1989), but sophisticated self-report measures can minimize these potential sources of error (Elliott & Huizinga, 1989; Huizinga & Elliott, 1986). The advantages of self-reports are that they capture not only unreported offenses but also details not found in arrest records. In addition, this measure of violent offending is not subject to any of the biases that might be involved in arrest processes.1 The general conclusion from studies evaluating the validity and reliability of self-reports is that they compare favorably with other standard, accepted social science indicators (Hindelang et al., 1981).

Both types of measures contribute to our understanding of violence. The key to using them is to understand their relative strengths and limitations, determine where they reinforce each other and where they diverge or conflict, and then interpret the differences in findings, if possible (Brener et al., 1995; Hindelang et al., 1981; Huizinga & Elliott, 1986; Snyder & Sickmund, 1999).


Arrest rates of young people for homicide and other violent crimes skyrocketed from 1983 to 1993. In response to the dramatic increase in the number of murders committed by young people, Congress and many state legislatures passed new gun control laws, established boot camps, and began waiving children as young as 10 out of the juvenile justice system and into adult criminal courts. Then, starting in the mid-1990s, overall arrest rates began to decline, returning by 1999 to rates only slightly higher than those in 1983.

Several important indicators were used to track youth violence during these years, but their findings did not always agree. Arrest rates, as noted above, provide strong evidence of both a violence epidemic between 1983 and 1993/1994 and a subsequent decline to 1999. Several other indicators of violence furnish similar, but not as robust evidence of a violence epidemic that later subsided. However, the decline in arrest rates is not uniform for all types of violent crime. Moreover, another key indicator—the volume of violent behavior, which is based on self-reports—does not show a decline in youth violence after 1993. As explained later, that indicator remained high and essentially level from 1993 to 1998. This chapter answers the questions raised by these disparate findings—namely, whether the epidemic of violence is really over and why leading indicators of youth violence do not agree.

A rise and subsequent decline in the use of firearms and other weapons by young people provides one potential explanation for the different trends in arrest records and self-reports. The violence epidemic was accompanied by an increase in weapons carrying and use. During this era, instant access to weapons, especially firearms, often turned an angry encounter into a seriously violent or lethal one, which, in turn, drew attention from the police in the form of an arrest. As weapons carrying declined, so too did arrest rates, perhaps because the violence was less injurious or lethal. But the amount of underlying violent behavior (on the basis of self-reports) did not change much—if anything, it appears to have increased in recent years. That undercurrent of violent behavior could reignite into a new epidemic if weapons carrying rises again. From a public health perspective, a resurgence of weapons carrying—and hence the potential for another epidemic of violence—poses a grave threat.


The Federal Bureau of Investigation (FBI) monitors arrests made by law enforcement agencies across the United States through the Uniform Crime Reporting (UCR) program. Since the 1930s, this program has compiled annual arrest information submitted voluntarily by thousands of city, county, and state police agencies. This information currently comes from police jurisdictions that represent only 68 percent of the population, so FBI figures represent projections of these data to the entire U.S. population (Snyder & Sickmund, 1999).

The UCR tabulates the number, rate, and certain features of arrests made by law enforcement agencies. Because some people are arrested more than once a year, the UCR cannot provide an accurate count of the number of people arrested or the proportion of the total population arrested (the prevalence). Nor can the UCR provide an accurate count of the number of crimes committed. A single arrest may account for a series of crimes, or a single crime may involve the arrest of more than one person. Young people tend to commit crimes in groups, so the number of youths arrested inflates the number of crimes committed (Snyder & Sickmund, 1999). As noted earlier, arrest rates are also prone to certain types of error. Unless indicated otherwise, the figures on arrests were assembled by the FBI.

Arrest Rates and Trends
As shown in Figure 2-1, overall arrest rates for violent crimes by youths between the ages of 10 and 17 rose sharply from 1983 to 1993/1994. Rates then declined until 1999, the most recent year for which figures are available.

Figure 2-1. Arrest rates of youths age 10-17 for serious violent crime, 1980-1999

Figure 2-2 shows arrest rates for each of the four violent crimes considered in this report. In 1999, arrests of young people for all crimes totaled 2.4 million (Snyder, unpublished), with 104,000 arrests for violent crimes. Arrests for aggravated assault (69,600) and robbery (28,000) were the most frequent, with arrests for forcible rape (5,000) and murder (1,400) trailing significantly behind. In 1998, youths accounted for one out of six arrests for all violent crimes, a share that has decreased slightly (16 percent) in recent years (Snyder, unpublished). Although the 1999 arrest rate for violent crimes was the lowest in this decade, it is still 15 percent higher than the 1983 rate (Snyder, unpublished). As seen in Figure 2-2, the 1999 rates for homicide, robbery, and rape are below the 1983 rates; however, arrests for aggravated assault are still nearly 70 percent higher than 1983 rates.

Figure 2-2. Arrest rates of youths age 10-17 for serious violent crime, by type of crime, 1980-1999

Figure 2-2. Arrest rates of youths age 10-17 for serious violent crime, by type of crime, 1980-1999
Figure 2-2 (click to enlarge)

The sheer magnitude of the increase in arrest rates between 1983 and 1993/1994 is striking. Overall, arrest rates of youths for violent offenses grew by about 70 percent. The increase in homicides committed by young people was particularly alarming. Both the rate of homicide arrests and the actual number of young people who were arrested for a homicide nearly tripled (Snyder & Sickmund, 1999). This increase was consistent for adolescents at each age between 14 and 17 (Snyder & Sickmund, 1999).

The Role of Firearms
The decade-long upsurge in homicides was tied to an increased use of firearms in the commission of crimes (Cherry et al., 1998; Snyder & Sickmund, 1999). Likewise, the downward trend in homicide arrests from 1993 to 1999 can be traced largely to a decline in firearm usage. The critical role of firearms in homicide and other violent crimes is supported by arrest, victimization, hospitalization, and self-report data.

Analysis of arrest data (Figure 2-3) shows an unequivocal upsurge in firearm usage by young people who committed homicide. In 1983, youths were equally likely to use firearms and other weapons, such as a knife or club, to kill someone. By 1994, 82 percent of homicides by young people were committed with firearms (Snyder & Sickmund, 1999). Virtually all of the increase in firearm-related homicides involved African American youths (Snyder & Sickmund, 1999). The precipitous drop in homicides between 1994 and 1998 coincided with a decline in firearm usage, again mostly by African American youths (Snyder & Sickmund, 1999).

Figure 2-3. Firearm- and nonfirearm-related homicides by youths, 1980-1997

Figure 2-3. Firearm- and nonfirearm-related homicides by youths, 1980-1997
Figure 2-3 (click to enlarge)

Analysis of Supplementary Homicide Report data on young victims of homicide2 reinforces this pattern of firearm use. A large increase in the number of young people killed by firearms between 1987 and 1993 was followed by a decrease. More than 2,000 youths were homicide victims in 1993, the peak year (Snyder & Sickmund, 1999). Most victims were male, and a disproportionately high percentage were African American males (Snyder & Sickmund, 1999).

The use of firearms in violent crimes other than homicide cannot readily be tracked in youth arrest statistics, but for Americans of all ages, firearm use in violent crimes increased from 1985 to 1992 and then declined from 1993 to 1998. Firearm use during robberies increased 33 percent between 1985 and 1992; the decline in firearm use from 1993 to 1998 was nearly 20 percent for aggravated assaults but only 6 percent for robbery (Cook & Laub, 1998; Maguire & Pastore, 1995, 1999).

Firearm use can also be tracked indirectly, through victims treated in hospital emergency departments. Since 1992, injuries related to firearms have been monitored through an emergency department surveillance system.3 Although there are no data from this source to corroborate the growing pattern of firearm injuries before 1992, there are data to corroborate the decline since then. Figure 2-4 presents a special analysis of emergency department surveillance data on youths age 10 to 19. It shows that the rate of firearms-related injuries among young people treated in hospital emergency departments dropped by almost 50 percent from 1993 to 1998. Data on male youths alone reveal a similarly dramatic drop.

Figure 2-4. Nonfatal firearm-related injuries of youths age 10-19 treated in hospital emergency departments, 1993-1998

Figure 2-4. Nonfatal firearm-related injuries of youths age 10-19 treated in hospital emergency departments, 1993-1998
Figure 2-4 (click to enlarge)

In the early 1990s, high school students began to report that they were increasingly less likely to carry guns anywhere and specifically less likely to carry them to school. Figure 2-5 illustrates these trends, as well as trends in general weapons carrying, based on data from the Youth Risk Behavior Survey (YRBS).4 Each trend shows a significant linear decrease, although the decline in weapons carrying in general leveled off in 1999 (Brener et al., 1999; CDC, 2000a; Kann et al., 2000).

Figure 2-5. High school students who carried weapons,* 1991-1999

Figure 2-5. High school students who carried weapons,* 1991-1999
Figure 2-5 (click to enlarge)

Thus, there has been an upsurge and then a decline in the use of firearms and weapons over the past two decades. The easy availability of guns and the resulting rise in lethal violence was caused at least in part by the emerging crack cocaine markets in the mid-1980s and the recruitment of youths into these markets, where carrying guns became routine (Blumstein & Wallman, 2000). It also resulted from changes in the types of guns manufactured, with cheaper, larger caliber guns flooding the gun markets (Wintemute, 2000).

The reasons for the decline are complex and not well understood, but they do involve changes in the carrying and use of guns in violent encounters (Blumstein & Wallman, 2000). The explanations most often given are a decline in youth involvement in the crack market and in gang involvement in crack distribution, police crackdowns on gun carrying and illegal gun purchases, longer sentences for violent crimes involving a gun, a strong economy, and expanded crime and violence prevention programs. After reviewing these and other potential explanations for the drop in violence, Blumstein and Wallman (2000) concluded that no single factor was responsible; rather, the decrease in violence resulted from the combination of many factors.

Comparing Arrests to Other Trends
As noted above, the steep rise and fall in arrest rates over the past two decades has been matched to some extent by changes in leading indicators of violence. Figure 2-6 tracks the trends in four indicators: arrest rates for homicide only, arrest rates for all serious violent crimes, incident rates from victims' self-reports, and incident rates from offenders' self-reports.

The incident rate is a measure of the volume of violence. It refers to the number of self-reported violent acts within a given-sized population—in this case, the number of violent acts per 1,000 young people. In contrast, the prevalence rate indicates what proportion of that population is involved in one or more violent behaviors. Figure 2-6 compares arrest rates with self-reported incident rates (rather than with prevalence rates) because both measure the volume of violent events. Even though arrest and incident rates measure different events and have different absolute magnitudes, the degree of change in these rates over time can be compared.

Arrests Versus Self-Reported Incidents
The sharpest increases in Figure 2-6 are for the two arrest indicators. Homicide arrest rates were roughly 170 percent higher in 1993 than in 1983, and arrest rates for all serious violent crimes were 70 percent higher. The incident rates of serious violent crimes reported by victims and the rates of serious assault and robbery reported by offenders increased to a lesser extent, by about 50 percent.

Figure 2-6. Trends in youth violence since 1983

Figure 2-6. Trends in youth violence since 1983
Figure 2-6 (click to enlarge)

By 1999, arrest rates for homicide, robbery, and rape had dropped below their 1983 levels; by 1997, victim-reported incident rates had dropped back to roughly their 1983 levels. Arrests for aggravated assaults remained high, however—at almost 70 percent above their 1983 level. Since the peak year of arrests for aggravated assault (1994), arrests for this violent crime have declined only 24 percent.

Self-reported violent offending showed no decline at all. After rising by about 50 percent, the incident rate of self-reported serious assaults and robbery remained essentially level through 1998. The leveling off of these rates after 1993 is troubling, for it indicates that the rise and fall in arrest rates are set against a backdrop of ongoing violent behavior. This picture of ongoing violence is borne out by prevalence rates and trends from the Monitoring the Future survey (MTF). Trends in the incident rate of serious violence are shown again in Figure 2-7, this time graphed according to magnitude rather than percentage of change.5

Figure 2-7. Trends in incident rates of serious violence among 12th graders, assault with injury and robbery with a weapon combined, 1980-1998

Figure 2-7. Trends in incident rates of serious violence among 12th graders, assault with injury and robbery with a weapon combined, 1980-1998
Figure 2-7 (click to enlarge)


Questions have been raised about potential racial/ethnic biases in both types of measures. There is evidence that arrests of whites, compared to those of African Americans, are underrepresented in local arrest records and archives (Geerken, 1994). Some studies find racial/ethnic bias in arrests and other justice system processing, while others do not (for reviews, see Austin & Allen, 2000; Hawkins et al., 1998; Sampson & Lauritsen, 1997). Comparisons of an individual's arrest and self-reported offenses reveal a greater discrepancy for African Americans than whites, with African American males self-reporting fewer of the offenses found in their official records (Hindelang et al., 1981; Huizinga & Elliott, 1986). If one accepts the accuracy of arrest records, this finding would indicate an underreporting on the part of African American males, but there are reasons to question this assumption (see Elliott, 1982; Huizinga & Elliott, 1986). The question of racial/ethnic bias in both measures remains controversial.
Youths are victims in about 27 percent of homicides committed by other youths (Snyder & Sickmund, 1999).
The National Electronic Injury Surveillance System (NEISS); NEISS is operated by the U.S. Consumer Product Safety Commission.
Begun in 1990, the YRBS is a national school-based survey conducted every 2 years by the Centers for Disease Control and Prevention in collaboration with Federal, state, and local partners. It is representative of students in grades 9 through 12 in both public and private schools. YRBS monitors six important health behaviors, including those that may result in violent injuries. The survey is voluntary, anonymous, provides for parental consent for minors, and oversamples minorities (Kolbe et al., 1993). The 1999 survey included more than 15,000 respondents (Kann et al., 2000).
This self-reported incident rate appears to be much higher (e.g., almost 400 assaults with injury and robberies with a weapon were reported per 1,000 high school seniors in 1998) than the arrest rate for aggravated assault and robbery (about 350 arrests per 100,000 youth, see Figure 2-2), but the two are not strictly comparable: high school seniors (17- and 18-year-olds) have much higher arrest rates as a group than do 10- to 17-year-olds.

Back to Top

Home | Contents | Previous | Next